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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 744 - (show annotations) (download)
Sun Nov 13 16:31:38 2011 UTC (2 years, 9 months ago) by zherczeg
File size: 218279 byte(s)
Correctly supporting \x and \u in JavaScript compatibility mode
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.21
5 ------------
6
7 1. Updating the JIT compiler.
8
9 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10 are added as well.
11
12 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14 calling _pcre_jit_exec. Some extra comments are added.
15
16 4. Mark settings inside atomic groups that do not contain any capturing
17 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18 was introduced by change 18 for 8.20.
19
20 5. Supporting of \x and \u in JavaScript compatibility mode based on the
21 ECMA-262 standard.
22
23
24 Version 8.20 21-Oct-2011
25 ------------------------
26
27 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
28 a POSIX class. After further experiments with Perl, which convinced me that
29 Perl has bugs and confusions, a closing square bracket is no longer allowed
30 in a POSIX name. This bug also affected patterns with classes that started
31 with full stops.
32
33 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
34 captured substring, but while checking the failing first alternative,
35 substring 1 is temporarily captured. If the output vector supplied to
36 pcre_exec() was not big enough for this capture, the yield of the function
37 was still zero ("insufficient space for captured substrings"). This cannot
38 be totally fixed without adding another stack variable, which seems a lot
39 of expense for a edge case. However, I have improved the situation in cases
40 such as /(a)(b)x|abc/ matched against "abc", where the return code
41 indicates that fewer than the maximum number of slots in the ovector have
42 been set.
43
44 3. Related to (2) above: when there are more back references in a pattern than
45 slots in the output vector, pcre_exec() uses temporary memory during
46 matching, and copies in the captures as far as possible afterwards. It was
47 using the entire output vector, but this conflicts with the specification
48 that only 2/3 is used for passing back captured substrings. Now it uses
49 only the first 2/3, for compatibility. This is, of course, another edge
50 case.
51
52 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
53 main code base, and can be used by building with --enable-jit. When this is
54 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
55 runtime --no-jit option is given.
56
57 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
58 ovector, the return from the function was zero, implying that there were
59 other matches that did not fit. The correct "exactly full" value is now
60 returned.
61
62 6. If a subpattern that was called recursively or as a subroutine contained
63 (*PRUNE) or any other control that caused it to give a non-standard return,
64 invalid errors such as "Error -26 (nested recursion at the same subject
65 position)" or even infinite loops could occur.
66
67 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
68 computing the minimum length on reaching *ACCEPT, and so ended up with the
69 wrong value of 1 rather than 0. Further investigation indicates that
70 computing a minimum subject length in the presence of *ACCEPT is difficult
71 (think back references, subroutine calls), and so I have changed the code
72 so that no minimum is registered for a pattern that contains *ACCEPT.
73
74 8. If (*THEN) was present in the first (true) branch of a conditional group,
75 it was not handled as intended. [But see 16 below.]
76
77 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
78 Sheri Pierce.
79
80 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
81 the first byte in a match must be "a".
82
83 11. Change 17 for 8.13 increased the recursion depth for patterns like
84 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
85 pattern contains any instances of (*THEN). If it does not, the old
86 optimizations are restored. It would be nice to do this on a per-group
87 basis, but at the moment that is not feasible.
88
89 12. In some environments, the output of pcretest -C is CRLF terminated. This
90 broke RunTest's code that checks for the link size. A single white space
91 character after the value is now allowed for.
92
93 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
94 For "fr", it uses the Windows-specific input and output files.
95
96 14. If (*THEN) appeared in a group that was called recursively or as a
97 subroutine, it did not work as intended. [But see next item.]
98
99 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
100 pattern fragments (but not containing any | characters). If A and B are
101 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
102 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
103 D. In other words, Perl considers parentheses that do not contain any |
104 characters to be part of a surrounding alternative, whereas PCRE was
105 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
106 differently. PCRE now behaves in the same way as Perl, except in the case
107 of subroutine/recursion calls such as (?1) which have in any case always
108 been different (but PCRE had them first :-).
109
110 16. Related to 15 above: Perl does not treat the | in a conditional group as
111 creating alternatives. Such a group is treated in the same way as an
112 ordinary group without any | characters when processing (*THEN). PCRE has
113 been changed to match Perl's behaviour.
114
115 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
116 RunGrepTest script failed.
117
118 18. Change 22 for version 13 caused atomic groups to use more stack. This is
119 inevitable for groups that contain captures, but it can lead to a lot of
120 stack use in large patterns. The old behaviour has been restored for atomic
121 groups that do not contain any capturing parentheses.
122
123 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
124 suppress the check for a minimum subject length at run time. (If it was
125 given to pcre_exec() or pcre_dfa_exec() it did work.)
126
127 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
128 fail to work when decoding hex characters in data strings in EBCDIC
129 environments.
130
131 21. It appears that in at least one Mac OS environment, the isxdigit() function
132 is implemented as a macro that evaluates to its argument more than once,
133 contravening the C 90 Standard (I haven't checked a later standard). There
134 was an instance in pcretest which caused it to go wrong when processing
135 \x{...} escapes in subject strings. The has been rewritten to avoid using
136 things like p++ in the argument of isxdigit().
137
138
139 Version 8.13 16-Aug-2011
140 ------------------------
141
142 1. The Unicode data tables have been updated to Unicode 6.0.0.
143
144 2. Two minor typos in pcre_internal.h have been fixed.
145
146 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
147 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
148 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
149
150 4. There were a number of related bugs in the code for matching backrefences
151 caselessly in UTF-8 mode when codes for the characters concerned were
152 different numbers of bytes. For example, U+023A and U+2C65 are an upper
153 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
154 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
155 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
156 2-byte code at the end of the subject (it thought there wasn't enough data
157 left).
158
159 5. Comprehensive information about what went wrong is now returned by
160 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
161 as the output vector has at least 2 elements. The offset of the start of
162 the failing character and a reason code are placed in the vector.
163
164 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
165 now returned is for the first byte of the failing character, instead of the
166 last byte inspected. This is an incompatible change, but I hope it is small
167 enough not to be a problem. It makes the returned offset consistent with
168 pcre_exec() and pcre_dfa_exec().
169
170 7. pcretest now gives a text phrase as well as the error number when
171 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
172 failure, the offset and reason code are output.
173
174 8. When \R was used with a maximizing quantifier it failed to skip backwards
175 over a \r\n pair if the subsequent match failed. Instead, it just skipped
176 back over a single character (\n). This seems wrong (because it treated the
177 two characters as a single entity when going forwards), conflicts with the
178 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
179 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
180 has been changed.
181
182 9. Some internal refactoring has changed the processing so that the handling
183 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
184 time (the PCRE_DOTALL option was changed this way some time ago: version
185 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
186 which was always a bit of a fudge. It also means that there is one less
187 argument for the match() function, which reduces its stack requirements
188 slightly. This change also fixes an incompatibility with Perl: the pattern
189 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
190
191 10. More internal refactoring has drastically reduced the number of recursive
192 calls to match() for possessively repeated groups such as (abc)++ when
193 using pcre_exec().
194
195 11. While implementing 10, a number of bugs in the handling of groups were
196 discovered and fixed:
197
198 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
199 (a|)*(?1) gave a compile-time internal error.
200 ((a|)+)+ did not notice that the outer group could match an empty string.
201 (^a|^)+ was not marked as anchored.
202 (.*a|.*)+ was not marked as matching at start or after a newline.
203
204 12. Yet more internal refactoring has removed another argument from the match()
205 function. Special calls to this function are now indicated by setting a
206 value in a variable in the "match data" data block.
207
208 13. Be more explicit in pcre_study() instead of relying on "default" for
209 opcodes that mean there is no starting character; this means that when new
210 ones are added and accidentally left out of pcre_study(), testing should
211 pick them up.
212
213 14. The -s option of pcretest has been documented for ages as being an old
214 synonym of -m (show memory usage). I have changed it to mean "force study
215 for every regex", that is, assume /S for every regex. This is similar to -i
216 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
217 using it. It makes it easier to run collections of tests with and without
218 study enabled, and thereby test pcre_study() more easily. All the standard
219 tests are now run with and without -s (but some patterns can be marked as
220 "never study" - see 20 below).
221
222 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
223 restoration of the capturing data to the outer values was not happening
224 correctly.
225
226 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
227 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
228 pattern had matched an empty string, and so incorrectly returned a no
229 match.
230
231 17. There was optimizing code for the last branch of non-capturing parentheses,
232 and also for the obeyed branch of a conditional subexpression, which used
233 tail recursion to cut down on stack usage. Unfortunately, now that there is
234 the possibility of (*THEN) occurring in these branches, tail recursion is
235 no longer possible because the return has to be checked for (*THEN). These
236 two optimizations have therefore been removed. [But see 8.20/11 above.]
237
238 18. If a pattern containing \R was studied, it was assumed that \R always
239 matched two bytes, thus causing the minimum subject length to be
240 incorrectly computed because \R can also match just one byte.
241
242 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
243 was incorrectly computed.
244
245 20. If /S is present twice on a test pattern in pcretest input, it now
246 *disables* studying, thereby overriding the use of -s on the command line
247 (see 14 above). This is necessary for one or two tests to keep the output
248 identical in both cases.
249
250 21. When (*ACCEPT) was used in an assertion that matched an empty string and
251 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
252
253 22. When an atomic group that contained a capturing parenthesis was
254 successfully matched, but the branch in which it appeared failed, the
255 capturing was not being forgotten if a higher numbered group was later
256 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
257 group 1 to "a", when in fact it should be unset. This applied to multi-
258 branched capturing and non-capturing groups, repeated or not, and also to
259 positive assertions (capturing in negative assertions does not happen
260 in PCRE) and also to nested atomic groups.
261
262 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
263 subject after a captured substring, to make it easier to tell which of a
264 number of identical substrings has been captured.
265
266 24. The way atomic groups are processed by pcre_exec() has been changed so that
267 if they are repeated, backtracking one repetition now resets captured
268 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
269 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
270 "aaa". Previously, it would have been "a". As part of this code
271 refactoring, the way recursive calls are handled has also been changed.
272
273 25. If an assertion condition captured any substrings, they were not passed
274 back unless some other capturing happened later. For example, if
275 (?(?=(a))a) was matched against "a", no capturing was returned.
276
277 26. When studying a pattern that contained subroutine calls or assertions,
278 the code for finding the minimum length of a possible match was handling
279 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
280 group 1 called group 2 while simultaneously a separate group 2 called group
281 1). A stack overflow occurred in this case. I have fixed this by limiting
282 the recursion depth to 10.
283
284 27. Updated RunTest.bat in the distribution to the version supplied by Tom
285 Fortmann. This supports explicit test numbers on the command line, and has
286 argument validation and error reporting.
287
288 28. An instance of \X with an unlimited repeat could fail if at any point the
289 first character it looked at was a mark character.
290
291 29. Some minor code refactoring concerning Unicode properties and scripts
292 should reduce the stack requirement of match() slightly.
293
294 30. Added the '=' option to pcretest to check the setting of unused capturing
295 slots at the end of the pattern, which are documented as being -1, but are
296 not included in the return count.
297
298 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
299 compiled something random. Now it gives a compile-time error (as does
300 Perl).
301
302 32. A *MARK encountered during the processing of a positive assertion is now
303 recorded and passed back (compatible with Perl).
304
305 33. If --only-matching or --colour was set on a pcregrep call whose pattern
306 had alternative anchored branches, the search for a second match in a line
307 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
308 matched the line "0102" twice. The same bug affected patterns that started
309 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
310 twice.
311
312 34. Previously, PCRE did not allow quantification of assertions. However, Perl
313 does, and because of capturing effects, quantifying parenthesized
314 assertions may at times be useful. Quantifiers are now allowed for
315 parenthesized assertions.
316
317 35. A minor code tidy in pcre_compile() when checking options for \R usage.
318
319 36. \g was being checked for fancy things in a character class, when it should
320 just be a literal "g".
321
322 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
323 appearance of a nested POSIX class supersedes an apparent external class.
324 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
325 unescaped square brackets may also appear as part of class names. For
326 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
327 more like Perl. (But see 8.20/1 above.)
328
329 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
330 was because it thought it was \N{name}, which is not supported).
331
332 39. Add minix to OS list not supporting the -S option in pcretest.
333
334 40. PCRE tries to detect cases of infinite recursion at compile time, but it
335 cannot analyze patterns in sufficient detail to catch mutual recursions
336 such as ((?1))((?2)). There is now a runtime test that gives an error if a
337 subgroup is called recursively as a subpattern for a second time at the
338 same position in the subject string. In previous releases this might have
339 been caught by the recursion limit, or it might have run out of stack.
340
341 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
342 happen only once. PCRE was, however incorrectly giving a compile time error
343 "recursive call could loop indefinitely" because it cannot analyze the
344 pattern in sufficient detail. The compile time test no longer happens when
345 PCRE is compiling a conditional subpattern, but actual runaway loops are
346 now caught at runtime (see 40 above).
347
348 42. It seems that Perl allows any characters other than a closing parenthesis
349 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
350 has been changed to be the same.
351
352 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
353 as not to get warnings when autogen.sh is called. Also changed
354 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
355
356 44. To help people who use pcregrep to scan files containing exceedingly long
357 lines, the following changes have been made:
358
359 (a) The default value of the buffer size parameter has been increased from
360 8K to 20K. (The actual buffer used is three times this size.)
361
362 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
363 PCRE is built.
364
365 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
366 to be set at run time.
367
368 (d) Numerical values in pcregrep options can be followed by K or M, for
369 example --buffer-size=50K.
370
371 (e) If a line being scanned overflows pcregrep's buffer, an error is now
372 given and the return code is set to 2.
373
374 45. Add a pointer to the latest mark to the callout data block.
375
376 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
377 partial match of an empty string instead of no match. This was specific to
378 the use of ".".
379
380 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
381 complete match instead of a partial match. This bug was dependent on both
382 the PCRE_UTF8 and PCRE_DOTALL options being set.
383
384 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
385 starting byte set, because \b was not being ignored.
386
387
388 Version 8.12 15-Jan-2011
389 ------------------------
390
391 1. Fixed some typos in the markup of the man pages, and wrote a script that
392 checks for such things as part of the documentation building process.
393
394 2. On a big-endian 64-bit system, pcregrep did not correctly process the
395 --match-limit and --recursion-limit options (added for 8.11). In
396 particular, this made one of the standard tests fail. (The integer value
397 went into the wrong half of a long int.)
398
399 3. If the --colour option was given to pcregrep with -v (invert match), it
400 did strange things, either producing crazy output, or crashing. It should,
401 of course, ignore a request for colour when reporting lines that do not
402 match.
403
404 4. Another pcregrep bug caused similar problems if --colour was specified with
405 -M (multiline) and the pattern match finished with a line ending.
406
407 5. In pcregrep, when a pattern that ended with a literal newline sequence was
408 matched in multiline mode, the following line was shown as part of the
409 match. This seems wrong, so I have changed it.
410
411 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
412 the check for further matches in the same line (so they could be coloured)
413 to overrun the end of the current line. If another match was found, it was
414 incorrectly shown (and then shown again when found in the next line).
415
416 7. If pcregrep was compiled under Windows, there was a reference to the
417 function pcregrep_exit() before it was defined. I am assuming this was
418 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
419 reported by a user. I've moved the definition above the reference.
420
421
422 Version 8.11 10-Dec-2010
423 ------------------------
424
425 1. (*THEN) was not working properly if there were untried alternatives prior
426 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
427 backtracked to try for "b" instead of moving to the next alternative branch
428 at the same level (in this case, to look for "c"). The Perl documentation
429 is clear that when (*THEN) is backtracked onto, it goes to the "next
430 alternative in the innermost enclosing group".
431
432 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
433 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
434 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
435 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
436 (*THEN).
437
438 3. If \s appeared in a character class, it removed the VT character from
439 the class, even if it had been included by some previous item, for example
440 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
441 of \s, but is part of the POSIX "space" class.)
442
443 4. A partial match never returns an empty string (because you can always
444 match an empty string at the end of the subject); however the checking for
445 an empty string was starting at the "start of match" point. This has been
446 changed to the "earliest inspected character" point, because the returned
447 data for a partial match starts at this character. This means that, for
448 example, /(?<=abc)def/ gives a partial match for the subject "abc"
449 (previously it gave "no match").
450
451 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
452 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
453 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
454 has an implication that the given string is incomplete (because a partial
455 match is preferred over a full match). For this reason, these items now
456 give a partial match in this situation. [Aside: previously, the one case
457 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
458 match rather than a full match, which was wrong by the old rules, but is
459 now correct.]
460
461 6. There was a bug in the handling of #-introduced comments, recognized when
462 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
463 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
464 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
465 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
466 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
467 places in pcre_compile().
468
469 7. Related to (6) above, when pcre_compile() was skipping #-introduced
470 comments when looking ahead for named forward references to subpatterns,
471 the only newline sequence it recognized was NL. It now handles newlines
472 according to the set newline convention.
473
474 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
475 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
476 cater for a lack of strerror(). These oversights have been fixed.
477
478 9. Added --match-limit and --recursion-limit to pcregrep.
479
480 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
481
482 11. When the -o option was used, pcregrep was setting a return code of 1, even
483 when matches were found, and --line-buffered was not being honoured.
484
485 12. Added an optional parentheses number to the -o and --only-matching options
486 of pcregrep.
487
488 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
489 can match an empty string. The code to do it in pcretest and pcredemo
490 needed fixing:
491
492 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
493 only one byte after an empty string match just before CRLF (this case
494 just got forgotten; "any" and "anycrlf" were OK).
495
496 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
497 mode when an empty string match preceded an ASCII character followed by
498 a non-ASCII character. (The code for advancing by one character rather
499 than one byte was nonsense.)
500
501 (c) The pcredemo.c sample program did not have any code at all to handle
502 the cases when CRLF is a valid newline sequence.
503
504 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
505 as a starting offset was within the subject string. There is now a new
506 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
507 negative or greater than the length of the string. In order to test this,
508 pcretest is extended to allow the setting of negative starting offsets.
509
510 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
511 starting offset points to the beginning of a UTF-8 character was
512 unnecessarily clumsy. I tidied it up.
513
514 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
515 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
516
517 17. Nobody had reported that the --include_dir option, which was added in
518 release 7.7 should have been called --include-dir (hyphen, not underscore)
519 for compatibility with GNU grep. I have changed it to --include-dir, but
520 left --include_dir as an undocumented synonym, and the same for
521 --exclude-dir, though that is not available in GNU grep, at least as of
522 release 2.5.4.
523
524 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
525 characters from a string of bytes) have been redefined so as not to use
526 loops, in order to improve performance in some environments. At the same
527 time, I abstracted some of the common code into auxiliary macros to save
528 repetition (this should not affect the compiled code).
529
530 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
531 compile-time error is now given if \c is not followed by an ASCII
532 character, that is, a byte less than 128. (In EBCDIC mode, the code is
533 different, and any byte value is allowed.)
534
535 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
536 START_OPTIMIZE option, which is now allowed at compile time - but just
537 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
538 to pcregrep and other applications that have no direct access to PCRE
539 options. The new /Y option in pcretest sets this option when calling
540 pcre_compile().
541
542 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
543 back references. Groups containing recursive back references were forced to
544 be atomic by that change, but in the case of named groups, the amount of
545 memory required was incorrectly computed, leading to "Failed: internal
546 error: code overflow". This has been fixed.
547
548 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
549 pcretest.c, to avoid build problems in some Borland environments.
550
551
552 Version 8.10 25-Jun-2010
553 ------------------------
554
555 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
556 THEN.
557
558 2. (*ACCEPT) was not working when inside an atomic group.
559
560 3. Inside a character class, \B is treated as a literal by default, but
561 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
562 causes the error). The code is unchanged, but I tidied the documentation.
563
564 4. Inside a character class, PCRE always treated \R and \X as literals,
565 whereas Perl faults them if its -w option is set. I have changed PCRE so
566 that it faults them when PCRE_EXTRA is set.
567
568 5. Added support for \N, which always matches any character other than
569 newline. (It is the same as "." when PCRE_DOTALL is not set.)
570
571 6. When compiling pcregrep with newer versions of gcc which may have
572 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
573 declared with attribute warn_unused_result" were given. Just casting the
574 result to (void) does not stop the warnings; a more elaborate fudge is
575 needed. I've used a macro to implement this.
576
577 7. Minor change to pcretest.c to avoid a compiler warning.
578
579 8. Added four artifical Unicode properties to help with an option to make
580 \s etc use properties (see next item). The new properties are: Xan
581 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
582
583 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
584 use Unicode properties. (*UCP) at the start of a pattern can be used to set
585 this option. Modified pcretest to add /W to test this facility. Added
586 REG_UCP to make it available via the POSIX interface.
587
588 10. Added --line-buffered to pcregrep.
589
590 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
591 studied, and the match started with a letter with a code point greater than
592 127 whose first byte was different to the first byte of the other case of
593 the letter, the other case of this starting letter was not recognized
594 (#976).
595
596 12. If a pattern that was studied started with a repeated Unicode property
597 test, for example, \p{Nd}+, there was the theoretical possibility of
598 setting up an incorrect bitmap of starting bytes, but fortunately it could
599 not have actually happened in practice until change 8 above was made (it
600 added property types that matched character-matching opcodes).
601
602 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
603 possible starting bytes for non-anchored patterns.
604
605 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
606 \R, and also a number of cases that involve Unicode properties, both
607 explicit and implicit when PCRE_UCP is set.
608
609 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
610 input, it could crash or give wrong results if characters with values
611 greater than 0xc0 were present in the subject string. (Detail: it assumed
612 UTF-8 input when processing these items.)
613
614 16. Added a lot of (int) casts to avoid compiler warnings in systems where
615 size_t is 64-bit (#991).
616
617 17. Added a check for running out of memory when PCRE is compiled with
618 --disable-stack-for-recursion (#990).
619
620 18. If the last data line in a file for pcretest does not have a newline on
621 the end, a newline was missing in the output.
622
623 19. The default pcre_chartables.c file recognizes only ASCII characters (values
624 less than 128) in its various bitmaps. However, there is a facility for
625 generating tables according to the current locale when PCRE is compiled. It
626 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
627 characters, are recognized by isspace() and therefore were getting set in
628 these tables, and indeed these tables seem to approximate to ISO 8859. This
629 caused a problem in UTF-8 mode when pcre_study() was used to create a list
630 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
631 which of course cannot start UTF-8 characters. I have changed the code so
632 that only real ASCII characters (less than 128) and the correct starting
633 bytes for UTF-8 encodings are set for characters greater than 127 when in
634 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
635 altogether.)
636
637 20. Added the /T option to pcretest so as to be able to run tests with non-
638 standard character tables, thus making it possible to include the tests
639 used for 19 above in the standard set of tests.
640
641 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
642 reference to a subpattern the other side of a comment that contains an
643 opening parenthesis caused either an internal compiling error, or a
644 reference to the wrong subpattern.
645
646
647 Version 8.02 19-Mar-2010
648 ------------------------
649
650 1. The Unicode data tables have been updated to Unicode 5.2.0.
651
652 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
653 configured.
654
655 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
656 original author of that file, following a query about its status.
657
658 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
659 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
660
661 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
662 quantifier applied to a forward-referencing subroutine call, could compile
663 incorrect code or give the error "internal error: previously-checked
664 referenced subpattern not found".
665
666 6. Both MS Visual Studio and Symbian OS have problems with initializing
667 variables to point to external functions. For these systems, therefore,
668 pcre_malloc etc. are now initialized to local functions that call the
669 relevant global functions.
670
671 7. There were two entries missing in the vectors called coptable and poptable
672 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
673 I've fixed the data, and added a kludgy way of testing at compile time that
674 the lengths are correct (equal to the number of opcodes).
675
676 8. Following on from 7, I added a similar kludge to check the length of the
677 eint vector in pcreposix.c.
678
679 9. Error texts for pcre_compile() are held as one long string to avoid too
680 much relocation at load time. To find a text, the string is searched,
681 counting zeros. There was no check for running off the end of the string,
682 which could happen if a new error number was added without updating the
683 string.
684
685 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
686
687 11. \K was not working if it appeared in an atomic group or in a group that
688 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
689 \K is "not well defined" if used in an assertion. PCRE now accepts it if
690 the assertion is positive, but not if it is negative.
691
692 12. Change 11 fortuitously reduced the size of the stack frame used in the
693 "match()" function of pcre_exec.c by one pointer. Forthcoming
694 implementation of support for (*MARK) will need an extra pointer on the
695 stack; I have reserved it now, so that the stack frame size does not
696 decrease.
697
698 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
699 item in branch that calls a recursion is a subroutine call - as in the
700 second branch in the above example - was incorrectly given the compile-
701 time error "recursive call could loop indefinitely" because pcre_compile()
702 was not correctly checking the subroutine for matching a non-empty string.
703
704 14. The checks for overrunning compiling workspace could trigger after an
705 overrun had occurred. This is a "should never occur" error, but it can be
706 triggered by pathological patterns such as hundreds of nested parentheses.
707 The checks now trigger 100 bytes before the end of the workspace.
708
709 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
710
711
712 Version 8.01 19-Jan-2010
713 ------------------------
714
715 1. If a pattern contained a conditional subpattern with only one branch (in
716 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
717 computed the wrong minimum data length (which is of course zero for such
718 subpatterns). This could cause incorrect "no match" results.
719
720 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
721 the pattern is reset in the first branch, pcre_compile() failed with
722 "internal error: code overflow at offset...". This happened only when
723 the reset was to the original external option setting. (An optimization
724 abstracts leading options settings into an external setting, which was the
725 cause of this.)
726
727 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
728 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
729 assertion pattern did not match (meaning that the assertion was true), it
730 was incorrectly treated as false if the SKIP had been reached during the
731 matching. This also applied to assertions used as conditions.
732
733 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
734 assertion subpattern, including such a pattern used as a condition,
735 unpredictable results occurred, instead of the error return
736 PCRE_ERROR_DFA_UITEM.
737
738 5. The C++ GlobalReplace function was not working like Perl for the special
739 situation when an empty string is matched. It now does the fancy magic
740 stuff that is necessary.
741
742 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
743 removed. (These were left over from very, very early versions of PCRE.)
744
745 7. Some cosmetic changes to the code to make life easier when compiling it
746 as part of something else:
747
748 (a) Change DEBUG to PCRE_DEBUG.
749
750 (b) In pcre_compile(), rename the member of the "branch_chain" structure
751 called "current" as "current_branch", to prevent a collision with the
752 Linux macro when compiled as a kernel module.
753
754 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
755 prevent a collision with the Linux macro when compiled as a kernel
756 module.
757
758 8. In pcre_compile() there are some checks for integer overflows that used to
759 cast potentially large values to (double). This has been changed to that
760 when building, a check for int64_t is made, and if it is found, it is used
761 instead, thus avoiding the use of floating point arithmetic. (There is no
762 other use of FP in PCRE.) If int64_t is not found, the fallback is to
763 double.
764
765 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
766 2005 (difference between two addresses compared to an unsigned value).
767
768 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
769 custom one, because of the following reported problem in Windows:
770
771 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
772 under Win32.
773 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
774 therefore missing the function definition.
775 - The compiler thus generates a "C" signature for the test function.
776 - The linker fails to find the "C" function.
777 - PCRE fails to configure if asked to do so against libbz2.
778
779 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
780 messages were output:
781
782 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
783 rerunning libtoolize, to keep the correct libtool macros in-tree.
784 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
785
786 I have done both of these things.
787
788 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
789 most of the time, it *can* run out if it is given a pattern that contains a
790 runaway infinite recursion. I updated the discussion in the pcrestack man
791 page.
792
793 13. Now that we have gone to the x.xx style of version numbers, the minor
794 version may start with zero. Using 08 or 09 is a bad idea because users
795 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
796 interpreted as invalid octal numbers. I've updated the previous comment in
797 configure.ac, and also added a check that gives an error if 08 or 09 are
798 used.
799
800 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
801 causing partial matching to fail when the end of the subject matched \W
802 in a UTF-8 pattern where \W was quantified with a minimum of 3.
803
804 15. There were some discrepancies between the declarations in pcre_internal.h
805 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
806 their definitions. The declarations used "const uschar *" and the
807 definitions used USPTR. Even though USPTR is normally defined as "const
808 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
809 reported that: "This difference in casting confuses some C++ compilers, for
810 example, SunCC recognizes above declarations as different functions and
811 generates broken code for hbpcre." I have changed the declarations to use
812 USPTR.
813
814 16. GNU libtool is named differently on some systems. The autogen.sh script now
815 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
816 (FreeBSD).
817
818 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
819 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
820 comment: "Figure out how to create a longlong from a string: strtoll and
821 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
822 instance, but it only takes 2 args instead of 3!"
823
824 18. A subtle bug concerned with back references has been fixed by a change of
825 specification, with a corresponding code fix. A pattern such as
826 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
827 refers, was giving matches when it shouldn't. For example, xa=xaaa would
828 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
829 same bug. Such groups have to be quantified to be useful, or contained
830 inside another quantified group. (If there's no repetition, the reference
831 can never match.) The problem arises because, having left the group and
832 moved on to the rest of the pattern, a later failure that backtracks into
833 the group uses the captured value from the final iteration of the group
834 rather than the correct earlier one. I have fixed this in PCRE by forcing
835 any group that contains a reference to itself to be an atomic group; that
836 is, there cannot be any backtracking into it once it has completed. This is
837 similar to recursive and subroutine calls.
838
839
840 Version 8.00 19-Oct-09
841 ----------------------
842
843 1. The table for translating pcre_compile() error codes into POSIX error codes
844 was out-of-date, and there was no check on the pcre_compile() error code
845 being within the table. This could lead to an OK return being given in
846 error.
847
848 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
849 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
850 in a Windows environment.
851
852 3. The pcregrep --count option prints the count for each file even when it is
853 zero, as does GNU grep. However, pcregrep was also printing all files when
854 --files-with-matches was added. Now, when both options are given, it prints
855 counts only for those files that have at least one match. (GNU grep just
856 prints the file name in this circumstance, but including the count seems
857 more useful - otherwise, why use --count?) Also ensured that the
858 combination -clh just lists non-zero counts, with no names.
859
860 4. The long form of the pcregrep -F option was incorrectly implemented as
861 --fixed_strings instead of --fixed-strings. This is an incompatible change,
862 but it seems right to fix it, and I didn't think it was worth preserving
863 the old behaviour.
864
865 5. The command line items --regex=pattern and --regexp=pattern were not
866 recognized by pcregrep, which required --regex pattern or --regexp pattern
867 (with a space rather than an '='). The man page documented the '=' forms,
868 which are compatible with GNU grep; these now work.
869
870 6. No libpcreposix.pc file was created for pkg-config; there was just
871 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
872
873 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
874 when UCP support is not needed, by modifying the Python script that
875 generates it from Unicode data files. This should not matter if the module
876 is correctly used as a library, but I received one complaint about 50K of
877 unwanted data. My guess is that the person linked everything into his
878 program rather than using a library. Anyway, it does no harm.
879
880 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
881 was a minimum greater than 1 for a wide character in a possessive
882 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
883 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
884 character. Chaos in the form of incorrect output or a compiling loop could
885 result.
886
887 9. The restrictions on what a pattern can contain when partial matching is
888 requested for pcre_exec() have been removed. All patterns can now be
889 partially matched by this function. In addition, if there are at least two
890 slots in the offset vector, the offset of the earliest inspected character
891 for the match and the offset of the end of the subject are set in them when
892 PCRE_ERROR_PARTIAL is returned.
893
894 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
895 synonymous with PCRE_PARTIAL, for backwards compatibility, and
896 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
897 and may be more useful for multi-segment matching.
898
899 11. Partial matching with pcre_exec() is now more intuitive. A partial match
900 used to be given if ever the end of the subject was reached; now it is
901 given only if matching could not proceed because another character was
902 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
903 string "Z", which now yields "no match" instead of "partial match". In the
904 case of pcre_dfa_exec(), "no match" is given if every matching path for the
905 final character ended with (*FAIL).
906
907 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
908 if the pattern had a "must contain" character that was already found in the
909 earlier partial match, unless partial matching was again requested. For
910 example, with the pattern /dog.(body)?/, the "must contain" character is
911 "g". If the first part-match was for the string "dog", restarting with
912 "sbody" failed. This bug has been fixed.
913
914 13. The string returned by pcre_dfa_exec() after a partial match has been
915 changed so that it starts at the first inspected character rather than the
916 first character of the match. This makes a difference only if the pattern
917 starts with a lookbehind assertion or \b or \B (\K is not supported by
918 pcre_dfa_exec()). It's an incompatible change, but it makes the two
919 matching functions compatible, and I think it's the right thing to do.
920
921 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
922 so that the demonstration program is easily available in environments where
923 PCRE has not been installed from source.
924
925 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
926 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
927 library.
928
929 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
930 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
931 is not the first non-POSIX option to be added. Clearly some people find
932 these options useful.
933
934 17. If a caller to the POSIX matching function regexec() passes a non-zero
935 value for nmatch with a NULL value for pmatch, the value of
936 nmatch is forced to zero.
937
938 18. RunGrepTest did not have a test for the availability of the -u option of
939 the diff command, as RunTest does. It now checks in the same way as
940 RunTest, and also checks for the -b option.
941
942 19. If an odd number of negated classes containing just a single character
943 interposed, within parentheses, between a forward reference to a named
944 subpattern and the definition of the subpattern, compilation crashed with
945 an internal error, complaining that it could not find the referenced
946 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
947 [The bug was that it was starting one character too far in when skipping
948 over the character class, thus treating the ] as data rather than
949 terminating the class. This meant it could skip too much.]
950
951 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
952 /g option in pcretest when the pattern contains \K, which makes it possible
953 to have an empty string match not at the start, even when the pattern is
954 anchored. Updated pcretest and pcredemo to use this option.
955
956 21. If the maximum number of capturing subpatterns in a recursion was greater
957 than the maximum at the outer level, the higher number was returned, but
958 with unset values at the outer level. The correct (outer level) value is
959 now given.
960
961 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
962 PCRE did not set those parentheses (unlike Perl). I have now found a way to
963 make it do so. The string so far is captured, making this feature
964 compatible with Perl.
965
966 23. The tests have been re-organized, adding tests 11 and 12, to make it
967 possible to check the Perl 5.10 features against Perl 5.10.
968
969 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
970 pattern matches a fixed length string. PCRE did not allow this; now it
971 does. Neither allows recursion.
972
973 25. I finally figured out how to implement a request to provide the minimum
974 length of subject string that was needed in order to match a given pattern.
975 (It was back references and recursion that I had previously got hung up
976 on.) This code has now been added to pcre_study(); it finds a lower bound
977 to the length of subject needed. It is not necessarily the greatest lower
978 bound, but using it to avoid searching strings that are too short does give
979 some useful speed-ups. The value is available to calling programs via
980 pcre_fullinfo().
981
982 26. While implementing 25, I discovered to my embarrassment that pcretest had
983 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
984 study optimizations had never been tested with that matching function.
985 Oops. What is worse, even when it was passed study data, there was a bug in
986 pcre_dfa_exec() that meant it never actually used it. Double oops. There
987 were also very few tests of studied patterns with pcre_dfa_exec().
988
989 27. If (?| is used to create subpatterns with duplicate numbers, they are now
990 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
991 on the other side of the coin, they are no longer allowed to have different
992 names, because these cannot be distinguished in PCRE, and this has caused
993 confusion. (This is a difference from Perl.)
994
995 28. When duplicate subpattern names are present (necessarily with different
996 numbers, as required by 27 above), and a test is made by name in a
997 conditional pattern, either for a subpattern having been matched, or for
998 recursion in such a pattern, all the associated numbered subpatterns are
999 tested, and the overall condition is true if the condition is true for any
1000 one of them. This is the way Perl works, and is also more like the way
1001 testing by number works.
1002
1003
1004 Version 7.9 11-Apr-09
1005 ---------------------
1006
1007 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1008 (pcretest), all targets were linked against these libraries. This included
1009 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1010 libraries. This caused unwanted dependencies to be created. This problem
1011 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1012 pcretest is linked with readline.
1013
1014 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1015 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1016 moved outside it again, because FALSE and TRUE are already defined in AIX,
1017 but BOOL is not.
1018
1019 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1020 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1021
1022 4. The pcregrep documentation said spaces were inserted as well as colons (or
1023 hyphens) following file names and line numbers when outputting matching
1024 lines. This is not true; no spaces are inserted. I have also clarified the
1025 wording for the --colour (or --color) option.
1026
1027 5. In pcregrep, when --colour was used with -o, the list of matching strings
1028 was not coloured; this is different to GNU grep, so I have changed it to be
1029 the same.
1030
1031 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1032 each matching line was coloured. Now it goes on to look for further matches
1033 of any of the test patterns, which is the same behaviour as GNU grep.
1034
1035 7. A pattern that could match an empty string could cause pcregrep to loop; it
1036 doesn't make sense to accept an empty string match in pcregrep, so I have
1037 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1038 seems to be how GNU grep behaves.
1039
1040 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1041 start or after a newline", because the conditional assertion was not being
1042 correctly handled. The rule now is that both the assertion and what follows
1043 in the first alternative must satisfy the test.
1044
1045 9. If auto-callout was enabled in a pattern with a conditional group whose
1046 condition was an assertion, PCRE could crash during matching, both with
1047 pcre_exec() and pcre_dfa_exec().
1048
1049 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1050 used for matching.
1051
1052 11. Unicode property support in character classes was not working for
1053 characters (bytes) greater than 127 when not in UTF-8 mode.
1054
1055 12. Added the -M command line option to pcretest.
1056
1057 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1058
1059 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1060
1061 16. Added comments and documentation about mis-use of no_arg in the C++
1062 wrapper.
1063
1064 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1065 from Martin Jerabek that uses macro names for all relevant character and
1066 string constants.
1067
1068 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1069 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1070 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1071 these, but not everybody uses configure.
1072
1073 19. A conditional group that had only one branch was not being correctly
1074 recognized as an item that could match an empty string. This meant that an
1075 enclosing group might also not be so recognized, causing infinite looping
1076 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1077 with the subject "ab", where knowledge that the repeated group can match
1078 nothing is needed in order to break the loop.
1079
1080 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1081 exec(), but without supplying a callout function, matching went wrong.
1082
1083 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1084 leak if the size of the offset vector was greater than 30. When the vector
1085 is smaller, the saved offsets during recursion go onto a local stack
1086 vector, but for larger vectors malloc() is used. It was failing to free
1087 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1088 error, in fact).
1089
1090 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1091 heapframe that is used only when UTF-8 support is enabled. This caused no
1092 problem, but was untidy.
1093
1094 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1095 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1096 included within another project.
1097
1098 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1099 slightly modified by me:
1100
1101 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1102 not building pcregrep.
1103
1104 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1105 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1106
1107 25. Forward references, both numeric and by name, in patterns that made use of
1108 duplicate group numbers, could behave incorrectly or give incorrect errors,
1109 because when scanning forward to find the reference group, PCRE was not
1110 taking into account the duplicate group numbers. A pattern such as
1111 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1112
1113 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1114 the feature of a custom pointer more persuasive (as requested by a user).
1115
1116 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1117 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1118 pre-defined.
1119
1120 28. Added support for (*UTF8) at the start of a pattern.
1121
1122 29. Arrange for flags added by the "release type" setting in CMake to be shown
1123 in the configuration summary.
1124
1125
1126 Version 7.8 05-Sep-08
1127 ---------------------
1128
1129 1. Replaced UCP searching code with optimized version as implemented for Ad
1130 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1131 stage table and inline lookup instead of a function, giving speed ups of 2
1132 to 5 times on some simple patterns that I tested. Permission was given to
1133 distribute the MultiStage2.py script that generates the tables (it's not in
1134 the tarball, but is in the Subversion repository).
1135
1136 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1137 scripts.
1138
1139 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1140 a group with a zero qualifier. The result of the study could be incorrect,
1141 or the function might crash, depending on the pattern.
1142
1143 4. Caseless matching was not working for non-ASCII characters in back
1144 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1145 It now works when Unicode Property Support is available.
1146
1147 5. In pcretest, an escape such as \x{de} in the data was always generating
1148 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1149 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1150 truncation.
1151
1152 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1153
1154 7. Added two (int) casts to pcregrep when printing the difference of two
1155 pointers, in case they are 64-bit values.
1156
1157 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1158 test 2 if it fails.
1159
1160 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1161 and a #define of that name to empty if it is not externally set. This is to
1162 allow users of MSVC to set it if necessary.
1163
1164 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1165 the convenience functions in the pcre_get.c source file.
1166
1167 11. An option change at the start of a pattern that had top-level alternatives
1168 could cause overwriting and/or a crash. This command provoked a crash in
1169 some environments:
1170
1171 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1172
1173 This potential security problem was recorded as CVE-2008-2371.
1174
1175 12. For a pattern where the match had to start at the beginning or immediately
1176 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1177 pcre_dfa_exec() could read past the end of the passed subject if there was
1178 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1179 pcretest so that it places the subject at the end of its malloc-ed buffer.
1180
1181 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1182 exec() might read past the end of the data buffer in UTF-8 mode.
1183
1184 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1185 the data contained the byte 0x85 as part of a UTF-8 character within its
1186 first line. This applied both to normal and DFA matching.
1187
1188 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1189 /^[^d]*?$/8 failed to match "abc".
1190
1191 16. Added a missing copyright notice to pcrecpp_internal.h.
1192
1193 17. Make it more clear in the documentation that values returned from
1194 pcre_exec() in ovector are byte offsets, not character counts.
1195
1196 18. Tidied a few places to stop certain compilers from issuing warnings.
1197
1198 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1199 supplied by Stefan Weber. I made a further small update for 7.8 because
1200 there is a change of source arrangements: the pcre_searchfuncs.c module is
1201 replaced by pcre_ucd.c.
1202
1203
1204 Version 7.7 07-May-08
1205 ---------------------
1206
1207 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1208 a string to a long long, pretend we don't even have a long long." This is
1209 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1210
1211 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1212 pre-7.6 versions, which defined a global no_arg variable instead of putting
1213 it in the RE class. (See also #8 below.)
1214
1215 3. Remove a line of dead code, identified by coverity and reported by Nuno
1216 Lopes.
1217
1218 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1219
1220 (1) The include/exclude patterns were being applied to the whole pathnames
1221 of files, instead of just to the final components.
1222
1223 (2) If there was more than one level of directory, the subdirectories were
1224 skipped unless they satisfied the include/exclude conditions. This is
1225 inconsistent with GNU grep (and could even be seen as contrary to the
1226 pcregrep specification - which I improved to make it absolutely clear).
1227 The action now is always to scan all levels of directory, and just
1228 apply the include/exclude patterns to regular files.
1229
1230 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1231 --exclude_dir in the tests to avoid scanning .svn directories.
1232
1233 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1234 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1235 doesn't support NULs in patterns.
1236
1237 7. Added some missing "const"s to declarations of static tables in
1238 pcre_compile.c and pcre_dfa_exec.c.
1239
1240 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1241 caused by fix #2 above. (Subsequently also a second patch to fix the
1242 first patch. And a third patch - this was a messy problem.)
1243
1244 9. Applied Craig's patch to remove the use of push_back().
1245
1246 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1247 matching function regexec().
1248
1249 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1250 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1251 references. PCRE supports relative numbers with this syntax (I don't think
1252 Oniguruma does).
1253
1254 12. Previously, a group with a zero repeat such as (...){0} was completely
1255 omitted from the compiled regex. However, this means that if the group
1256 was called as a subroutine from elsewhere in the pattern, things went wrong
1257 (an internal error was given). Such groups are now left in the compiled
1258 pattern, with a new opcode that causes them to be skipped at execution
1259 time.
1260
1261 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1262 to the way PCRE behaves:
1263
1264 (a) A lone ] character is dis-allowed (Perl treats it as data).
1265
1266 (b) A back reference to an unmatched subpattern matches an empty string
1267 (Perl fails the current match path).
1268
1269 (c) A data ] in a character class must be notated as \] because if the
1270 first data character in a class is ], it defines an empty class. (In
1271 Perl it is not possible to have an empty class.) The empty class []
1272 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1273 The negative empty class [^] matches any one character, independently
1274 of the DOTALL setting.
1275
1276 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1277 non-existent subpattern following a character class starting with ']' and
1278 containing () gave an internal compiling error instead of "reference to
1279 non-existent subpattern". Fortunately, when the pattern did exist, the
1280 compiled code was correct. (When scanning forwards to check for the
1281 existencd of the subpattern, it was treating the data ']' as terminating
1282 the class, so got the count wrong. When actually compiling, the reference
1283 was subsequently set up correctly.)
1284
1285 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1286 it was being rejected as not supported by pcre_dfa_exec(), even though
1287 other assertions are supported. I have made pcre_dfa_exec() support
1288 (*FAIL).
1289
1290 16. The implementation of 13c above involved the invention of a new opcode,
1291 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1292 cannot be changed at match time, I realized I could make a small
1293 improvement to matching performance by compiling OP_ALLANY instead of
1294 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1295 on the OP_ANY path.
1296
1297 17. Compiling pcretest on Windows with readline support failed without the
1298 following two fixes: (1) Make the unistd.h include conditional on
1299 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1300
1301 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1302 ncurses library to be included for pcretest when ReadLine support is
1303 requested, but also to allow for it to be overridden. This patch came from
1304 Daniel Bergström.
1305
1306 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1307 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1308 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1309 spotting this.
1310
1311
1312 Version 7.6 28-Jan-08
1313 ---------------------
1314
1315 1. A character class containing a very large number of characters with
1316 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1317 overflow.
1318
1319 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1320 HAVE_LONG_LONG is not defined.
1321
1322 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1323 bring them up to date and include new features. This patch includes:
1324
1325 - Fixed PH's badly added libz and libbz2 support.
1326 - Fixed a problem with static linking.
1327 - Added pcredemo. [But later removed - see 7 below.]
1328 - Fixed dftables problem and added an option.
1329 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1330 HAVE_LONG_LONG.
1331 - Added readline support for pcretest.
1332 - Added an listing of the option settings after cmake has run.
1333
1334 4. A user submitted a patch to Makefile that makes it easy to create
1335 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1336 Makefile.am that cause it to include this special target, without
1337 affecting anything else. Note that the same mingw target plus all
1338 the other distribution libraries and programs are now supported
1339 when configuring with CMake (see 6 below) instead of with
1340 Configure/Make.
1341
1342 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1343 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1344 exported in the Windows port". It has not yet been confirmed that the patch
1345 solves the problem, but it does no harm.
1346
1347 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1348 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1349 with CMake, and also correct the comment about stack recursion.
1350
1351 7. Remove the automatic building of pcredemo from the ./configure system and
1352 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1353 of a program that users should build themselves after PCRE is installed, so
1354 building it automatically is not really right. What is more, it gave
1355 trouble in some build environments.
1356
1357 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1358
1359
1360 Version 7.5 10-Jan-08
1361 ---------------------
1362
1363 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1364 values in parens when parsing an RE using the C++ wrapper."
1365
1366 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1367 Characters greater than 255 were excluded from the class instead of being
1368 included.
1369
1370 3. The same bug as (2) above applied to negated POSIX classes such as
1371 [:^space:].
1372
1373 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1374 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1375 I have changed it.
1376
1377 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1378 first named subpattern) and a construct such as (?&a) would reference the
1379 first named subpattern whose name started with "a" (in other words, the
1380 length check was missing). Both these problems are fixed. "Subpattern name
1381 expected" is now given for (?&) (a zero-length name), and this patch also
1382 makes it give the same error for \k'' (previously it complained that that
1383 was a reference to a non-existent subpattern).
1384
1385 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1386 this is right because (?- can be followed by option settings as well as by
1387 digits. I have, however, made the messages clearer.
1388
1389 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1390 than the number used in the conditional) now cause a compile-time error.
1391 This is actually not compatible with Perl, which accepts such patterns, but
1392 treats the conditional as always being FALSE (as PCRE used to), but it
1393 seems to me that giving a diagnostic is better.
1394
1395 8. Change "alphameric" to the more common word "alphanumeric" in comments
1396 and messages.
1397
1398 9. Fix two occurrences of "backslash" in comments that should have been
1399 "backspace".
1400
1401 10. Remove two redundant lines of code that can never be obeyed (their function
1402 was moved elsewhere).
1403
1404 11. The program that makes PCRE's Unicode character property table had a bug
1405 which caused it to generate incorrect table entries for sequences of
1406 characters that have the same character type, but are in different scripts.
1407 It amalgamated them into a single range, with the script of the first of
1408 them. In other words, some characters were in the wrong script. There were
1409 thirteen such cases, affecting characters in the following ranges:
1410
1411 U+002b0 - U+002c1
1412 U+0060c - U+0060d
1413 U+0061e - U+00612
1414 U+0064b - U+0065e
1415 U+0074d - U+0076d
1416 U+01800 - U+01805
1417 U+01d00 - U+01d77
1418 U+01d9b - U+01dbf
1419 U+0200b - U+0200f
1420 U+030fc - U+030fe
1421 U+03260 - U+0327f
1422 U+0fb46 - U+0fbb1
1423 U+10450 - U+1049d
1424
1425 12. The -o option (show only the matching part of a line) for pcregrep was not
1426 compatible with GNU grep in that, if there was more than one match in a
1427 line, it showed only the first of them. It now behaves in the same way as
1428 GNU grep.
1429
1430 13. If the -o and -v options were combined for pcregrep, it printed a blank
1431 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1432 does the same. The return code can be used to tell if there were any
1433 non-matching lines.
1434
1435 14. Added --file-offsets and --line-offsets to pcregrep.
1436
1437 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1438 infinitely looping recursion. The bug was that positive lookaheads were not
1439 being skipped when checking for a possible empty match (negative lookaheads
1440 and both kinds of lookbehind were skipped).
1441
1442 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1443 inclusion of <windows.h> to before rather than after the definition of
1444 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1445
1446 17. Specifying a possessive quantifier with a specific limit for a Unicode
1447 character property caused pcre_compile() to compile bad code, which led at
1448 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1449 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1450 caused the error; without that there was no problem.
1451
1452 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1453
1454 19. Added --enable-pcretest-libreadline.
1455
1456 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1457 RE::GlobalReplace(). As a result, the number of replacements returned was
1458 double what it should be. I removed one of the increments, but Craig sent a
1459 later patch that removed the other one (the right fix) and added unit tests
1460 that check the return values (which was not done before).
1461
1462 21. Several CMake things:
1463
1464 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1465 the names libpcre and libpcreposix, not just pcre and pcreposix.
1466
1467 (2) The above change means that pcretest and pcregrep are now correctly
1468 linked with the newly-built libraries, not previously installed ones.
1469
1470 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1471
1472 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1473 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1474 UTF-8 newline character). The key issue is that the pattern starts .*;
1475 this means that the match must be either at the beginning, or after a
1476 newline. The bug was in the code for advancing after a failed match and
1477 checking that the new position followed a newline. It was not taking
1478 account of UTF-8 characters correctly.
1479
1480 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1481 character classes. PCRE was not treating the sequence [:...:] as a
1482 character class unless the ... were all letters. Perl, however, seems to
1483 allow any characters between [: and :], though of course it rejects as
1484 unknown any "names" that contain non-letters, because all the known class
1485 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1486 for example, whereas PCRE did not - it did not recognize a POSIX character
1487 class. This seemed a bit dangerous, so the code has been changed to be
1488 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1489 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1490 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1491 Perl does, and where it didn't before.
1492
1493 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1494 Windows environments %n is disabled by default.
1495
1496
1497 Version 7.4 21-Sep-07
1498 ---------------------
1499
1500 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1501 means that a class such as [\s] counted as "explicit reference to CR or
1502 LF". That isn't really right - the whole point of the change was to try to
1503 help when there was an actual mention of one of the two characters. So now
1504 the change happens only if \r or \n (or a literal CR or LF) character is
1505 encountered.
1506
1507 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1508 of both had grown to the point where there were only 3 bits left.
1509 Fortunately, there was spare space in the data structure, and so I have
1510 moved the internal flags into a new 16-bit field to free up more option
1511 bits.
1512
1513 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1514 but did not set the internal JCHANGED flag - either of these is enough to
1515 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1516 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1517 start sets both bits.
1518
1519 4. Added options (at build time, compile time, exec time) to change \R from
1520 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1521
1522 5. doc/pcresyntax.html was missing from the distribution.
1523
1524 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1525 compatibility, even though it is no longer used.
1526
1527 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1528 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1529 windows.h file is present (where different names are used). [This was
1530 reversed later after testing - see 16 below.]
1531
1532 8. Changed all #include <config.h> to #include "config.h". There were also
1533 some further <pcre.h> cases that I changed to "pcre.h".
1534
1535 9. When pcregrep was used with the --colour option, it missed the line ending
1536 sequence off the lines that it output.
1537
1538 10. It was pointed out to me that arrays of string pointers cause lots of
1539 relocations when a shared library is dynamically loaded. A technique of
1540 using a single long string with a table of offsets can drastically reduce
1541 these. I have refactored PCRE in four places to do this. The result is
1542 dramatic:
1543
1544 Originally: 290
1545 After changing UCP table: 187
1546 After changing error message table: 43
1547 After changing table of "verbs" 36
1548 After changing table of Posix names 22
1549
1550 Thanks to the folks working on Gregex for glib for this insight.
1551
1552 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1553 unicode-properties was also set.
1554
1555 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1556
1557 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1558 checked only for CRLF.
1559
1560 14. Added casts to pcretest.c to avoid compiler warnings.
1561
1562 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1563
1564 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1565 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1566 entirely. This removes changes made in 7 above.
1567
1568 17. The CMake files have been updated, and there is now more information about
1569 building with CMake in the NON-UNIX-USE document.
1570
1571
1572 Version 7.3 28-Aug-07
1573 ---------------------
1574
1575 1. In the rejigging of the build system that eventually resulted in 7.1, the
1576 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1577 brackets there is not right, since it causes compilers to look for an
1578 installed pcre.h, not the version that is in the source that is being
1579 compiled (which of course may be different). I have changed it back to:
1580
1581 #include "pcre.h"
1582
1583 I have a vague recollection that the change was concerned with compiling in
1584 different directories, but in the new build system, that is taken care of
1585 by the VPATH setting the Makefile.
1586
1587 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1588 when the subject happened to end in the byte 0x85 (e.g. if the last
1589 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1590 characters but of course it shouldn't be taken as a newline when it is part
1591 of another character. The bug was that, for an unlimited repeat of . in
1592 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1593 characters when looking for a newline.
1594
1595 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1596
1597 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1598 in debug output.
1599
1600 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1601 long printing in the pcrecpp unittest when running under MinGW.
1602
1603 6. ESC_K was left out of the EBCDIC table.
1604
1605 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1606 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1607 limit also applies to "virtual nesting" when a pattern is recursive, and in
1608 this case 1000 isn't so big. I have been able to remove this limit at the
1609 expense of backing off one optimization in certain circumstances. Normally,
1610 when pcre_exec() would call its internal match() function recursively and
1611 immediately return the result unconditionally, it uses a "tail recursion"
1612 feature to save stack. However, when a subpattern that can match an empty
1613 string has an unlimited repetition quantifier, it no longer makes this
1614 optimization. That gives it a stack frame in which to save the data for
1615 checking that an empty string has been matched. Previously this was taken
1616 from the 1000-entry workspace that had been reserved. So now there is no
1617 explicit limit, but more stack is used.
1618
1619 8. Applied Daniel's patches to solve problems with the import/export magic
1620 syntax that is required for Windows, and which was going wrong for the
1621 pcreposix and pcrecpp parts of the library. These were overlooked when this
1622 problem was solved for the main library.
1623
1624 9. There were some crude static tests to avoid integer overflow when computing
1625 the size of patterns that contain repeated groups with explicit upper
1626 limits. As the maximum quantifier is 65535, the maximum group length was
1627 set at 30,000 so that the product of these two numbers did not overflow a
1628 32-bit integer. However, it turns out that people want to use groups that
1629 are longer than 30,000 bytes (though not repeat them that many times).
1630 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1631 made it possible to implement the integer overflow checks in a much more
1632 dynamic way, which I have now done. The artificial limitation on group
1633 length has been removed - we now have only the limit on the total length of
1634 the compiled pattern, which depends on the LINK_SIZE setting.
1635
1636 10. Fixed a bug in the documentation for get/copy named substring when
1637 duplicate names are permitted. If none of the named substrings are set, the
1638 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1639 empty string.
1640
1641 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1642 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1643 because the ] is interpreted as the first data character and the
1644 terminating ] is not found. PCRE has been made compatible with Perl in this
1645 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1646 cause memory overwriting.
1647
1648 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1649 string has been matched (to stop an infinite loop). It was not recognizing
1650 a conditional subpattern that could match an empty string if that
1651 subpattern was within another subpattern. For example, it looped when
1652 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1653 condition was not nested. This bug has been fixed.
1654
1655 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1656 past the start of the subject in the presence of bytes with the top bit
1657 set, for example "\x8aBCD".
1658
1659 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1660 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1661
1662 14. Optimized (?!) to (*FAIL).
1663
1664 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1665 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1666 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1667 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1668 does: it's just the validity check that is more restrictive.
1669
1670 16. Inserted checks for integer overflows during escape sequence (backslash)
1671 processing, and also fixed erroneous offset values for syntax errors during
1672 backslash processing.
1673
1674 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1675 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1676
1677 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1678 caused an overrun.
1679
1680 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1681 something other than just ASCII characters) inside a group that had an
1682 unlimited repeat caused a loop at compile time (while checking to see
1683 whether the group could match an empty string).
1684
1685 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1686 [\P{Any}] did so. (Error in the code for printing property names.)
1687
1688 21. An orphan \E inside a character class could cause a crash.
1689
1690 22. A repeated capturing bracket such as (A)? could cause a wild memory
1691 reference during compilation.
1692
1693 23. There are several functions in pcre_compile() that scan along a compiled
1694 expression for various reasons (e.g. to see if it's fixed length for look
1695 behind). There were bugs in these functions when a repeated \p or \P was
1696 present in the pattern. These operators have additional parameters compared
1697 with \d, etc, and these were not being taken into account when moving along
1698 the compiled data. Specifically:
1699
1700 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1701 length.
1702
1703 (b) An item such as \pL+ within a repeated group could cause crashes or
1704 loops.
1705
1706 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1707 "reference to non-existent subpattern" error.
1708
1709 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1710
1711 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1712 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1713
1714 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1715
1716 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1717 character were causing crashes (broken optimization).
1718
1719 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1720 \p or \P) caused a compile-time loop.
1721
1722 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1723 break. For example, the unstudied pattern [\r\n]A does not match the string
1724 "\r\nA" because change 7.0/46 below moves the current point on by two
1725 characters after failing to match at the start. However, the pattern \nA
1726 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1727 the same is true. There doesn't seem any very clean way out of this, but
1728 what I have chosen to do makes the common cases work: PCRE now takes note
1729 of whether there can be an explicit match for \r or \n anywhere in the
1730 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1731 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1732 pattern has explicit CR or LF references.
1733
1734 29. Added (*CR) etc for changing newline setting at start of pattern.
1735
1736
1737 Version 7.2 19-Jun-07
1738 ---------------------
1739
1740 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1741 which is apparently normally available under Windows.
1742
1743 2. Re-jig the pcregrep tests with different newline settings in an attempt
1744 to make them independent of the local environment's newline setting.
1745
1746 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1747
1748 4. Some of the "internals" tests were previously cut out when the link size
1749 was not 2, because the output contained actual offsets. The recent new
1750 "Z" feature of pcretest means that these can be cut out, making the tests
1751 usable with all link sizes.
1752
1753 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1754 stack recursion. This gives a massive performance boost under BSD, but just
1755 a small improvement under Linux. However, it saves one field in the frame
1756 in all cases.
1757
1758 6. Added more features from the forthcoming Perl 5.10:
1759
1760 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1761 recursion call. It refers to the nth most recently opened parentheses.
1762
1763 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1764 to be opened parentheses.
1765
1766 (c) Conditions that refer to capturing parentheses can be specified
1767 relatively, for example, (?(-2)... or (?(+3)...
1768
1769 (d) \K resets the start of the current match so that everything before
1770 is not part of it.
1771
1772 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1773
1774 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1775 reference syntax.
1776
1777 (g) (?| introduces a group in which the numbering of parentheses in each
1778 alternative starts with the same number.
1779
1780 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1781
1782 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1783 PCRE_INFO_JCHANGED.
1784
1785 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1786 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1787 for detecting groups that can match an empty string.
1788
1789 9. A pattern with a very large number of alternatives (more than several
1790 hundred) was running out of internal workspace during the pre-compile
1791 phase, where pcre_compile() figures out how much memory will be needed. A
1792 bit of new cunning has reduced the workspace needed for groups with
1793 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1794 workspace instead of running out of the 4096 that are available.
1795
1796 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1797
1798 11. Applied patch from Google to remove an optimization that didn't quite work.
1799 The report of the bug said:
1800
1801 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1802 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1803 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1804
1805 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1806 it matched the wrong number of bytes.
1807
1808
1809 Version 7.1 24-Apr-07
1810 ---------------------
1811
1812 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1813 that is more "standard", making use of automake and other Autotools. There
1814 is some re-arrangement of the files and adjustment of comments consequent
1815 on this.
1816
1817 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1818 for recursive directory scanning broke on some systems because the files
1819 are not scanned in any specific order and on different systems the order
1820 was different. A call to "sort" has been inserted into RunGrepTest for the
1821 approprate test as a short-term fix. In the longer term there may be an
1822 alternative.
1823
1824 3. I had an email from Eric Raymond about problems translating some of PCRE's
1825 man pages to HTML (despite the fact that I distribute HTML pages, some
1826 people do their own conversions for various reasons). The problems
1827 concerned the use of low-level troff macros .br and .in. I have therefore
1828 removed all such uses from the man pages (some were redundant, some could
1829 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1830 HTML has been updated to handle .nf/.fi and to complain if it encounters
1831 .br or .in.
1832
1833 4. Updated comments in configure.ac that get placed in config.h.in and also
1834 arranged for config.h to be included in the distribution, with the name
1835 config.h.generic, for the benefit of those who have to compile without
1836 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1837
1838 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1839 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1840 makevp.bat for the current PCRE, using the additional files
1841 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1842
1843 6. A Windows user reported a minor discrepancy with test 2, which turned out
1844 to be caused by a trailing space on an input line that had got lost in his
1845 copy. The trailing space was an accident, so I've just removed it.
1846
1847 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1848 that is needed.
1849
1850 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1851 as "const" (a) because they are and (b) because it helps the PHP
1852 maintainers who have recently made a script to detect big data structures
1853 in the php code that should be moved to the .rodata section. I remembered
1854 to update Builducptable as well, so it won't revert if ucptable.h is ever
1855 re-created.
1856
1857 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1858 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1859 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1860 support is not required. This saves 1.5-2K of code, which is important in
1861 some applications.
1862
1863 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1864 so as not to refer to the tables, even though these functions will never be
1865 called when UTF-8 support is disabled. Otherwise there are problems with a
1866 shared library.
1867
1868 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1869
1870 (a) It was defining its arguments as char * instead of void *.
1871
1872 (b) It was assuming that all moves were upwards in memory; this was true
1873 a long time ago when I wrote it, but is no longer the case.
1874
1875 The emulated memove() is provided for those environments that have neither
1876 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1877 is clearly not the case, as these two bugs were recently reported.
1878
1879 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1880 and Detrail to create the HTML documentation, the .txt form of the man
1881 pages, and it removes trailing spaces from listed files. It also creates
1882 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1883 case, it wraps all the #defines with #ifndefs. This script should be run
1884 before "make dist".
1885
1886 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1887 with Unicode property support.
1888
1889 (a) For a maximizing quantifier, if the two different cases of the
1890 character were of different lengths in their UTF-8 codings (there are
1891 some cases like this - I found 11), and the matching function had to
1892 back up over a mixture of the two cases, it incorrectly assumed they
1893 were both the same length.
1894
1895 (b) When PCRE was configured to use the heap rather than the stack for
1896 recursion during matching, it was not correctly preserving the data for
1897 the other case of a UTF-8 character when checking ahead for a match
1898 while processing a minimizing repeat. If the check also involved
1899 matching a wide character, but failed, corruption could cause an
1900 erroneous result when trying to check for a repeat of the original
1901 character.
1902
1903 13. Some tidying changes to the testing mechanism:
1904
1905 (a) The RunTest script now detects the internal link size and whether there
1906 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1907 values substituted by "configure". (The RunGrepTest script already did
1908 this for UTF-8.) The configure.ac script no longer substitutes the
1909 relevant variables.
1910
1911 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1912 with length and offset values. This means that the output is different
1913 for different internal link sizes. Test 2 is skipped for link sizes
1914 other than 2 because of this, bypassing the problem. Unfortunately,
1915 there was also a test in test 3 (the locale tests) that used /B and
1916 failed for link sizes other than 2. Rather than cut the whole test out,
1917 I have added a new /Z option to pcretest that replaces the length and
1918 offset values with spaces. This is now used to make test 3 independent
1919 of link size. (Test 2 will be tidied up later.)
1920
1921 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1922 segmentation fault instead of returning the appropriate error message.
1923
1924 15. In multiline mode when the newline sequence was set to "any", the pattern
1925 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1926 This doesn't seem right; it now treats the CRLF combination as the line
1927 ending, and so does not match in that case. It's only a pattern such as ^$
1928 that would hit this one: something like ^ABC$ would have failed after \r
1929 and then tried again after \r\n.
1930
1931 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1932 in an attempt to make files that differ only in their line terminators
1933 compare equal. This works on Linux.
1934
1935 17. Under certain error circumstances pcregrep might try to free random memory
1936 as it exited. This is now fixed, thanks to valgrind.
1937
1938 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1939 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1940 was because its rules for how to advance for /g after matching an empty
1941 string at the end of a line did not allow for this case. They now check for
1942 it specially.
1943
1944 20. pcretest is supposed to handle patterns and data of any length, by
1945 extending its buffers when necessary. It was getting this wrong when the
1946 buffer for a data line had to be extended.
1947
1948 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1949 CRLF as a newline sequence.
1950
1951 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1952 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1953 I have nevertheless tidied it up.
1954
1955 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1956
1957 24. Added a man page for pcre-config.
1958
1959
1960 Version 7.0 19-Dec-06
1961 ---------------------
1962
1963 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1964 moving to gcc 4.1.1.
1965
1966 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1967 sys/time.h, which is documented as needed for this function. It doesn't
1968 seem to matter on Linux, but it showed up on some releases of OS X.
1969
1970 3. It seems that there are systems where bytes whose values are greater than
1971 127 match isprint() in the "C" locale. The "C" locale should be the
1972 default when a C program starts up. In most systems, only ASCII printing
1973 characters match isprint(). This difference caused the output from pcretest
1974 to vary, making some of the tests fail. I have changed pcretest so that:
1975
1976 (a) When it is outputting text in the compiled version of a pattern, bytes
1977 other than 32-126 are always shown as hex escapes.
1978
1979 (b) When it is outputting text that is a matched part of a subject string,
1980 it does the same, unless a different locale has been set for the match
1981 (using the /L modifier). In this case, it uses isprint() to decide.
1982
1983 4. Fixed a major bug that caused incorrect computation of the amount of memory
1984 required for a compiled pattern when options that changed within the
1985 pattern affected the logic of the preliminary scan that determines the
1986 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1987 that the computed length was too small. The symptoms of this bug were
1988 either the PCRE error "internal error: code overflow" from pcre_compile(),
1989 or a glibc crash with a message such as "pcretest: free(): invalid next
1990 size (fast)". Examples of patterns that provoked this bug (shown in
1991 pcretest format) are:
1992
1993 /(?-x: )/x
1994 /(?x)(?-x: \s*#\s*)/
1995 /((?i)[\x{c0}])/8
1996 /(?i:[\x{c0}])/8
1997
1998 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1999 is now done differently.
2000
2001 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2002 wrapper classes; (b) implement a new function in the C++ scanner that is
2003 more efficient than the old way of doing things because it avoids levels of
2004 recursion in the regex matching; (c) add a paragraph to the documentation
2005 for the FullMatch() function.
2006
2007 6. The escape sequence \n was being treated as whatever was defined as
2008 "newline". Not only was this contrary to the documentation, which states
2009 that \n is character 10 (hex 0A), but it also went horribly wrong when
2010 "newline" was defined as CRLF. This has been fixed.
2011
2012 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2013 was being set to -1 for the "end of line" case (supposedly a value that no
2014 character can have). Though this value is never used (the check for end of
2015 line is "zero bytes in current character"), it caused compiler complaints.
2016 I've changed it to 0xffffffff.
2017
2018 8. In pcre_version.c, the version string was being built by a sequence of
2019 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2020 string (as it is for production releases) called a macro with an empty
2021 argument. The C standard says the result of this is undefined. The gcc
2022 compiler treats it as an empty string (which was what was wanted) but it is
2023 reported that Visual C gives an error. The source has been hacked around to
2024 avoid this problem.
2025
2026 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2027 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2028 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2029 of them did).
2030
2031 10. Originally, pcretest opened its input and output without "b"; then I was
2032 told that "b" was needed in some environments, so it was added for release
2033 5.0 to both the input and output. (It makes no difference on Unix-like
2034 systems.) Later I was told that it is wrong for the input on Windows. I've
2035 now abstracted the modes into two macros, to make it easier to fiddle with
2036 them, and removed "b" from the input mode under Windows.
2037
2038 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2039
2040 12. Added -help and --help to pcretest as an official way of being reminded
2041 of the options.
2042
2043 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2044 and pcrecpp.cc because they annoy compilers at high warning levels.
2045
2046 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2047
2048 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2049 scripts are not C programs :-) and which was not noticed because it works
2050 on Linux.
2051
2052 16. pcretest is supposed to handle any length of pattern and data line (as one
2053 line or as a continued sequence of lines) by extending its input buffer if
2054 necessary. This feature was broken for very long pattern lines, leading to
2055 a string of junk being passed to pcre_compile() if the pattern was longer
2056 than about 50K.
2057
2058 17. I have done a major re-factoring of the way pcre_compile() computes the
2059 amount of memory needed for a compiled pattern. Previously, there was code
2060 that made a preliminary scan of the pattern in order to do this. That was
2061 OK when PCRE was new, but as the facilities have expanded, it has become
2062 harder and harder to keep it in step with the real compile phase, and there
2063 have been a number of bugs (see for example, 4 above). I have now found a
2064 cunning way of running the real compile function in a "fake" mode that
2065 enables it to compute how much memory it would need, while actually only
2066 ever using a few hundred bytes of working memory and without too many
2067 tests of the mode. This should make future maintenance and development
2068 easier. A side effect of this work is that the limit of 200 on the nesting
2069 depth of parentheses has been removed (though this was never a serious
2070 limitation, I suspect). However, there is a downside: pcre_compile() now
2071 runs more slowly than before (30% or more, depending on the pattern). I
2072 hope this isn't a big issue. There is no effect on runtime performance.
2073
2074 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2075 newline (only possible for the last line of a file) and it was a
2076 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2077
2078 19. Added additional timing features to pcretest. (1) The -tm option now times
2079 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2080 separate command line item, by a number that specifies the number of
2081 repeats to use when timing. The default is 50000; this gives better
2082 precision, but takes uncomfortably long for very large patterns.
2083
2084 20. Extended pcre_study() to be more clever in cases where a branch of a
2085 subpattern has no definite first character. For example, (a*|b*)[cd] would
2086 previously give no result from pcre_study(). Now it recognizes that the
2087 first character must be a, b, c, or d.
2088
2089 21. There was an incorrect error "recursive call could loop indefinitely" if
2090 a subpattern (or the entire pattern) that was being tested for matching an
2091 empty string contained only one non-empty item after a nested subpattern.
2092 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2093 incorrectly, because the \d was being skipped in the check.
2094
2095 22. The pcretest program now has a new pattern option /B and a command line
2096 option -b, which is equivalent to adding /B to every pattern. This causes
2097 it to show the compiled bytecode, without the additional information that
2098 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2099 is the same as /B/I).
2100
2101 23. A new optimization is now able automatically to treat some sequences such
2102 as a*b as a*+b. More specifically, if something simple (such as a character
2103 or a simple class like \d) has an unlimited quantifier, and is followed by
2104 something that cannot possibly match the quantified thing, the quantifier
2105 is automatically "possessified".
2106
2107 24. A recursive reference to a subpattern whose number was greater than 39
2108 went wrong under certain circumstances in UTF-8 mode. This bug could also
2109 have affected the operation of pcre_study().
2110
2111 25. Realized that a little bit of performance could be had by replacing
2112 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2113
2114 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2115
2116 27. Possessive quantifiers such as a++ were previously implemented by turning
2117 them into atomic groups such as ($>a+). Now they have their own opcodes,
2118 which improves performance. This includes the automatically created ones
2119 from 23 above.
2120
2121 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2122 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2123 the first matched character to be a colon. This applied both to named and
2124 numbered groups.
2125
2126 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2127
2128 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2129 building PCRE on Windows easier, so I have included it in the distribution.
2130
2131 31. There is now a check in pcretest against a ridiculously large number being
2132 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2133 loop, the loop is abandoned.
2134
2135 32. Forward references to subpatterns in conditions such as (?(2)...) where
2136 subpattern 2 is defined later cause pcre_compile() to search forwards in
2137 the pattern for the relevant set of parentheses. This search went wrong
2138 when there were unescaped parentheses in a character class, parentheses
2139 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2140
2141 33. "Subroutine" calls and backreferences were previously restricted to
2142 referencing subpatterns earlier in the regex. This restriction has now
2143 been removed.
2144
2145 34. Added a number of extra features that are going to be in Perl 5.10. On the
2146 whole, these are just syntactic alternatives for features that PCRE had
2147 previously implemented using the Python syntax or my own invention. The
2148 other formats are all retained for compatibility.
2149
2150 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2151 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2152 also .NET compatible.
2153
2154 (b) A recursion or subroutine call to a named group can now be defined as
2155 (?&name) as well as (?P>name).
2156
2157 (c) A backreference to a named group can now be defined as \k<name> or
2158 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2159 5.10, are also .NET compatible.
2160
2161 (d) A conditional reference to a named group can now use the syntax
2162 (?(<name>) or (?('name') as well as (?(name).
2163
2164 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2165 groups (named and numbered) that are never evaluated inline, but can be
2166 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2167 is always false. There may be only one alternative in such a group.
2168
2169 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2170 as the simple (?(R). The condition is true only if the most recent
2171 recursion is that of the given number or name. It does not search out
2172 through the entire recursion stack.
2173
2174 (g) The escape \gN or \g{N} has been added, where N is a positive or
2175 negative number, specifying an absolute or relative reference.
2176
2177 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2178 some "unreachable code" warnings.
2179
2180 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2181 things, this adds five new scripts.
2182
2183 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2184 There were also incompatibilities regarding the handling of \Q..\E inside
2185 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2186 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2187
2188 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2189 matches an empty string, and forcibly breaks the loop. There were bugs in
2190 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2191 against aaaa the result was just "a" rather than "aaaa", for example. Two
2192 separate and independent bugs (that affected different cases) have been
2193 fixed.
2194
2195 39. Refactored the code to abolish the use of different opcodes for small
2196 capturing bracket numbers. This is a tidy that I avoided doing when I
2197 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2198 The new approach is not only tidier, it makes it possible to reduce the
2199 memory needed to fix the previous bug (38).
2200
2201 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2202 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2203 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2204 mode.
2205
2206 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2207 report.
2208
2209 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2210 copy construction and assignment in the C++ wrapper.
2211
2212 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2213 couple of bugs that could have given wrong results in the "--newline=crlf"
2214 case.
2215
2216 44. Added a number of casts and did some reorganization of signed/unsigned int
2217 variables following suggestions from Dair Grant. Also renamed the variable
2218 "this" as "item" because it is a C++ keyword.
2219
2220 45. Arranged for dftables to add
2221
2222 #include "pcre_internal.h"
2223
2224 to pcre_chartables.c because without it, gcc 4.x may remove the array
2225 definition from the final binary if PCRE is built into a static library and
2226 dead code stripping is activated.
2227
2228 46. For an unanchored pattern, if a match attempt fails at the start of a
2229 newline sequence, and the newline setting is CRLF or ANY, and the next two
2230 characters are CRLF, advance by two characters instead of one.
2231
2232
2233 Version 6.7 04-Jul-06
2234 ---------------------
2235
2236 1. In order to handle tests when input lines are enormously long, pcretest has
2237 been re-factored so that it automatically extends its buffers when
2238 necessary. The code is crude, but this _is_ just a test program. The
2239 default size has been increased from 32K to 50K.
2240
2241 2. The code in pcre_study() was using the value of the re argument before
2242 testing it for NULL. (Of course, in any sensible call of the function, it
2243 won't be NULL.)
2244
2245 3. The memmove() emulation function in pcre_internal.h, which is used on
2246 systems that lack both memmove() and bcopy() - that is, hardly ever -
2247 was missing a "static" storage class specifier.
2248
2249 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2250 containing an extended class (one that cannot be represented by a bitmap
2251 because it contains high-valued characters or Unicode property items, e.g.
2252 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2253 pattern, but PCRE should not loop if you do not (it no longer does).
2254 [Detail: two cases were found: (a) a repeated subpattern containing an
2255 extended class; (b) a recursive reference to a subpattern that followed a
2256 previous extended class. It wasn't skipping over the extended class
2257 correctly when UTF-8 mode was not set.]
2258
2259 5. A negated single-character class was not being recognized as fixed-length
2260 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2261 compile error "lookbehind assertion is not fixed length".
2262
2263 6. The RunPerlTest auxiliary script was showing an unexpected difference
2264 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2265 write a Perl script that can interpret lines of an input file either as
2266 byte characters or as UTF-8, which is what "perltest" was being required to
2267 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2268 can't do is switch easily at run time between having the "use utf8;" pragma
2269 or not. In the end, I fudged it by using the RunPerlTest script to insert
2270 "use utf8;" explicitly for the UTF-8 tests.
2271
2272 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2273 the end of the subject string, contrary to the documentation and to what
2274 Perl does. This was true of both matching functions. Now it matches only at
2275 the start of the subject and immediately after *internal* newlines.
2276
2277 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2278 a pointer to an int instead of a pointer to an unsigned long int. This
2279 caused problems on 64-bit systems.
2280
2281 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2282 instance of the 'standard' template library not being so standard".
2283
2284 10. There was no check on the number of named subpatterns nor the maximum
2285 length of a subpattern name. The product of these values is used to compute
2286 the size of the memory block for a compiled pattern. By supplying a very
2287 long subpattern name and a large number of named subpatterns, the size
2288 computation could be caused to overflow. This is now prevented by limiting
2289 the length of names to 32 characters, and the number of named subpatterns
2290 to 10,000.
2291
2292 11. Subpatterns that are repeated with specific counts have to be replicated in
2293 the compiled pattern. The size of memory for this was computed from the
2294 length of the subpattern and the repeat count. The latter is limited to
2295 65535, but there was no limit on the former, meaning that integer overflow
2296 could in principle occur. The compiled length of a repeated subpattern is
2297 now limited to 30,000 bytes in order to prevent this.
2298
2299 12. Added the optional facility to have named substrings with the same name.
2300
2301 13. Added the ability to use a named substring as a condition, using the
2302 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2303 are numbers (not recommended). Forward references are permitted.
2304
2305 14. Added forward references in named backreferences (if you see what I mean).
2306
2307 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2308 pattern could run off the end of the subject. For example, the pattern
2309 "(?s)(.{1,5})"8 did this with the subject "ab".
2310
2311 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2312 PCRE_CASELESS was set when matching characters that were quantified with ?
2313 or *.
2314
2315 17. A character class other than a single negated character that had a minimum
2316 but no maximum quantifier - for example [ab]{6,} - was not handled
2317 correctly by pce_dfa_exec(). It would match only one character.
2318
2319 18. A valid (though odd) pattern that looked like a POSIX character
2320 class but used an invalid character after [ (for example [[,abc,]]) caused
2321 pcre_compile() to give the error "Failed: internal error: code overflow" or
2322 in some cases to crash with a glibc free() error. This could even happen if
2323 the pattern terminated after [[ but there just happened to be a sequence of
2324 letters, a binary zero, and a closing ] in the memory that followed.
2325
2326 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2327 over the years. Originally (before any Unicode support), just the bottom 8
2328 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2329 output from "man perlunicode" includes this:
2330
2331 The regular expression compiler produces polymorphic opcodes. That
2332 is, the pattern adapts to the data and automatically switches to
2333 the Unicode character scheme when presented with Unicode data--or
2334 instead uses a traditional byte scheme when presented with byte
2335 data.
2336
2337 Sadly, a wide octal escape does not cause a switch, and in a string with
2338 no other multibyte characters, these octal escapes are treated as before.
2339 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2340 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2341 Unicode string.
2342
2343 I have not perpetrated such confusion in PCRE. Up till now, it took just
2344 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2345 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2346 translate to the appropriate multibyte character.
2347
2348 29. Applied some refactoring to reduce the number of warnings from Microsoft
2349 and Borland compilers. This has included removing the fudge introduced
2350 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2351 a warning about an unused variable.
2352
2353 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2354 characters since release 4.0, because Perl (from release 5.004) does not.
2355 [Or at least, is documented not to: some releases seem to be in conflict
2356 with the documentation.] However, when a pattern was studied with
2357 pcre_study() and all its branches started with \s, PCRE still included VT
2358 as a possible starting character. Of course, this did no harm; it just
2359 caused an unnecessary match attempt.
2360
2361 22. Removed a now-redundant internal flag bit that recorded the fact that case
2362 dependency changed within the pattern. This was once needed for "required
2363 byte" processing, but is no longer used. This recovers a now-scarce options
2364 bit. Also moved the least significant internal flag bit to the most-
2365 significant bit of the word, which was not previously used (hangover from
2366 the days when it was an int rather than a uint) to free up another bit for
2367 the future.
2368
2369 23. Added support for CRLF line endings as well as CR and LF. As well as the
2370 default being selectable at build time, it can now be changed at runtime
2371 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2372 specify that it is scanning data with non-default line endings.
2373
2374 24. Changed the definition of CXXLINK to make it agree with the definition of
2375 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2376
2377 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2378 recursions. This makes a big different to stack usage for some patterns.
2379
2380 26. If a subpattern containing a named recursion or subroutine reference such
2381 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2382 the space required for the compiled pattern went wrong and gave too small a
2383 value. Depending on the environment, this could lead to "Failed: internal
2384 error: code overflow at offset 49" or "glibc detected double free or
2385 corruption" errors.
2386
2387 27. Applied patches from Google (a) to support the new newline modes and (b) to
2388 advance over multibyte UTF-8 characters in GlobalReplace.
2389
2390 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2391 difference for some implementation of PCRE in some Windows version.
2392
2393 29. Added some extra testing facilities to pcretest:
2394
2395 \q<number> in a data line sets the "match limit" value
2396 \Q<number> in a data line sets the "match recursion limt" value
2397 -S <number> sets the stack size, where <number> is in megabytes
2398
2399 The -S option isn't available for Windows.
2400
2401
2402 Version 6.6 06-Feb-06
2403 ---------------------
2404
2405 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2406 in pcreposix.h. I have copied the definition from pcre.h.
2407
2408 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2409 because pcre.h is no longer a built file.
2410
2411 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2412 not normally included in the compiled code.
2413
2414
2415 Version 6.5 01-Feb-06
2416 ---------------------
2417
2418 1. When using the partial match feature with pcre_dfa_exec(), it was not
2419 anchoring the second and subsequent partial matches at the new starting
2420 point. This could lead to incorrect results. For example, with the pattern
2421 /1234/, partially matching against "123" and then "a4" gave a match.
2422
2423 2. Changes to pcregrep:
2424
2425 (a) All non-match returns from pcre_exec() were being treated as failures
2426 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2427 error message is output. Some extra information is given for the
2428 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2429 probably the only errors that are likely to be caused by users (by
2430 specifying a regex that has nested indefinite repeats, for instance).
2431 If there are more than 20 of these errors, pcregrep is abandoned.
2432
2433 (b) A binary zero was treated as data while matching, but terminated the
2434 output line if it was written out. This has been fixed: binary zeroes
2435 are now no different to any other data bytes.
2436
2437 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2438 used to set a locale for matching. The --locale=xxxx long option has
2439 been added (no short equivalent) to specify a locale explicitly on the
2440 pcregrep command, overriding the environment variables.
2441
2442 (d) When -B was used with -n, some line numbers in the output were one less
2443 than they should have been.
2444
2445 (e) Added the -o (--only-matching) option.
2446
2447 (f) If -A or -C was used with -c (count only), some lines of context were
2448 accidentally printed for the final match.
2449
2450 (g) Added the -H (--with-filename) option.
2451
2452 (h) The combination of options -rh failed to suppress file names for files
2453 that were found from directory arguments.
2454
2455 (i) Added the -D (--devices) and -d (--directories) options.
2456
2457 (j) Added the -F (--fixed-strings) option.
2458
2459 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2460
2461 (l) Added the --colo(u)r option.
2462
2463 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2464 is not present by default.
2465
2466 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2467 items such as (?R) or (?1), when the recursion could match a number of
2468 alternatives. If it matched one of the alternatives, but subsequently,
2469 outside the recursion, there was a failure, the code tried to back up into
2470 the recursion. However, because of the way PCRE is implemented, this is not
2471 possible, and the result was an incorrect result from the match.
2472
2473 In order to prevent this happening, the specification of recursion has
2474 been changed so that all such subpatterns are automatically treated as
2475 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2476
2477 4. I had overlooked the fact that, in some locales, there are characters for
2478 which isalpha() is true but neither isupper() nor islower() are true. In
2479 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2480 and ordfeminine) are like this. This affected the treatment of \w and \W
2481 when they appeared in character classes, but not when they appeared outside
2482 a character class. The bit map for "word" characters is now created
2483 separately from the results of isalnum() instead of just taking it from the
2484 upper, lower, and digit maps. (Plus the underscore character, of course.)
2485
2486 5. The above bug also affected the handling of POSIX character classes such as
2487 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2488 permanent tables. Instead, the bit maps for such a class were previously
2489 created as the appropriate unions of the upper, lower, and digit bitmaps.
2490 Now they are created by subtraction from the [[:word:]] class, which has
2491 its own bitmap.
2492
2493 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2494 It is created by subtracting the vertical space characters (\x09, \x0a,
2495 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2496 subtraction was done in the overall bitmap for a character class, meaning
2497 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2498 be recognized. This bug has been fixed.
2499
2500 7. Patches from the folks at Google:
2501
2502 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2503 real life, but is still worth protecting against".
2504
2505 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2506 regular expressions".
2507
2508 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2509 have it.
2510
2511 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2512 "configure" and the latter not, in order to fix a problem somebody had
2513 with compiling the Arg class on HP-UX.
2514
2515 (e) Improve the error-handling of the C++ wrapper a little bit.
2516
2517 (f) New tests for checking recursion limiting.
2518
2519 8. The pcre_memmove() function, which is used only if the environment does not
2520 have a standard memmove() function (and is therefore rarely compiled),
2521 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2522 returning a result (though PCRE never actually uses the result).
2523
2524 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2525 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2526 returned instead of calling malloc() with an overflowing number that would
2527 most likely cause subsequent chaos.
2528
2529 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2530
2531 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2532 with this option is matched, the nmatch and pmatch options of regexec() are
2533 ignored.
2534
2535 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2536 provided in case anyone wants to the the POSIX interface with UTF-8
2537 strings.
2538
2539 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2540 C++ linking (needed for some HP-UX environments).
2541
2542 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2543 (unused parameter) and in the pcre_printint() function (omitted "default"
2544 switch label when the default is to do nothing).
2545
2546 15. Added some code to make it possible, when PCRE is compiled as a C++
2547 library, to replace subject pointers for pcre_exec() with a smart pointer
2548 class, thus making it possible to process discontinuous strings.
2549
2550 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2551 much the same function. They were added by different people who were trying
2552 to make PCRE easy to compile on non-Unix systems. It has been suggested
2553 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2554 for compiling on Windows systems. I have therefore replaced it with
2555 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2556 defaults to "extern" for C or "extern C" for C++, which works fine on
2557 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2558 SCOPE with something explicit in config.h. In addition:
2559
2560 (a) pcreposix.h still had just "extern" instead of either of these macros;
2561 I have replaced it with PCRE_DATA_SCOPE.
2562
2563 (b) Functions such as _pcre_xclass(), which are internal to the library,
2564 but external in the C sense, all had PCRE_EXPORT in their definitions.
2565 This is apparently wrong for the Windows case, so I have removed it.
2566 (It makes no difference on Unix-like systems.)
2567
2568 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2569 of recursive calls to match(). This is different to MATCH_LIMIT because
2570 that limits the total number of calls to match(), not all of which increase
2571 the depth of recursion. Limiting the recursion depth limits the amount of
2572 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2573 when PCRE is compiled, and changed at run time. A patch from Google adds
2574 this functionality to the C++ interface.
2575
2576 18. Changes to the handling of Unicode character properties:
2577
2578 (a) Updated the table to Unicode 4.1.0.
2579
2580 (b) Recognize characters that are not in the table as "Cn" (undefined).
2581
2582 (c) I revised the way the table is implemented to a much improved format
2583 which includes recognition of ranges. It now supports the ranges that
2584 are defined in UnicodeData.txt, and it also amalgamates other
2585 characters into ranges. This has reduced the number of entries in the
2586 table from around 16,000 to around 3,000, thus reducing its size
2587 considerably. I realized I did not need to use a tree structure after
2588 all - a binary chop search is just as efficient. Having reduced the
2589 number of entries, I extended their size from 6 bytes to 8 bytes to
2590 allow for more data.
2591
2592 (d) Added support for Unicode script names via properties such as \p{Han}.
2593
2594 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2595 matching that character.
2596
2597 20. When matching a repeated Unicode property with a minimum greater than zero,
2598 (for example \pL{2,}), PCRE could look past the end of the subject if it
2599 reached it while seeking the minimum number of characters. This could
2600 happen only if some of the characters were more than one byte long, because
2601 there is a check for at least the minimum number of bytes.
2602
2603 21. Refactored the implementation of \p and \P so as to be more general, to
2604 allow for more different types of property in future. This has changed the
2605 compiled form incompatibly. Anybody with saved compiled patterns that use
2606 \p or \P will have to recompile them.
2607
2608 22. Added "Any" and "L&" to the supported property types.
2609
2610 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2611 but give a compile time error if the value is greater than 0xff.
2612
2613 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2614 accidentally not being installed or uninstalled.
2615
2616 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2617 made were to insert the current release number. This seemed silly, because
2618 it made things harder for people building PCRE on systems that don't run
2619 "configure". I have turned pcre.h into a distributed file, no longer built
2620 by "configure", with the version identification directly included. There is
2621 no longer a pcre.h.in file.
2622
2623 However, this change necessitated a change to the pcre-config script as
2624 well. It is built from pcre-config.in, and one of the substitutions was the
2625 release number. I have updated configure.ac so that ./configure now finds
2626 the release number by grepping pcre.h.
2627
2628 26. Added the ability to run the tests under valgrind.
2629
2630
2631 Version 6.4 05-Sep-05
2632 ---------------------
2633
2634 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2635 "--" to be printed when multiple files were scanned, even when none of the
2636 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2637 consider it to be a bug, and have restored the previous behaviour.
2638
2639 2. A couple of code tidies to get rid of compiler warnings.
2640
2641 3. The pcretest program used to cheat by referring to symbols in the library
2642 whose names begin with _pcre_. These are internal symbols that are not
2643 really supposed to be visible externally, and in some environments it is
2644 possible to suppress them. The cheating is now confined to including
2645 certain files from the library's source, which is a bit cleaner.
2646
2647 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2648 file's purpose clearer.
2649
2650 5. Reorganized pcre_ucp_findchar().
2651
2652
2653 Version 6.3 15-Aug-05
2654 ---------------------
2655
2656 1. The file libpcre.pc.in did not have general read permission in the tarball.
2657
2658 2. There were some problems when building without C++ support:
2659
2660 (a) If C++ support was not built, "make install" and "make test" still
2661 tried to test it.
2662
2663 (b) There were problems when the value of CXX was explicitly set. Some
2664 changes have been made to try to fix these, and ...
2665
2666 (c) --disable-cpp can now be used to explicitly disable C++ support.
2667
2668 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2669 backslash in a target when C++ was disabled. This confuses some
2670 versions of "make", apparently. Using an intermediate variable solves
2671 this. (Same for CPP_LOBJ.)
2672
2673 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2674 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2675 necessary on certain architectures.
2676
2677 4. Added a setting of -export-symbols-regex to the link command to remove
2678 those symbols that are exported in the C sense, but actually are local
2679 within the library, and not documented. Their names all begin with
2680 "_pcre_". This is not a perfect job, because (a) we have to except some
2681 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2682 available (and never for static libraries). I have made a note to try to
2683 find a way round (a) in the future.
2684
2685
2686 Version 6.2 01-Aug-05
2687 ---------------------
2688
2689 1. There was no test for integer overflow of quantifier values. A construction
2690 such as {1111111111111111} would give undefined results. What is worse, if
2691 a minimum quantifier for a parenthesized subpattern overflowed and became
2692 negative, the calculation of the memory size went wrong. This could have
2693 led to memory overwriting.
2694
2695 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2696
2697 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2698 operating environments where this matters.
2699
2700 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2701 PCRE options from within the C++ wrapper.
2702
2703 5. Named capturing subpatterns were not being correctly counted when a pattern
2704 was compiled. This caused two problems: (a) If there were more than 100
2705 such subpatterns, the calculation of the memory needed for the whole
2706 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2707 back references of the form \12, where the number was greater than 9, were
2708 not recognized as back references, even though there were sufficient
2709 previous subpatterns.
2710
2711 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2712 versions of gcc, e.g. 2.95.4.
2713
2714
2715 Version 6.1 21-Jun-05
2716 ---------------------
2717
2718 1. There was one reference to the variable "posix" in pcretest.c that was not
2719 surrounded by "#if !defined NOPOSIX".
2720
2721 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2722 the cross-check on the old pcre_info() function, for the benefit of the
2723 cut-down version of PCRE that is currently imported into Exim.
2724
2725 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2726 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2727 patterns that start (?i) but I don't think that matters. The use of (?i) is
2728 just an example; this all applies to the other options as well.
2729
2730 4. Since libtool seems to echo the compile commands it is issuing, the output
2731 from "make" can be reduced a bit by putting "@" in front of each libtool
2732 compile command.
2733
2734 5. Patch from the folks at Google for configure.in to be a bit more thorough
2735 in checking for a suitable C++ installation before trying to compile the
2736 C++ stuff. This should fix a reported problem when a compiler was present,
2737 but no suitable headers.
2738
2739 6. The man pages all had just "PCRE" as their title. I have changed them to
2740 be the relevant file name. I have also arranged that these names are
2741 retained in the file doc/pcre.txt, which is a concatenation in text format
2742 of all the man pages except the little individual ones for each function.
2743
2744 7. The NON-UNIX-USE file had not been updated for the different set of source
2745 files that come with release 6. I also added a few comments about the C++
2746 wrapper.
2747
2748
2749 Version 6.0 07-Jun-05
2750 ---------------------
2751
2752 1. Some minor internal re-organization to help with my DFA experiments.
2753
2754 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2755 didn't matter for the library itself when fully configured, but did matter
2756 when compiling without UCP support, or within Exim, where the ucp files are
2757 not imported.
2758
2759 3. Refactoring of the library code to split up the various functions into
2760 different source modules. The addition of the new DFA matching code (see
2761 below) to a single monolithic source would have made it really too
2762 unwieldy, quite apart from causing all the code to be include in a
2763 statically linked application, when only some functions are used. This is
2764 relevant even without the DFA addition now that patterns can be compiled in
2765 one application and matched in another.
2766
2767 The downside of splitting up is that there have to be some external
2768 functions and data tables that are used internally in different modules of
2769 the library but which are not part of the API. These have all had their
2770 names changed to start with "_pcre_" so that they are unlikely to clash
2771 with other external names.
2772
2773 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2774 a different (DFA) algorithm. Although it is slower than the original
2775 function, it does have some advantages for certain types of matching
2776 problem.
2777
2778 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2779 including restarting after a partial match.
2780
2781 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2782 defined when compiling for Windows was sent to me. I have put it into the
2783 code, though I have no means of testing or verifying it.
2784
2785 7. Added the pcre_refcount() auxiliary function.
2786
2787 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2788 match before or at the first newline in the subject string. In pcretest,
2789 the /f option on a pattern can be used to set this.
2790
2791 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2792 would behave wrongly. This has been present in PCRE since release 4.0.
2793
2794 10. A number of changes to the pcregrep command:
2795
2796 (a) Refactored how -x works; insert ^(...)$ instead of setting
2797 PCRE_ANCHORED and checking the length, in preparation for adding
2798 something similar for -w.
2799
2800 (b) Added the -w (match as a word) option.
2801
2802 (c) Refactored the way lines are read and buffered so as to have more
2803 than one at a time available.
2804
2805 (d) Implemented a pcregrep test script.
2806
2807 (e) Added the -M (multiline match) option. This allows patterns to match
2808 over several lines of the subject. The buffering ensures that at least
2809 8K, or the rest of the document (whichever is the shorter) is available
2810 for matching (and similarly the previous 8K for lookbehind assertions).
2811
2812 (f) Changed the --help output so that it now says
2813
2814 -w, --word-regex(p)
2815
2816 instead of two lines, one with "regex" and the other with "regexp"
2817 because that confused at least one person since the short forms are the
2818 same. (This required a bit of code, as the output is generated
2819 automatically from a table. It wasn't just a text change.)
2820
2821 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2822 option but starts with a hyphen. Could be a pattern or a path name
2823 starting with a hyphen, for instance.
2824
2825 (h) "-" can be given as a file name to represent stdin.
2826
2827 (i) When file names are being printed, "(standard input)" is used for
2828 the standard input, for compatibility with GNU grep. Previously
2829 "<stdin>" was used.
2830
2831 (j) The option --label=xxx can be used to supply a name to be used for
2832 stdin when file names are being printed. There is no short form.
2833
2834 (k) Re-factored the options decoding logic because we are going to add
2835 two more options that take data. Such options can now be given in four
2836 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2837
2838 (l) Added the -A, -B, and -C options for requesting that lines of context
2839 around matches be printed.
2840
2841 (m) Added the -L option to print the names of files that do not contain
2842 any matching lines, that is, the complement of -l.
2843
2844 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2845 continue to scan other files.
2846
2847 (o) The -s option was incorrectly implemented. For compatibility with other
2848 greps, it now suppresses the error message for a non-existent or non-
2849 accessible file (but not the return code). There is a new option called
2850 -q that suppresses the output of matching lines, which was what -s was
2851 previously doing.
2852
2853 (p) Added --include and --exclude options to specify files for inclusion
2854 and exclusion when recursing.
2855
2856 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2857 Hopefully, it now does.
2858
2859 12. Missing cast in pcre_study().
2860
2861 13. Added an "uninstall" target to the makefile.
2862
2863 14. Replaced "extern" in the function prototypes in Makefile.in with
2864 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2865 world, but is set differently for Windows.
2866
2867 15. Added a second compiling function called pcre_compile2(). The only
2868 difference is that it has an extra argument, which is a pointer to an
2869 integer error code. When there is a compile-time failure, this is set
2870 non-zero, in addition to the error test pointer being set to point to an
2871 error message. The new argument may be NULL if no error number is required
2872 (but then you may as well call pcre_compile(), which is now just a
2873 wrapper). This facility is provided because some applications need a
2874 numeric error indication, but it has also enabled me to tidy up the way
2875 compile-time errors are handled in the POSIX wrapper.
2876
2877 16. Added VPATH=.libs to the makefile; this should help when building with one
2878 prefix path and installing with another. (Or so I'm told by someone who
2879 knows more about this stuff than I do.)
2880
2881 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2882 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2883 match everything, including newlines. This is not POSIX-compatible, but
2884 somebody wanted the feature. From pcretest it can be activated by using
2885 both the P and the s flags.
2886
2887 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2888
2889 19. libpcre.pc was being incorrectly installed as executable.
2890
2891 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2892 it now also looks for '\r' so that it will work unmodified on Windows.
2893
2894 21. Added Google's contributed C++ wrapper to the distribution.
2895
2896 22. Added some untidy missing memory free() calls in pcretest, to keep
2897 Electric Fence happy when testing.
2898
2899
2900
2901 Version 5.0 13-Sep-04
2902 ---------------------
2903
2904 1. Internal change: literal characters are no longer packed up into items
2905 containing multiple characters in a single byte-string. Each character
2906 is now matched using a separate opcode. However, there may be more than one
2907 byte in the character in UTF-8 mode.
2908
2909 2. The pcre_callout_block structure has two new fields: pattern_position and
2910 next_item_length. These contain the offset in the pattern to the next match
2911 item, and its length, respectively.
2912
2913 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2914 insertion of callouts before each pattern item. Added the /C option to
2915 pcretest to make use of this.
2916
2917 4. On the advice of a Windows user, the lines
2918
2919 #if defined(_WIN32) || defined(WIN32)
2920 _setmode( _fileno( stdout ), 0x8000 );
2921 #endif /* defined(_WIN32) || defined(WIN32) */
2922
2923 have been added to the source of pcretest. This apparently does useful
2924 magic in relation to line terminators.
2925
2926 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2927 for the benefit of those environments where the "b" makes a difference.
2928
2929 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2930 to know about it. I have put a hack into configure.in that adds in code
2931 to set GCC=yes if CC=icc. This seems to end up at a point in the
2932 generated configure script that is early enough to affect the setting of
2933 compiler options, which is what is needed, but I have no means of testing
2934 whether it really works. (The user who reported this had patched the
2935 generated configure script, which of course I cannot do.)
2936
2937 LATER: After change 22 below (new libtool files), the configure script
2938 seems to know about icc (and also ecc). Therefore, I have commented out
2939 this hack in configure.in.
2940
2941 7. Added support for pkg-config (2 patches were sent in).
2942
2943 8. Negated POSIX character classes that used a combination of internal tables
2944 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2945 [[:^ascii]]. Typically, they would match almost any characters. The other
2946 POSIX classes were not broken in this way.
2947
2948 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2949 to find the match, as PCRE was deluded into thinking that the match had to
2950 start at the start point or following a newline. The same bug applied to
2951 patterns with negative forward assertions or any backward assertions
2952 preceding ".*" at the start, unless the pattern required a fixed first
2953 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2954
2955 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2956 starting at the last subject character, bytes beyond the end of the subject
2957 string were read.
2958
2959 11. Renamed the variable "class" as "classbits" to make life easier for C++
2960 users. (Previously there was a macro definition, but it apparently wasn't
2961 enough.)
2962
2963 12. Added the new field "tables" to the extra data so that tables can be passed
2964 in at exec time, or the internal tables can be re-selected. This allows
2965 a compiled regex to be saved and re-used at a later time by a different
2966 program that might have everything at different addresses.
2967
2968 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2969 -R library as well as a -L library.
2970
2971 14. The debugging options of pcretest (-d on the command line or D on a
2972 pattern) showed incorrect output for anything following an extended class
2973 that contained multibyte characters and which was followed by a quantifier.
2974
2975 15. Added optional support for general category Unicode character properties
2976 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2977 support. It adds about 90K to the size of the library. The meanings of the
2978 inbuilt class escapes such as \d and \s have NOT been changed.
2979
2980 16. Updated pcredemo.c to include calls to free() to release the memory for the
2981 compiled pattern.
2982
2983 17. The generated file chartables.c was being created in the source directory
2984 instead of in the building directory. This caused the build to fail if the
2985 source directory was different from the building directory, and was
2986 read-only.
2987
2988 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2989 file. No doubt somebody will tell me if they don't make sense... Also added
2990 Dan Mooney's comments about building on OpenVMS.
2991
2992 19. Added support for partial matching via the PCRE_PARTIAL option for
2993 pcre_exec() and the \P data escape in pcretest.
2994
2995 20. Extended pcretest with 3 new pattern features:
2996
2997 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2998 write the compiled pattern to the file whose name is "rest-of-line".
2999 This is a straight binary dump of the data, with the saved pointer to
3000 the character tables forced to be NULL. The study data, if any, is
3001 written too. After writing, pcretest reads a new pattern.
3002
3003 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3004 compiled pattern from the given file. There must not be any
3005 occurrences of "<" in the file name (pretty unlikely); if there are,
3006 pcretest will instead treat the initial "<" as a pattern delimiter.
3007 After reading in the pattern, pcretest goes on to read data lines as
3008 usual.
3009
3010 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3011 and 16-bit fields in a compiled pattern, to simulate a pattern that
3012 was compiled on a host of opposite endianness.
3013
3014 21. The pcre-exec() function can now cope with patterns that were compiled on
3015 hosts of opposite endianness, with this restriction:
3016
3017 As for any compiled expression that is saved and used later, the tables
3018 pointer field cannot be preserved; the extra_data field in the arguments
3019 to pcre_exec() should be used to pass in a tables address if a value
3020 other than the default internal tables were used at compile time.
3021
3022 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3023 now diagnosed as an error. Previously, most of the time, a negative number
3024 would have been treated as zero, but if in addition "ovector" was passed as
3025 NULL, a crash could occur.
3026
3027 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3028 new versions from the libtool 1.5 distribution (the last one is a copy of
3029 a file called libtool.m4). This seems to have fixed the need to patch
3030 "configure" to support Darwin 1.3 (which I used to do). However, I still
3031 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3032 workstation).
3033
3034 24. Changed the PCRE licence to be the more standard "BSD" licence.
3035
3036
3037 Version 4.5 01-Dec-03
3038 ---------------------
3039
3040 1. There has been some re-arrangement of the code for the match() function so
3041 that it can be compiled in a version that does not call itself recursively.
3042 Instead, it keeps those local variables that need separate instances for
3043 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3044 needs to "recurse". Keeping track of where control must go is done by means
3045 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3046 hide most of the details from the main code, and operates only if
3047 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3048 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3049 operating.
3050
3051 To make it easier for callers to provide specially tailored get/free
3052 functions for this usage, two new functions, pcre_stack_malloc, and
3053 pcre_stack_free, are used. They are always called in strict stacking order,
3054 and the size of block requested is always the same.
3055
3056 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3057 PCRE has been compiled to use the stack or the heap for recursion. The
3058 -C option of pcretest uses this to show which version is compiled.
3059
3060 A new data escape \S, is added to pcretest; it causes the amounts of store
3061 obtained and freed by both kinds of malloc/free at match time to be added
3062 to the output.
3063
3064 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3065 what's available on my current Linux desktop machine.
3066
3067 3. When matching a UTF-8 string, the test for a valid string at the start has
3068 been extended. If start_offset is not zero, PCRE now checks that it points
3069 to a byte that is the start of a UTF-8 character. If not, it returns
3070 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3071 this is necessary because there may be backward assertions in the pattern.
3072 When matching the same subject several times, it may save resources to use
3073 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3074
3075 4. The code for checking the validity of UTF-8 strings has been tightened so
3076 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3077 containing "overlong sequences".
3078
3079 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3080 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3081 should have been "&", but it just so happened that all the cases this let
3082 through by mistake were picked up later in the function.
3083
3084 6. I had used a variable called "isblank" - this is a C99 function, causing
3085 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3086
3087 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3088 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3089 so the version is shown; (c) stop "make test" echoing "./RunTest".
3090
3091 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3092
3093 9. The prototype for memmove() for systems that don't have it was using
3094 size_t, but the inclusion of the header that defines size_t was later. I've
3095 moved the #includes for the C headers earlier to avoid this.
3096
3097 10. Added some adjustments to the code to make it easier to compiler on certain
3098 special systems:
3099
3100 (a) Some "const" qualifiers were missing.
3101 (b) Added the macro EXPORT before all exported functions; by default this
3102 is defined to be empty.
3103 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3104 that it reads its output file name as an argument instead of writing
3105 to the standard output and assuming this can be redirected.
3106
3107 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3108 class containing characters with values greater than 255, PCRE compilation
3109 went into a loop.
3110
3111 12. A recursive reference to a subpattern that was within another subpattern
3112 that had a minimum quantifier of zero caused PCRE to crash. For example,
3113 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3114 recursion. If the recursively-called subpattern itself had a zero repeat,
3115 that was OK.
3116
3117 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3118 buffer into which it was copied (for escape processing) was still set at
3119 1024, so long lines caused crashes.
3120
3121 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3122 "internal error: code overflow...". This applied to any character class
3123 that was followed by a possessive quantifier.
3124
3125 15. Modified the Makefile to add libpcre.la as a prerequisite for
3126 libpcreposix.la because I was told this is needed for a parallel build to
3127 work.
3128
3129 16. If a pattern that contained .* following optional items at the start was
3130 studied, the wrong optimizing data was generated, leading to matching
3131 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3132 matching string must start with a or b or c. The correct conclusion for
3133 this pattern is that a match can start with any character.
3134
3135
3136 Version 4.4 13-Aug-03
3137 ---------------------
3138
3139 1. In UTF-8 mode, a character class containing characters with values between
3140 127 and 255 was not handled correctly if the compiled pattern was studied.
3141 In fixing this, I have also improved the studying algorithm for such
3142 classes (slightly).
3143
3144 2. Three internal functions had redundant arguments passed to them. Removal
3145 might give a very teeny performance improvement.
3146
3147 3. Documentation bug: the value of the capture_top field in a callout is *one
3148 more than* the number of the hightest numbered captured substring.
3149
3150 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3151 in incorrectly linking with a previously installed version. They now link
3152 explicitly with libpcre.la.
3153
3154 5. configure.in no longer needs to recognize Cygwin specially.
3155
3156 6. A problem in pcre.in for Windows platforms is fixed.
3157
3158 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3159 pcretest, it used to include the size of the study block as part of its
3160 output. Unfortunately, the structure contains a field that has a different
3161 size on different hardware architectures. This meant that the tests that
3162 showed this size failed. As the block is currently always of a fixed size,
3163 this information isn't actually particularly useful in pcretest output, so
3164 I have just removed it.
3165
3166 8. Three pre-processor statements accidentally did not start in column 1.
3167 Sadly, there are *still* compilers around that complain, even though
3168 standard C has not required this for well over a decade. Sigh.
3169
3170 9. In pcretest, the code for checking callouts passed small integers in the
3171 callout_data field, which is a void * field. However, some picky compilers
3172 complained about the casts involved for this on 64-bit systems. Now
3173 pcretest passes the address of the small integer instead, which should get
3174 rid of the warnings.
3175
3176 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3177 both compile and run time, and gives an error if an invalid UTF-8 sequence
3178 is found. There is a option for disabling this check in cases where the
3179 string is known to be correct and/or the maximum performance is wanted.
3180
3181 11. In response to a bug report, I changed one line in Makefile.in from
3182
3183 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3184 to
3185 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3186
3187 to look similar to other lines, but I have no way of telling whether this
3188 is the right thing to do, as I do not use Windows. No doubt I'll get told
3189 if it's wrong...
3190
3191
3192 Version 4.3 21-May-03
3193 ---------------------
3194
3195 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3196 Makefile.
3197
3198 2. Some refactoring to improve the quality of the code:
3199
3200 (i) The utf8_table... variables are now declared "const".
3201
3202 (ii) The code for \cx, which used the "case flipping" table to upper case
3203 lower case letters, now just substracts 32. This is ASCII-specific,
3204 but the whole concept of \cx is ASCII-specific, so it seems
3205 reasonable.
3206
3207 (iii) PCRE was using its character types table to recognize decimal and
3208 hexadecimal digits in the pattern. This is silly, because it handles
3209 only 0-9, a-f, and A-F, but the character types table is locale-
3210 specific, which means strange things might happen. A private
3211 table is now used for this - though it costs 256 bytes, a table is
3212 much faster than multiple explicit tests. Of course, the standard
3213 character types table is still used for matching digits in subject
3214 strings against \d.
3215
3216 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3217 ending in _t are). So I've renamed it as ESC_tee.
3218
3219 3. The first argument for regexec() in the POSIX wrapper should have been
3220 defined as "const".
3221
3222 4. Changed pcretest to use malloc() for its buffers so that they can be
3223 Electric Fenced for debugging.
3224
3225 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3226 to read one or more bytes before the start of the subject string. Often this
3227 had no effect on PCRE's behaviour, but in some circumstances it could
3228 provoke a segmentation fault.
3229
3230 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3231 to try to read one or more bytes before the start of the subject string.
3232
3233 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3234 UTF-8 support could misbehave in various ways if the subject string
3235 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3236 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3237 back over UTF-8 characters.)
3238
3239
3240 Version 4.2 14-Apr-03
3241 ---------------------
3242
3243 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3244
3245 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3246 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3247 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3248 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3249 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3250 and BUILD_EXEEXT
3251 Note: automatic setting of the BUILD variables is not yet working
3252 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3253 compile-time but not at link-time
3254 [LINK]: use for linking executables only
3255 make different versions for Windows and non-Windows
3256 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3257 libraries
3258 [LINK_FOR_BUILD]: new variable
3259 [OBJEXT]: use throughout
3260 [EXEEXT]: use throughout
3261 <winshared>: new target
3262 <wininstall>: new target
3263 <dftables.o>: use native compiler
3264 <dftables>: use native linker
3265 <install>: handle Windows platform correctly
3266 <clean>: ditto
3267 <check>: ditto
3268 copy DLL to top builddir before testing
3269
3270 As part of these changes, -no-undefined was removed again. This was reported
3271 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3272 in any case.
3273
3274 3. Some tidies to get rid of compiler warnings:
3275
3276 . In the match_data structure, match_limit was an unsigned long int, whereas
3277 match_call_count was an int. I've made them both unsigned long ints.
3278
3279 . In pcretest the fact that a const uschar * doesn't automatically cast to
3280 a void * provoked a warning.
3281
3282 . Turning on some more compiler warnings threw up some "shadow" variables
3283 and a few more missing casts.
3284
3285 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3286 option, a class that contained a single character with a value between 128
3287 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3288
3289 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3290 option, a class that contained several characters, but with at least one
3291 whose value was between 128 and 255 caused PCRE to crash.
3292
3293
3294 Version 4.1 12-Mar-03
3295 ---------------------
3296
3297 1. Compiling with gcc -pedantic found a couple of places where casts were
3298 needed, and a string in dftables.c that was longer than standard compilers are
3299 required to support.
3300
3301 2. Compiling with Sun's compiler found a few more places where the code could
3302 be tidied up in order to avoid warnings.
3303
3304 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3305 first of these names is deprecated in the latest Autoconf in favour of the name
3306 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3307 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3308 analogy I have changed it to CFLAGS_FOR_BUILD.
3309
3310 4. Added -no-undefined to the linking command in the Makefile, because this is
3311 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3312 linking step for the pcreposix library.
3313
3314 5. PCRE was failing to diagnose the case of two named groups with the same
3315 name.
3316
3317 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3318 literal character that is needed in the subject for a match, and scans along to
3319 ensure that it is present before embarking on the full matching process. This
3320 saves time in cases of nested unlimited repeats that are never going to match.
3321 Problem: the scan can take a lot of time if the subject is very long (e.g.
3322 megabytes), thus penalizing straightforward matches. It is now done only if the
3323 amount of subject to be scanned is less than 1000 bytes.
3324
3325 7. A lesser problem with the same optimization is that it was recording the
3326 first character of an anchored pattern as "needed", thus provoking a search
3327 right along the subject, even when the first match of the pattern was going to
3328 fail. The "needed" character is now not set for anchored patterns, unless it
3329 follows something in the pattern that is of non-fixed length. Thus, it still
3330 fulfils its original purpose of finding quick non-matches in cases of nested
3331 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3332
3333
3334 Version 4.0 17-Feb-03
3335 ---------------------
3336
3337 1. If a comment in an extended regex that started immediately after a meta-item
3338 extended to the end of string, PCRE compiled incorrect data. This could lead to
3339 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3340
3341 2. Moved to autoconf 2.53 and libtool 1.4.2.
3342
3343 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3344 the special perltest8 script is no longer needed - all the tests can be run
3345 from a single perltest script.
3346
3347 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3348 by \s. It has now been removed in PCRE. This means it isn't recognized as
3349 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3350 class [:space:] *does* include VT, thereby creating a mess.
3351
3352 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3353 space and tab.
3354
3355 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3356 its new features into the main test script, reducing the number of scripts.
3357
3358 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3359 were backward compatible, and made the (?i) apply to the whole pattern, as if
3360 /i were given. Now it behaves more logically, and applies the option setting
3361 only to what follows. PCRE has been changed to follow suit. However, if it
3362 finds options settings right at the start of the pattern, it extracts them into
3363 the global options, as before. Thus, they show up in the info data.
3364
3365 8. Added support for the \Q...\E escape sequence. Characters in between are
3366 treated as literals. This is slightly different from Perl in that $ and @ are
3367 also handled as literals inside the quotes. In Perl, they will cause variable
3368 interpolation. Note the following examples:
3369
3370 Pattern PCRE matches Perl matches
3371
3372 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3373 \Qabc\$xyz\E abc\$xyz abc\$xyz
3374 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3375
3376 For compatibility with Perl, \Q...\E sequences are recognized inside character
3377 classes as well as outside them.
3378
3379 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3380 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3381 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3382 signed/unsigned warnings.
3383
3384 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3385 option for pcretest, so I've replaced it by a simple function that does just
3386 that job.
3387
3388 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3389 "pcregrep -".
3390
3391 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3392 Java package. This provides some syntactic sugar for simple cases of what my
3393 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3394 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3395 item, you can use this simplified notation. Note that only makes sense with
3396 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3397 greediness, whatever the setting of the PCRE_UNGREEDY option.
3398
3399 13. A change of greediness default within a pattern was not taking effect at
3400 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3401 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3402 was abstracted outside.
3403
3404 14. PCRE now supports the \G assertion. It is true when the current matching
3405 position is at the start point of the match. This differs from \A when the
3406 starting offset is non-zero. Used with the /g option of pcretest (or similar
3407 code), it works in the same way as it does for Perl's /g option. If all
3408 alternatives of a regex begin with \G, the expression is anchored to the start
3409 match position, and the "anchored" flag is set in the compiled expression.
3410
3411 15. Some bugs concerning the handling of certain option changes within patterns
3412 have been fixed. These applied to options other than (?ims). For example,
3413 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3414 been the other way round. Some of this was related to change 7 above.
3415
3416 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3417 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3418 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3419 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3420
3421 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3422 mode. Unlike ".", it always matches newline, whatever the setting of
3423 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3424 assertions. Perl allows it, but it doesn't (in general) work because it can't
3425 calculate the length of the lookbehind. At least, that's the case for Perl
3426 5.8.0 - I've been told they are going to document that it doesn't work in
3427 future.
3428
3429 18. Added an error diagnosis for escapes that PCRE does not support: these are
3430 \L, \l, \N, \P, \p, \U, \u, and \X.
3431
3432 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3433 reading past the end of the pattern in cases such as /[abcd/.
3434
3435 20. PCRE was getting more memory than necessary for patterns with classes that
3436 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3437
3438 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3439 compiling PCRE for use with Virtual Pascal.
3440
3441 22. Small fix to the Makefile to make it work properly if the build is done
3442 outside the source tree.
3443
3444 23. Added a new extension: a condition to go with recursion. If a conditional
3445 subpattern starts with (?(R) the "true" branch is used if recursion has
3446 happened, whereas the "false" branch is used only at the top level.
3447
3448 24. When there was a very long string of literal characters (over 255 bytes
3449 without UTF support, over 250 bytes with UTF support), the computation of how
3450 much memory was required could be incorrect, leading to segfaults or other
3451 strange effects.
3452
3453 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3454 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3455 there was a subsequent back reference to those brackets. This meant that, for
3456 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3457 possible to check for precisely this case. All we can do is abandon the
3458 optimization if .* occurs inside capturing brackets when there are any back
3459 references whatsoever. (See below for a better fix that came later.)
3460
3461 26. The handling of the optimization for finding the first character of a
3462 non-anchored pattern, and for finding a character that is required later in the
3463 match were failing in some cases. This didn't break the matching; it just
3464 failed to optimize when it could. The way this is done has been re-implemented.
3465
3466 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3467
3468 28. Added a new feature that provides some of the functionality that Perl
3469 provides with (?{...}). The facility is termed a "callout". The way it is done
3470 in PCRE is for the caller to provide an optional function, by setting
3471 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3472 global variable. By default it is unset, which disables all calling out. To get
3473 the function called, the regex must include (?C) at appropriate points. This
3474 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3475 This provides a means of identifying different callout points. When PCRE
3476 reaches such a point in the regex, if pcre_callout has been set, the external
3477 function is called. It is provided with data in a structure called
3478 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3479 matching continues; if it returns a non-zero value, the match at the current
3480 point fails. However, backtracking will occur if possible. [This was changed
3481 later and other features added - see item 49 below.]
3482
3483 29. pcretest is upgraded to test the callout functionality. It provides a
3484 callout function that displays information. By default, it shows the start of
3485 the match and the current position in the text. There are some new data escapes
3486 to vary what happens:
3487
3488 \C+ in addition, show current contents of captured substrings
3489 \C- do not supply a callout function
3490 \C!n return 1 when callout number n is reached
3491 \C!n!m return 1 when callout number n is reached for the mth time
3492
3493 30. If pcregrep was called with the -l option and just a single file name, it
3494 output "<stdin>" if a match was found, instead of the file name.
3495
3496 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3497 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3498 pcre_exec(). This saves a malloc/free per call. The default value of
3499 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3500 when configuring.
3501
3502 32. The default maximum size of a compiled pattern is 64K. There have been a
3503 few cases of people hitting this limit. The code now uses macros to handle the
3504 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3505 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3506 configuring. Tests 2 and 5 work only with 2-byte links because they output
3507 debugging information about compiled patterns.
3508
3509 33. Internal code re-arrangements:
3510
3511 (a) Moved the debugging function for printing out a compiled regex into
3512 its own source file (printint.c) and used #include to pull it into
3513 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3514 separate copies.
3515
3516 (b) Defined the list of op-code names for debugging as a macro in
3517 internal.h so that it is next to the definition of the opcodes.
3518
3519 (c) Defined a table of op-code lengths for simpler skipping along compiled
3520 code. This is again a macro in internal.h so that it is next to the
3521 definition of the opcodes.
3522
3523 34. Added support for recursive calls to individual subpatterns, along the
3524 lines of Robin Houston's patch (but implemented somewhat differently).
3525
3526 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3527 allow it to read and process whole directories in Win32. This code was
3528 contributed by Lionel Fourquaux; it has not been tested by me.
3529
3530 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3531 used to name a group. Names consist of alphanumerics and underscores, and must
3532 be unique. Back references use the syntax (?P=name) and recursive calls use
3533 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3534 numbers. The function pcre_fullinfo() can be used after compilation to extract
3535 a name/number map. There are three relevant calls:
3536
3537 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3538 PCRE_INFO_NAMECOUNT yields the number of entries
3539 PCRE_INFO_NAMETABLE yields a pointer to the map.
3540
3541 The map is a vector of fixed-size entries. The size of each entry depends on
3542 the length of the longest name used. The first two bytes of each entry are the
3543 group number, most significant byte first. There follows the corresponding
3544 name, zero terminated. The names are in alphabetical order.
3545
3546 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3547 case instead of 255. Making it the same both with and without UTF-8 support
3548 means that the same test output works with both.
3549
3550 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3551 calling malloc() with a zero argument.
3552
3553 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3554 optimization. I've improved things by keeping a bitmap of backreferences with
3555 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3556 fact referenced, the optimization can be applied. It is unlikely that a
3557 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3558 the match to follow \n) will appear inside brackets with a number greater than
3559 31, but if it does, any back reference > 31 suppresses the optimization.
3560
3561 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3562 of disabling numbered capturing parentheses. Any opening parenthesis that is
3563 not followed by ? behaves as if it were followed by ?: but named parentheses
3564 can still be used for capturing (and they will acquire numbers in the usual
3565 way).
3566
3567 41. Redesigned the return codes from the match() function into yes/no/error so
3568 that errors can be passed back from deep inside the nested calls. A malloc
3569 failure while inside a recursive subpattern call now causes the
3570 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3571
3572 42. It is now possible to set a limit on the number of times the match()
3573 function is called in a call to pcre_exec(). This facility makes it possible to
3574 limit the amount of recursion and backtracking, though not in a directly
3575 obvious way, because the match() function is used in a number of different
3576 circumstances. The count starts from zero for each position in the subject
3577 string (for non-anchored patterns). The default limit is, for compatibility, a
3578 large number, namely 10 000 000. You can change this in two ways:
3579
3580 (a) When configuring PCRE before making, you can use --with-match-limit=n
3581 to set a default value for the compiled library.
3582
3583 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3584 a different value is set. See 45 below.
3585
3586 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3587
3588 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3589 of things that can be changed at compile time. The first argument specifies
3590 what is wanted and the second points to where the information is to be placed.
3591 The current list of available information is:
3592
3593 PCRE_CONFIG_UTF8
3594
3595 The output is an integer that is set to one if UTF-8 support is available;
3596 otherwise it is set to zero.
3597
3598 PCRE_CONFIG_NEWLINE
3599
3600 The output is an integer that it set to the value of the code that is used for
3601 newline. It is either LF (10) or CR (13).
3602
3603 PCRE_CONFIG_LINK_SIZE
3604
3605 The output is an integer that contains the number of bytes used for internal
3606 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3607
3608 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3609
3610 The output is an integer that contains the threshold above which the POSIX
3611 interface uses malloc() for output vectors. See item 31 above.
3612
3613 PCRE_CONFIG_MATCH_LIMIT
3614
3615 The output is an unsigned integer that contains the default limit of the number
3616 of match() calls in a pcre_exec() execution. See 42 above.
3617
3618 44. pcretest has been upgraded by the addition of the -C option. This causes it
3619 to extract all the available output from the new pcre_config() function, and to
3620 output it. The program then exits immediately.
3621
3622 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3623 order to support additional features. One way would have been to define
3624 pcre_exec2() (for example) with extra arguments, but this would not have been
3625 extensible, and would also have required all calls to the original function to
3626 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3627 is used for passing in "extra" data from pcre_study().
3628
3629 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3630 contains the following fields:
3631
3632 flags a bitmap indicating which of the following fields are set
3633 study_data opaque data from pcre_study()
3634 match_limit a way of specifying a limit on match() calls for a specific
3635 call to pcre_exec()
3636 callout_data data for callouts (see 49 below)
3637
3638 The flag bits are also defined in pcre.h, and are
3639
3640 PCRE_EXTRA_STUDY_DATA
3641 PCRE_EXTRA_MATCH_LIMIT
3642 PCRE_EXTRA_CALLOUT_DATA
3643
3644 The pcre_study() function now returns one of these new pcre_extra blocks, with
3645 the actual study data pointed to by the study_data field, and the
3646 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3647 before. That is, this change is entirely upwards-compatible and requires no
3648 change to existing code.
3649
3650 If you want to pass in additional data to pcre_exec(), you can either place it
3651 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3652 block.
3653
3654 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3655 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3656 times with different match limits, until it finds the minimum value needed for
3657 pcre_exec() to complete. The value is then output. This can be instructive; for
3658 most simple matches the number is quite small, but for pathological cases it
3659 gets very large very quickly.
3660
3661 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3662 returns the size of the data block pointed to by the study_data field in a
3663 pcre_extra block, that is, the value that was passed as the argument to
3664 pcre_malloc() when PCRE was getting memory in which to place the information
3665 created by pcre_study(). The fourth argument should point to a size_t variable.
3666 pcretest has been extended so that this information is shown after a successful
3667 pcre_study() call when information about the compiled regex is being displayed.
3668
3669 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3670 because what follows is always an absolute path. (Later: it turns out that this
3671 is more than cosmetic for MinGW, because it doesn't like empty path
3672 components.)
3673
3674 49. Some changes have been made to the callout feature (see 28 above):
3675
3676 (i) A callout function now has three choices for what it returns:
3677
3678 0 => success, carry on matching
3679 > 0 => failure at this point, but backtrack if possible
3680 < 0 => serious error, return this value from pcre_exec()
3681
3682 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3683 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3684 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3685 use by callout functions. It will never be used by PCRE itself.
3686
3687 (ii) The pcre_extra structure (see 45 above) has a void * field called
3688 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3689 pcre_callout_block structure has a field of the same name. The contents of
3690 the field passed in the pcre_extra structure are passed to the callout
3691 function in the corresponding field in the callout block. This makes it
3692 easier to use the same callout-containing regex from multiple threads. For
3693 testing, the pcretest program has a new data escape
3694
3695 \C*n pass the number n (may be negative) as callout_data
3696
3697 If the callout function in pcretest receives a non-zero value as
3698 callout_data, it returns that value.
3699
3700 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3701 there were some redundant $(CFLAGS) in commands that are now specified as
3702 $(LINK), which already includes $(CFLAGS).
3703
3704 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3705 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3706 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3707 one-byte characters throughout. Note that case-insensitive matching applies
3708 only to characters whose values are less than 256. PCRE doesn't support the
3709 notion of cases for higher-valued characters.
3710
3711 (i) A character class whose characters are all within 0-255 is handled as
3712 a bit map, and the map is inverted for negative classes. Previously, a
3713 character > 255 always failed to match such a class; however it should
3714 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3715
3716 (ii) A negated character class with a single character < 255 is coded as
3717 "not this character" (OP_NOT). This wasn't working properly when the test
3718 character was multibyte, either singly or repeated.
3719
3720 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3721 mode, for example: \x{100}{2,3}.
3722
3723 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3724 singly or repeated) now correctly test multibyte characters. However,
3725 PCRE doesn't recognize any characters with values greater than 255 as
3726 digits, spaces, or word characters. Such characters always match \D, \S,
3727 and \W, and never match \d, \s, or \w.
3728
3729 (v) Classes may now contain characters and character ranges with values
3730 greater than 255. For example: [ab\x{100}-\x{400}].
3731
3732 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3733 PCRE in UTF-8 mode.
3734
3735 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3736 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3737 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3738 value.)
3739
3740 53. The single man page has become too large. I have therefore split it up into
3741 a number of separate man pages. These also give rise to individual HTML pages;
3742 these are now put in a separate directory, and there is an index.html page that
3743 lists them all. Some hyperlinking between the pages has been installed.
3744
3745 54. Added convenience functions for handling named capturing parentheses.
3746
3747 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3748 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3749 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3750 are faulted.
3751
3752 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3753 calling configure. These values are used when compiling the dftables.c program
3754 which is run to generate the source of the default character tables. They
3755 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3756 you will need to set these values.
3757
3758 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3759
3760
3761 Version 3.9 02-Jan-02
3762 ---------------------
3763
3764 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3765
3766 2. If --disable-static was given, the building process failed when trying to
3767 build pcretest and pcregrep. (For some reason it was using libtool to compile
3768 them, which is not right, as they aren't part of the library.)
3769
3770
3771 Version 3.8 18-Dec-01
3772 ---------------------
3773
3774 1. The experimental UTF-8 code was completely screwed up. It was packing the
3775 bytes in the wrong order. How dumb can you get?
3776
3777
3778 Version 3.7 29-Oct-01
3779 ---------------------
3780
3781 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3782 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3783 this didn't happen under Solaris 8, where I normally test things.
3784
3785 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3786 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3787 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3788 this fix an hour or so after the initial 3.7 release.)
3789
3790
3791 Version 3.6 23-Oct-01
3792 ---------------------
3793
3794 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3795 offsets passed as NULL with zero offset count.
3796
3797 2. The config.guess and config.sub files had not been updated when I moved to
3798 the latest autoconf.
3799
3800
3801 Version 3.5 15-Aug-01
3802 ---------------------
3803
3804 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3805 had been forgotten.
3806
3807 2. By using declared but undefined structures, we can avoid using "void"
3808 definitions in pcre.h while keeping the internal definitions of the structures
3809 private.
3810
3811 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3812 user point of view, this means that both static and shared libraries are built
3813 by default, but this can be individually controlled. More of the work of
3814 handling this static/shared cases is now inside libtool instead of PCRE's make
3815 file.
3816
3817 4. The pcretest utility is now installed along with pcregrep because it is
3818 useful for users (to test regexs) and by doing this, it automatically gets
3819 relinked by libtool. The documentation has been turned into a man page, so
3820 there are now .1, .txt, and .html versions in /doc.
3821
3822 5. Upgrades to pcregrep:
3823 (i) Added long-form option names like gnu grep.
3824 (ii) Added --help to list all options with an explanatory phrase.
3825 (iii) Added -r, --recursive to recurse into sub-directories.
3826 (iv) Added -f, --file to read patterns from a file.
3827
3828 6. pcre_exec() was referring to its "code" argument before testing that
3829 argument for NULL (and giving an error if it was NULL).
3830
3831 7. Upgraded Makefile.in to allow for compiling in a different directory from
3832 the source directory.
3833
3834 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3835 options bits, the pointer it was passed was to an int instead of to an unsigned
3836 long int. This mattered only on 64-bit systems.
3837
3838 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3839 generated) instead of pcre.in, which it its source. Also made the same change
3840 in several of the .c files.
3841
3842 10. A new release of gcc defines printf() as a macro, which broke pcretest
3843 because it had an ifdef in the middle of a string argument for printf(). Fixed
3844 by using separate calls to printf().
3845
3846 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3847 script, to force use of CR or LF instead of \n in the source. On non-Unix
3848 systems, the value can be set in config.h.
3849
3850 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3851 absolute limit. Changed the text of the error message to make this clear, and
3852 likewise updated the man page.
3853
3854 13. The limit of 99 on the number of capturing subpatterns has been removed.
3855 The new limit is 65535, which I hope will not be a "real" limit.
3856
3857
3858 Version 3.4 22-Aug-00
3859 ---------------------
3860
3861 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3862
3863 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3864
3865
3866 Version 3.3 01-Aug-00
3867 ---------------------
3868
3869 1. If an octal character was given, but the value was greater than \377, it
3870 was not getting masked to the least significant bits, as documented. This could
3871 lead to crashes in some systems.
3872
3873 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3874 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3875
3876 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3877 These just pass their arguments on to (pcre_free)(), but they are provided
3878 because some uses of PCRE bind it to non-C systems that can call its functions,
3879 but cannot call free() or pcre_free() directly.
3880
3881 4. Add "make test" as a synonym for "make check". Corrected some comments in
3882 the Makefile.
3883
3884 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3885 Makefile.
3886
3887 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3888 command called pgrep for grepping around the active processes.
3889
3890 7. Added the beginnings of support for UTF-8 character strings.
3891
3892 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3893 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3894 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3895 out for the ar command.)
3896
3897
3898 Version 3.2 12-May-00
3899 ---------------------
3900
3901 This is purely a bug fixing release.
3902
3903 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3904 of ZA. This was just one example of several cases that could provoke this bug,
3905 which was introduced by change 9 of version 2.00. The code for breaking
3906 infinite loops after an iteration that matches an empty string was't working
3907 correctly.
3908
3909 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3910 when matched against abbab (for example). After matching an empty string, it
3911 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3912 caused it to match further down the string than it should.
3913
3914 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3915 was there because it doesn't seem to be needed, and it causes trouble on some
3916 systems, as it is not a Standard C header. It has been removed.
3917
3918 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3919 were reported on the Macintosh. The changes were from
3920
3921 while ((c = *(++ptr)) != 0 && c != '\n');
3922 to
3923 while ((c = *(++ptr)) != 0 && c != '\n') ;
3924
3925 Totally extraordinary, but if that's what it takes...
3926
3927 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3928 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3929 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3930 assumes the way PCRE uses memmove() (always moving upwards).
3931
3932 6. PCRE is being used in one environment where strchr() is not available. There
3933 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3934 faster code anyway.
3935
3936
3937 Version 3.1 09-Feb-00
3938 ---------------------
3939
3940 The only change in this release is the fixing of some bugs in Makefile.in for
3941 the "install" target:
3942
3943 (1) It was failing to install pcreposix.h.
3944
3945 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3946
3947
3948 Version 3.0 01-Feb-00
3949 ---------------------
3950
3951 1. Add support for the /+ modifier to perltest (to output $` like it does in
3952 pcretest).
3953
3954 2. Add support for the /g modifier to perltest.
3955
3956 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3957 matches null strings.
3958
3959 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3960 pattern. Perl treats empty patterns specially - it reuses the most recent
3961 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3962 effect.
3963
3964 5. The POSIX interface was broken in that it was just handing over the POSIX
3965 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3966 required a bigger vector, with some working space on the end. This means that
3967 the POSIX wrapper now has to get and free some memory, and copy the results.
3968
3969 6. Added some simple autoconf support, placing the test data and the
3970 documentation in separate directories, re-organizing some of the
3971 information files, and making it build pcre-config (a GNU standard). Also added
3972 libtool support for building PCRE as a shared library, which is now the
3973 default.
3974
3975 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3976 09 are not valid octal constants. Single digits will be used for minor values
3977 less than 10.
3978
3979 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3980 existing programs that set these in the POSIX interface can use PCRE without
3981 modification.
3982
3983 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3984 return all that pcre_info() returns, plus additional data. The pcre_info()
3985 function is retained for compatibility, but is considered to be obsolete.
3986
3987 10. Added experimental recursion feature (?R) to handle one common case that
3988 Perl 5.6 will be able to do with (?p{...}).
3989
3990 11. Added support for POSIX character classes like [:alpha:], which Perl is
3991 adopting.
3992
3993
3994 Version 2.08 31-Aug-99
3995 ----------------------
3996
3997 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3998 trying to match at the startoffset position, but instead was moving forward to
3999 the next newline as if a previous match had failed.
4000
4001 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4002 and could get into a loop if a null string was matched other than at the start
4003 of the subject.
4004
4005 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4006 be distinguished at compile time, and for completeness also added PCRE_DATE.
4007
4008 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4009 in GnuWin32 environments.
4010
4011
4012 Version 2.07 29-Jul-99
4013 ----------------------
4014
4015 1. The documentation is now supplied in plain text form and HTML as well as in
4016 the form of man page sources.
4017
4018 2. C++ compilers don't like assigning (void *) values to other pointer types.
4019 In particular this affects malloc(). Although there is no problem in Standard
4020 C, I've put in casts to keep C++ compilers happy.
4021
4022 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4023 should be (const char *).
4024
4025 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4026 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4027 However, I haven't made this a standard facility. The documentation doesn't
4028 mention it, and the Makefile doesn't support it.
4029
4030 5. The Makefile now contains an "install" target, with editable destinations at
4031 the top of the file. The pcretest program is not installed.
4032
4033 6. pgrep -V now gives the PCRE version number and date.
4034
4035 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4036 causing the entire string to be ignored, instead of just the last character.
4037
4038 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4039 non-matching string, it can take a very, very long time, even for strings of
4040 quite modest length, because of the nested recursion. PCRE now does better in
4041 some of these cases. It does this by remembering the last required literal
4042 character in the pattern, and pre-searching the subject to ensure it is present
4043 before running the real match. In other words, it applies a heuristic to detect
4044 some types of certain failure quickly, and in the above example, if presented
4045 with a string that has no trailing " it gives "no match" very quickly.
4046
4047 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4048 other alternatives are tried instead.
4049
4050
4051 Version 2.06 09-Jun-99
4052 ----------------------
4053
4054 1. Change pcretest's output for amount of store used to show just the code
4055 space, because the remainder (the data block) varies in size between 32-bit and
4056 64-bit systems.
4057
4058 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4059 start matching at. This allows lookbehinds to work when searching for multiple
4060 occurrences in a string.
4061
4062 3. Added additional options to pcretest for testing multiple occurrences:
4063
4064 /+ outputs the rest of the string that follows a match
4065 /g loops for multiple occurrences, using the new startoffset argument
4066 /G loops for multiple occurrences by passing an incremented pointer
4067
4068 4. PCRE wasn't doing the "first character" optimization for patterns starting
4069 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4070 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4071 the letter 'x'. On long subject strings, this gives a significant speed-up.
4072
4073
4074 Version 2.05 21-Apr-99
4075 ----------------------
4076
4077 1. Changed the type of magic_number from int to long int so that it works
4078 properly on 16-bit systems.
4079
4080 2. Fixed a bug which caused patterns starting with .* not to work correctly
4081 when the subject string contained newline characters. PCRE was assuming
4082 anchoring for such patterns in all cases, which is not correct because .* will
4083 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4084 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4085 must be retried after every newline in the subject.
4086
4087
4088 Version 2.04 18-Feb-99
4089 ----------------------
4090
4091 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4092 computation of the store needed to hold the pattern was incorrect (too large).
4093 If such patterns were nested a few deep, this could multiply and become a real
4094 problem.
4095
4096 2. Added /M option to pcretest to show the memory requirement of a specific
4097 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4098
4099 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4100 compiled in such a way that the backtracking after subsequent failure was
4101 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4102 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4103
4104
4105 Version 2.03 02-Feb-99
4106 ----------------------
4107
4108 1. Fixed typo and small mistake in man page.
4109
4110 2. Added 4th condition (GPL supersedes if conflict) and created separate
4111 LICENCE file containing the conditions.
4112
4113 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4114 Perl, that is the internal \ allows the delimiter to be included in the
4115 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4116 the final delimiter, add \ to the end of the pattern (to test the error).
4117
4118 4. Added the convenience functions for extracting substrings after a successful
4119 match. Updated pcretest to make it able to test these functions.
4120
4121
4122 Version 2.02 14-Jan-99
4123 ----------------------
4124
4125 1. Initialized the working variables associated with each extraction so that
4126 their saving and restoring doesn't refer to uninitialized store.
4127
4128 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4129 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4130 fix the problem.
4131
4132 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4133 calls, and wasn't printing the correct value for compiling calls. Increased the
4134 default value of LOOPREPEAT, and the number of significant figures in the
4135 times.
4136
4137 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4138
4139 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4140 a building problem on Windows NT with a FAT file system.
4141
4142
4143 Version 2.01 21-Oct-98
4144 ----------------------
4145
4146 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4147 to character tables built by pcre_maketables() in the current locale. If NULL
4148 is passed, the default tables are used.
4149
4150
4151 Version 2.00 24-Sep-98
4152 ----------------------
4153
4154 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4155 it any more.
4156
4157 2. Allow quantification of (?>) groups, and make it work correctly.
4158
4159 3. The first character computation wasn't working for (?>) groups.
4160
4161 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4162 end of the subject) and add 5.005's \z, which really does match only at the
4163 very end of the subject.
4164
4165 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4166
4167 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4168 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4169 localized options. All options to pcre_study() were also removed.
4170
4171 7. Add other new features from 5.005:
4172
4173 $(?<= positive lookbehind
4174 $(?<! negative lookbehind
4175 (?imsx-imsx) added the unsetting capability
4176 such a setting is global if at outer level; local otherwise
4177 (?imsx-imsx:) non-capturing groups with option setting
4178 (?(cond)re|re) conditional pattern matching
4179
4180 A backreference to itself in a repeated group matches the previous
4181 captured string.
4182
4183 8. General tidying up of studying (both automatic and via "study")
4184 consequential on the addition of new assertions.
4185
4186 9. As in 5.005, unlimited repeated groups that could match an empty substring
4187 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4188 runtime if any iteration does actually match an empty substring.
4189
4190 10. Include the RunTest script in the distribution.
4191
4192 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4193 discrepancies, some of which were old and were also with respect to 5.004. They
4194 have now been fixed.
4195
4196
4197 Version 1.09 28-Apr-98
4198 ----------------------
4199
4200 1. A negated single character class followed by a quantifier with a minimum
4201 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4202 program crashes, or just wrong answers. This did not apply to negated classes
4203 containing more than one character, or to minima other than one.
4204
4205
4206 Version 1.08 27-Mar-98
4207 ----------------------
4208
4209 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4210
4211 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4212 latter must appear before anything that relies on it in the pattern.
4213
4214
4215 Version 1.07 16-Feb-98
4216 ----------------------
4217
4218 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4219 repeat of a potentially empty string).
4220
4221
4222 Version 1.06 23-Jan-98
4223 ----------------------
4224
4225 1. Added Markus Oberhumer's little patches for C++.
4226
4227 2. Literal strings longer than 255 characters were broken.
4228
4229
4230 Version 1.05 23-Dec-97
4231 ----------------------
4232
4233 1. Negated character classes containing more than one character were failing if
4234 PCRE_CASELESS was set at run time.
4235
4236
4237 Version 1.04 19-Dec-97
4238 ----------------------
4239
4240 1. Corrected the man page, where some "const" qualifiers had been omitted.
4241
4242 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4243 input syntax.
4244
4245 3. Fixed memory leak which occurred when a regex with back references was
4246 matched with an offsets vector that wasn't big enough. The temporary memory
4247 that is used in this case wasn't being freed if the match failed.
4248
4249 4. Tidied pcretest to ensure it frees memory that it gets.
4250
4251 5. Temporary memory was being obtained in the case where the passed offsets
4252 vector was exactly big enough.
4253
4254 6. Corrected definition of offsetof() from change 5 below.
4255
4256 7. I had screwed up change 6 below and broken the rules for the use of
4257 setjmp(). Now fixed.
4258
4259
4260 Version 1.03 18-Dec-97
4261 ----------------------
4262
4263 1. A erroneous regex with a missing opening parenthesis was correctly
4264 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4265 on some systems.
4266
4267 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4268 it was reported that one broken compiler failed on the former because "code" is
4269 also an independent variable.
4270
4271 3. The erroneous regex a[]b caused an array overrun reference.
4272
4273 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4274 fail on data ending with that character. (It was going on too far, and checking
4275 the next character, typically a binary zero.) This was specific to the
4276 optimized code for single-character negative classes.
4277
4278 5. Added a contributed patch from the TIN world which does the following:
4279
4280 + Add an undef for memmove, in case the the system defines a macro for it.
4281
4282 + Add a definition of offsetof(), in case there isn't one. (I don't know
4283 the reason behind this - offsetof() is part of the ANSI standard - but
4284 it does no harm).
4285
4286 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4287 most of the places where whitespace preceded '#'. I have given up and
4288 allowed the remaining 2 cases to be at the margin.
4289
4290 + Rename some variables in pcre to eliminate shadowing. This seems very
4291 pedantic, but does no harm, of course.
4292
4293 6. Moved the call to setjmp() into its own function, to get rid of warnings
4294 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4295
4296 7. Constructs such as \d{8,} were compiling into the equivalent of
4297 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4298 outcome, but in this particular case used more store than had been allocated,
4299 which caused the bug to be discovered because it threw up an internal error.
4300
4301 8. The debugging code in both pcre and pcretest for outputting the compiled
4302 form of a regex was going wrong in the case of back references followed by
4303 curly-bracketed repeats.
4304
4305
4306 Version 1.02 12-Dec-97
4307 ----------------------
4308
4309 1. Typos in pcre.3 and comments in the source fixed.
4310
4311 2. Applied a contributed patch to get rid of places where it used to remove
4312 'const' from variables, and fixed some signed/unsigned and uninitialized
4313 variable warnings.
4314
4315 3. Added the "runtest" target to Makefile.
4316
4317 4. Set default compiler flag to -O2 rather than just -O.
4318
4319
4320 Version 1.01 19-Nov-97
4321 ----------------------
4322
4323 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4324 like /([ab]*)*/, that is, for classes with more than one character in them.
4325
4326 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4327 as /((?>a*))*/ (a PCRE_EXTRA facility).
4328
4329
4330 Version 1.00 18-Nov-97
4331 ----------------------
4332
4333 1. Added compile-time macros to support systems such as SunOS4 which don't have
4334 memmove() or strerror() but have other things that can be used instead.
4335
4336 2. Arranged that "make clean" removes the executables.
4337
4338
4339 Version 0.99 27-Oct-97
4340 ----------------------
4341
4342 1. Fixed bug in code for optimizing classes with only one character. It was
4343 initializing a 32-byte map regardless, which could cause it to run off the end
4344 of the memory it had got.
4345
4346 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4347
4348
4349 Version 0.98 22-Oct-97
4350 ----------------------
4351
4352 1. Fixed bug in code for handling temporary memory usage when there are more
4353 back references than supplied space in the ovector. This could cause segfaults.
4354
4355
4356 Version 0.97 21-Oct-97
4357 ----------------------
4358
4359 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4360
4361 2. Optimized negated single characters not to use a bit map.
4362
4363 3. Brought error texts together as macro definitions; clarified some of them;
4364 fixed one that was wrong - it said "range out of order" when it meant "invalid
4365 escape sequence".
4366
4367 4. Changed some char * arguments to const char *.
4368
4369 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4370
4371 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4372 pcretest.
4373
4374
4375 Version 0.96 16-Oct-97
4376 ----------------------
4377
4378 1. Added a simple "pgrep" utility to the distribution.
4379
4380 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4381 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4382 where "ddd" means "one or more decimal digits".
4383
4384 3. Fixed serious bug. If a pattern had a back reference, but the call to
4385 pcre_exec() didn't supply a large enough ovector to record the related
4386 identifying subpattern, the match always failed. PCRE now remembers the number
4387 of the largest back reference, and gets some temporary memory in which to save
4388 the offsets during matching if necessary, in order to ensure that
4389 backreferences always work.
4390
4391 4. Increased the compatibility with Perl in a number of ways:
4392
4393 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4394 to request this handling. The option can be set at compile or exec time.
4395
4396 (b) $ matches before a terminating newline by default; an option
4397 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4398 mode). The option can be set at compile or exec time.
4399
4400 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4401 the same as Perl's. If the decimal number it represents is less than 10
4402 or there aren't that many previous left capturing parentheses, an octal
4403 escape is read. Inside a character class, it's always an octal escape,
4404 even if it is a single digit.
4405
4406 (d) An escaped but undefined alphabetic character is taken as a literal,
4407 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4408 escapes.
4409
4410 (e) {0} is now permitted. (The previous item is removed from the compiled
4411 pattern).
4412
4413 5. Changed all the names of code files so that the basic parts are no longer
4414 than 10 characters, and abolished the teeny "globals.c" file.
4415
4416 6. Changed the handling of character classes; they are now done with a 32-byte
4417 bit map always.
4418
4419 7. Added the -d and /D options to pcretest to make it possible to look at the
4420 internals of compilation without having to recompile pcre.
4421
4422
4423 Version 0.95 23-Sep-97
4424 ----------------------
4425
4426 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4427 \x20 at the start of a run of normal characters. These were being treated as
4428 real characters, instead of the source characters being re-checked.
4429
4430
4431 Version 0.94 18-Sep-97
4432 ----------------------
4433
4434 1. The functions are now thread-safe, with the caveat that the global variables
4435 containing pointers to malloc() and free() or alternative functions are the
4436 same for all threads.
4437
4438 2. Get pcre_study() to generate a bitmap of initial characters for non-
4439 anchored patterns when this is possible, and use it if passed to pcre_exec().
4440
4441
4442 Version 0.93 15-Sep-97
4443 ----------------------
4444
4445 1. /(b)|(:+)/ was computing an incorrect first character.
4446
4447 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4448 but not actually doing anything yet.
4449
4450 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4451 as Perl does (e.g. [-az] or [az-]).
4452
4453 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4454 all possible positions.
4455
4456 5. Split up into different modules to avoid including unneeded functions in a
4457 compiled binary. However, compile and exec are still in one module. The "study"
4458 function is split off.
4459
4460 6. The character tables are now in a separate module whose source is generated
4461 by an auxiliary program - but can then be edited by hand if required. There are
4462 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4463 toupper() in the code.
4464
4465 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4466 make them global. Abolish the function for setting them, as the caller can now
4467 set them directly.
4468
4469
4470 Version 0.92 11-Sep-97
4471 ----------------------
4472
4473 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4474 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4475
4476 2. Caseless matching was not working in character classes if the characters in
4477 the pattern were in upper case.
4478
4479 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4480
4481 4. Make PCRE_ANCHORED public and accept as a compile option.
4482
4483 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4484 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4485 pass them.
4486
4487 6. Give an error if bad option bits passed at compile or run time.
4488
4489 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4490 pcretest to cause it to pass that flag.
4491
4492 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4493 options, and the first character, if set.
4494
4495 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4496
4497
4498 Version 0.91 10-Sep-97
4499 ----------------------
4500
4501 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4502 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4503
4504 2. PCRE was looping on encountering an indefinitely repeated back reference to
4505 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4506 Perl does - treats the match as successful.
4507
4508 ****

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