/[pcre]/code/branches/pcre16/ChangeLog
ViewVC logotype

Contents of /code/branches/pcre16/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 723 - (show annotations)
Sat Oct 8 15:55:23 2011 UTC (8 years, 1 month ago) by ph10
Original Path: code/trunk/ChangeLog
File size: 216825 byte(s)
Revert handling of atomic groups that do not include captures to the old way of 
handling, thus reducing stack usage.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.20 23-Sep-2011
5 ------------------------
6
7 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
8 a POSIX class. After further experiments with Perl, which convinced me that
9 Perl has bugs and confusions, a closing square bracket is no longer allowed
10 in a POSIX name. This bug also affected patterns with classes that started
11 with full stops.
12
13 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
14 captured substring, but while checking the failing first alternative,
15 substring 1 is temporarily captured. If the output vector supplied to
16 pcre_exec() was not big enough for this capture, the yield of the function
17 was still zero ("insufficient space for captured substrings"). This cannot
18 be totally fixed without adding another stack variable, which seems a lot
19 of expense for a edge case. However, I have improved the situation in cases
20 such as /(a)(b)x|abc/ matched against "abc", where the return code
21 indicates that fewer than the maximum number of slots in the ovector have
22 been set.
23
24 3. Related to (2) above: when there are more back references in a pattern than
25 slots in the output vector, pcre_exec() uses temporary memory during
26 matching, and copies in the captures as far as possible afterwards. It was
27 using the entire output vector, but this conflicts with the specification
28 that only 2/3 is used for passing back captured substrings. Now it uses
29 only the first 2/3, for compatibility. This is, of course, another edge
30 case.
31
32 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
33 main code base, and can be used by building with --enable-jit. When this is
34 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
35 runtime --no-jit option is given.
36
37 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
38 ovector, the return from the function was zero, implying that there were
39 other matches that did not fit. The correct "exactly full" value is now
40 returned.
41
42 6. If a subpattern that was called recursively or as a subroutine contained
43 (*PRUNE) or any other control that caused it to give a non-standard return,
44 invalid errors such as "Error -26 (nested recursion at the same subject
45 position)" or even infinite loops could occur.
46
47 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
48 computing the minimum length on reaching *ACCEPT, and so ended up with the
49 wrong value of 1 rather than 0. Further investigation indicates that
50 computing a minimum subject length in the presence of *ACCEPT is difficult
51 (think back references, subroutine calls), and so I have changed the code
52 so that no minimum is registered for a pattern that contains *ACCEPT.
53
54 8. If (*THEN) was present in the first (true) branch of a conditional group,
55 it was not handled as intended. [But see 16 below.]
56
57 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
58 Sheri Pierce.
59
60 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
61 the first byte in a match must be "a".
62
63 11. Change 17 for 8.13 increased the recursion depth for patterns like
64 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
65 pattern contains any instances of (*THEN). If it does not, the old
66 optimizations are restored. It would be nice to do this on a per-group
67 basis, but at the moment that is not feasible.
68
69 12. In some environments, the output of pcretest -C is CRLF terminated. This
70 broke RunTest's code that checks for the link size. A single white space
71 after the value is now allowed for.
72
73 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
74 For "fr", it uses the Windows-specific input and output files.
75
76 14. If (*THEN) appeared in a group that was called recursively or as a
77 subroutine, it did not work as intended. [But see next item.]
78
79 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
80 pattern fragments (but not containing any | characters). If A and B are
81 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
82 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
83 D. In other words, Perl considers parentheses that do not contain any |
84 characters to be part of a surrounding alternative, whereas PCRE was
85 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
86 differently. PCRE now behaves in the same way as Perl, except in the case
87 of subroutine/recursion calls such as (?1) which have in any case always
88 been different (but PCRE had them first :-).
89
90 16. Related to 15 above: Perl does not treat the | in a conditional group as
91 creating alternatives. Such a group is treated in the same way as an
92 ordinary group without any | characters when processing (*THEN). PCRE has
93 been changed to match Perl's behaviour.
94
95 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
96 RunGrepTest script failed.
97
98 18. Change 22 for version 13 caused atomic groups to use more stack. This is
99 inevitable for groups that contain captures, but it can lead to a lot of
100 stack use in large patterns. The old behaviour has been restored for atomic
101 groups that do not contain any capturing parentheses.
102
103
104 Version 8.13 16-Aug-2011
105 ------------------------
106
107 1. The Unicode data tables have been updated to Unicode 6.0.0.
108
109 2. Two minor typos in pcre_internal.h have been fixed.
110
111 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
112 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
113 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
114
115 4. There were a number of related bugs in the code for matching backrefences
116 caselessly in UTF-8 mode when codes for the characters concerned were
117 different numbers of bytes. For example, U+023A and U+2C65 are an upper
118 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
119 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
120 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
121 2-byte code at the end of the subject (it thought there wasn't enough data
122 left).
123
124 5. Comprehensive information about what went wrong is now returned by
125 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
126 as the output vector has at least 2 elements. The offset of the start of
127 the failing character and a reason code are placed in the vector.
128
129 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
130 now returned is for the first byte of the failing character, instead of the
131 last byte inspected. This is an incompatible change, but I hope it is small
132 enough not to be a problem. It makes the returned offset consistent with
133 pcre_exec() and pcre_dfa_exec().
134
135 7. pcretest now gives a text phrase as well as the error number when
136 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
137 failure, the offset and reason code are output.
138
139 8. When \R was used with a maximizing quantifier it failed to skip backwards
140 over a \r\n pair if the subsequent match failed. Instead, it just skipped
141 back over a single character (\n). This seems wrong (because it treated the
142 two characters as a single entity when going forwards), conflicts with the
143 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
144 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
145 has been changed.
146
147 9. Some internal refactoring has changed the processing so that the handling
148 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
149 time (the PCRE_DOTALL option was changed this way some time ago: version
150 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
151 which was always a bit of a fudge. It also means that there is one less
152 argument for the match() function, which reduces its stack requirements
153 slightly. This change also fixes an incompatibility with Perl: the pattern
154 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
155
156 10. More internal refactoring has drastically reduced the number of recursive
157 calls to match() for possessively repeated groups such as (abc)++ when
158 using pcre_exec().
159
160 11. While implementing 10, a number of bugs in the handling of groups were
161 discovered and fixed:
162
163 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
164 (a|)*(?1) gave a compile-time internal error.
165 ((a|)+)+ did not notice that the outer group could match an empty string.
166 (^a|^)+ was not marked as anchored.
167 (.*a|.*)+ was not marked as matching at start or after a newline.
168
169 12. Yet more internal refactoring has removed another argument from the match()
170 function. Special calls to this function are now indicated by setting a
171 value in a variable in the "match data" data block.
172
173 13. Be more explicit in pcre_study() instead of relying on "default" for
174 opcodes that mean there is no starting character; this means that when new
175 ones are added and accidentally left out of pcre_study(), testing should
176 pick them up.
177
178 14. The -s option of pcretest has been documented for ages as being an old
179 synonym of -m (show memory usage). I have changed it to mean "force study
180 for every regex", that is, assume /S for every regex. This is similar to -i
181 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
182 using it. It makes it easier to run collections of tests with and without
183 study enabled, and thereby test pcre_study() more easily. All the standard
184 tests are now run with and without -s (but some patterns can be marked as
185 "never study" - see 20 below).
186
187 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
188 restoration of the capturing data to the outer values was not happening
189 correctly.
190
191 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
192 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
193 pattern had matched an empty string, and so incorrectly returned a no
194 match.
195
196 17. There was optimizing code for the last branch of non-capturing parentheses,
197 and also for the obeyed branch of a conditional subexpression, which used
198 tail recursion to cut down on stack usage. Unfortunately, now that there is
199 the possibility of (*THEN) occurring in these branches, tail recursion is
200 no longer possible because the return has to be checked for (*THEN). These
201 two optimizations have therefore been removed. [But see 8.20/11 above.]
202
203 18. If a pattern containing \R was studied, it was assumed that \R always
204 matched two bytes, thus causing the minimum subject length to be
205 incorrectly computed because \R can also match just one byte.
206
207 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
208 was incorrectly computed.
209
210 20. If /S is present twice on a test pattern in pcretest input, it now
211 *disables* studying, thereby overriding the use of -s on the command line
212 (see 14 above). This is necessary for one or two tests to keep the output
213 identical in both cases.
214
215 21. When (*ACCEPT) was used in an assertion that matched an empty string and
216 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
217
218 22. When an atomic group that contained a capturing parenthesis was
219 successfully matched, but the branch in which it appeared failed, the
220 capturing was not being forgotten if a higher numbered group was later
221 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
222 group 1 to "a", when in fact it should be unset. This applied to multi-
223 branched capturing and non-capturing groups, repeated or not, and also to
224 positive assertions (capturing in negative assertions does not happen
225 in PCRE) and also to nested atomic groups.
226
227 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
228 subject after a captured substring, to make it easier to tell which of a
229 number of identical substrings has been captured.
230
231 24. The way atomic groups are processed by pcre_exec() has been changed so that
232 if they are repeated, backtracking one repetition now resets captured
233 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
234 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
235 "aaa". Previously, it would have been "a". As part of this code
236 refactoring, the way recursive calls are handled has also been changed.
237
238 25. If an assertion condition captured any substrings, they were not passed
239 back unless some other capturing happened later. For example, if
240 (?(?=(a))a) was matched against "a", no capturing was returned.
241
242 26. When studying a pattern that contained subroutine calls or assertions,
243 the code for finding the minimum length of a possible match was handling
244 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
245 group 1 called group 2 while simultaneously a separate group 2 called group
246 1). A stack overflow occurred in this case. I have fixed this by limiting
247 the recursion depth to 10.
248
249 27. Updated RunTest.bat in the distribution to the version supplied by Tom
250 Fortmann. This supports explicit test numbers on the command line, and has
251 argument validation and error reporting.
252
253 28. An instance of \X with an unlimited repeat could fail if at any point the
254 first character it looked at was a mark character.
255
256 29. Some minor code refactoring concerning Unicode properties and scripts
257 should reduce the stack requirement of match() slightly.
258
259 30. Added the '=' option to pcretest to check the setting of unused capturing
260 slots at the end of the pattern, which are documented as being -1, but are
261 not included in the return count.
262
263 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
264 compiled something random. Now it gives a compile-time error (as does
265 Perl).
266
267 32. A *MARK encountered during the processing of a positive assertion is now
268 recorded and passed back (compatible with Perl).
269
270 33. If --only-matching or --colour was set on a pcregrep call whose pattern
271 had alternative anchored branches, the search for a second match in a line
272 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
273 matched the line "0102" twice. The same bug affected patterns that started
274 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
275 twice.
276
277 34. Previously, PCRE did not allow quantification of assertions. However, Perl
278 does, and because of capturing effects, quantifying parenthesized
279 assertions may at times be useful. Quantifiers are now allowed for
280 parenthesized assertions.
281
282 35. A minor code tidy in pcre_compile() when checking options for \R usage.
283
284 36. \g was being checked for fancy things in a character class, when it should
285 just be a literal "g".
286
287 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
288 appearance of a nested POSIX class supersedes an apparent external class.
289 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
290 unescaped square brackets may also appear as part of class names. For
291 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
292 more like Perl. (But see 8.20/1 above.)
293
294 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
295 was because it thought it was \N{name}, which is not supported).
296
297 39. Add minix to OS list not supporting the -S option in pcretest.
298
299 40. PCRE tries to detect cases of infinite recursion at compile time, but it
300 cannot analyze patterns in sufficient detail to catch mutual recursions
301 such as ((?1))((?2)). There is now a runtime test that gives an error if a
302 subgroup is called recursively as a subpattern for a second time at the
303 same position in the subject string. In previous releases this might have
304 been caught by the recursion limit, or it might have run out of stack.
305
306 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
307 happen only once. PCRE was, however incorrectly giving a compile time error
308 "recursive call could loop indefinitely" because it cannot analyze the
309 pattern in sufficient detail. The compile time test no longer happens when
310 PCRE is compiling a conditional subpattern, but actual runaway loops are
311 now caught at runtime (see 40 above).
312
313 42. It seems that Perl allows any characters other than a closing parenthesis
314 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
315 has been changed to be the same.
316
317 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
318 as not to get warnings when autogen.sh is called. Also changed
319 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
320
321 44. To help people who use pcregrep to scan files containing exceedingly long
322 lines, the following changes have been made:
323
324 (a) The default value of the buffer size parameter has been increased from
325 8K to 20K. (The actual buffer used is three times this size.)
326
327 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
328 PCRE is built.
329
330 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
331 to be set at run time.
332
333 (d) Numerical values in pcregrep options can be followed by K or M, for
334 example --buffer-size=50K.
335
336 (e) If a line being scanned overflows pcregrep's buffer, an error is now
337 given and the return code is set to 2.
338
339 45. Add a pointer to the latest mark to the callout data block.
340
341 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
342 partial match of an empty string instead of no match. This was specific to
343 the use of ".".
344
345 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
346 complete match instead of a partial match. This bug was dependent on both
347 the PCRE_UTF8 and PCRE_DOTALL options being set.
348
349 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
350 starting byte set, because \b was not being ignored.
351
352
353 Version 8.12 15-Jan-2011
354 ------------------------
355
356 1. Fixed some typos in the markup of the man pages, and wrote a script that
357 checks for such things as part of the documentation building process.
358
359 2. On a big-endian 64-bit system, pcregrep did not correctly process the
360 --match-limit and --recursion-limit options (added for 8.11). In
361 particular, this made one of the standard tests fail. (The integer value
362 went into the wrong half of a long int.)
363
364 3. If the --colour option was given to pcregrep with -v (invert match), it
365 did strange things, either producing crazy output, or crashing. It should,
366 of course, ignore a request for colour when reporting lines that do not
367 match.
368
369 4. Another pcregrep bug caused similar problems if --colour was specified with
370 -M (multiline) and the pattern match finished with a line ending.
371
372 5. In pcregrep, when a pattern that ended with a literal newline sequence was
373 matched in multiline mode, the following line was shown as part of the
374 match. This seems wrong, so I have changed it.
375
376 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
377 the check for further matches in the same line (so they could be coloured)
378 to overrun the end of the current line. If another match was found, it was
379 incorrectly shown (and then shown again when found in the next line).
380
381 7. If pcregrep was compiled under Windows, there was a reference to the
382 function pcregrep_exit() before it was defined. I am assuming this was
383 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
384 reported by a user. I've moved the definition above the reference.
385
386
387 Version 8.11 10-Dec-2010
388 ------------------------
389
390 1. (*THEN) was not working properly if there were untried alternatives prior
391 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
392 backtracked to try for "b" instead of moving to the next alternative branch
393 at the same level (in this case, to look for "c"). The Perl documentation
394 is clear that when (*THEN) is backtracked onto, it goes to the "next
395 alternative in the innermost enclosing group".
396
397 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
398 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
399 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
400 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
401 (*THEN).
402
403 3. If \s appeared in a character class, it removed the VT character from
404 the class, even if it had been included by some previous item, for example
405 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
406 of \s, but is part of the POSIX "space" class.)
407
408 4. A partial match never returns an empty string (because you can always
409 match an empty string at the end of the subject); however the checking for
410 an empty string was starting at the "start of match" point. This has been
411 changed to the "earliest inspected character" point, because the returned
412 data for a partial match starts at this character. This means that, for
413 example, /(?<=abc)def/ gives a partial match for the subject "abc"
414 (previously it gave "no match").
415
416 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
417 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
418 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
419 has an implication that the given string is incomplete (because a partial
420 match is preferred over a full match). For this reason, these items now
421 give a partial match in this situation. [Aside: previously, the one case
422 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
423 match rather than a full match, which was wrong by the old rules, but is
424 now correct.]
425
426 6. There was a bug in the handling of #-introduced comments, recognized when
427 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
428 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
429 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
430 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
431 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
432 places in pcre_compile().
433
434 7. Related to (6) above, when pcre_compile() was skipping #-introduced
435 comments when looking ahead for named forward references to subpatterns,
436 the only newline sequence it recognized was NL. It now handles newlines
437 according to the set newline convention.
438
439 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
440 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
441 cater for a lack of strerror(). These oversights have been fixed.
442
443 9. Added --match-limit and --recursion-limit to pcregrep.
444
445 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
446
447 11. When the -o option was used, pcregrep was setting a return code of 1, even
448 when matches were found, and --line-buffered was not being honoured.
449
450 12. Added an optional parentheses number to the -o and --only-matching options
451 of pcregrep.
452
453 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
454 can match an empty string. The code to do it in pcretest and pcredemo
455 needed fixing:
456
457 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
458 only one byte after an empty string match just before CRLF (this case
459 just got forgotten; "any" and "anycrlf" were OK).
460
461 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
462 mode when an empty string match preceded an ASCII character followed by
463 a non-ASCII character. (The code for advancing by one character rather
464 than one byte was nonsense.)
465
466 (c) The pcredemo.c sample program did not have any code at all to handle
467 the cases when CRLF is a valid newline sequence.
468
469 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
470 as a starting offset was within the subject string. There is now a new
471 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
472 negative or greater than the length of the string. In order to test this,
473 pcretest is extended to allow the setting of negative starting offsets.
474
475 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
476 starting offset points to the beginning of a UTF-8 character was
477 unnecessarily clumsy. I tidied it up.
478
479 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
480 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
481
482 17. Nobody had reported that the --include_dir option, which was added in
483 release 7.7 should have been called --include-dir (hyphen, not underscore)
484 for compatibility with GNU grep. I have changed it to --include-dir, but
485 left --include_dir as an undocumented synonym, and the same for
486 --exclude-dir, though that is not available in GNU grep, at least as of
487 release 2.5.4.
488
489 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
490 characters from a string of bytes) have been redefined so as not to use
491 loops, in order to improve performance in some environments. At the same
492 time, I abstracted some of the common code into auxiliary macros to save
493 repetition (this should not affect the compiled code).
494
495 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
496 compile-time error is now given if \c is not followed by an ASCII
497 character, that is, a byte less than 128. (In EBCDIC mode, the code is
498 different, and any byte value is allowed.)
499
500 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
501 START_OPTIMIZE option, which is now allowed at compile time - but just
502 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
503 to pcregrep and other applications that have no direct access to PCRE
504 options. The new /Y option in pcretest sets this option when calling
505 pcre_compile().
506
507 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
508 back references. Groups containing recursive back references were forced to
509 be atomic by that change, but in the case of named groups, the amount of
510 memory required was incorrectly computed, leading to "Failed: internal
511 error: code overflow". This has been fixed.
512
513 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
514 pcretest.c, to avoid build problems in some Borland environments.
515
516
517 Version 8.10 25-Jun-2010
518 ------------------------
519
520 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
521 THEN.
522
523 2. (*ACCEPT) was not working when inside an atomic group.
524
525 3. Inside a character class, \B is treated as a literal by default, but
526 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
527 causes the error). The code is unchanged, but I tidied the documentation.
528
529 4. Inside a character class, PCRE always treated \R and \X as literals,
530 whereas Perl faults them if its -w option is set. I have changed PCRE so
531 that it faults them when PCRE_EXTRA is set.
532
533 5. Added support for \N, which always matches any character other than
534 newline. (It is the same as "." when PCRE_DOTALL is not set.)
535
536 6. When compiling pcregrep with newer versions of gcc which may have
537 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
538 declared with attribute warn_unused_result" were given. Just casting the
539 result to (void) does not stop the warnings; a more elaborate fudge is
540 needed. I've used a macro to implement this.
541
542 7. Minor change to pcretest.c to avoid a compiler warning.
543
544 8. Added four artifical Unicode properties to help with an option to make
545 \s etc use properties (see next item). The new properties are: Xan
546 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
547
548 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
549 use Unicode properties. (*UCP) at the start of a pattern can be used to set
550 this option. Modified pcretest to add /W to test this facility. Added
551 REG_UCP to make it available via the POSIX interface.
552
553 10. Added --line-buffered to pcregrep.
554
555 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
556 studied, and the match started with a letter with a code point greater than
557 127 whose first byte was different to the first byte of the other case of
558 the letter, the other case of this starting letter was not recognized
559 (#976).
560
561 12. If a pattern that was studied started with a repeated Unicode property
562 test, for example, \p{Nd}+, there was the theoretical possibility of
563 setting up an incorrect bitmap of starting bytes, but fortunately it could
564 not have actually happened in practice until change 8 above was made (it
565 added property types that matched character-matching opcodes).
566
567 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
568 possible starting bytes for non-anchored patterns.
569
570 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
571 \R, and also a number of cases that involve Unicode properties, both
572 explicit and implicit when PCRE_UCP is set.
573
574 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
575 input, it could crash or give wrong results if characters with values
576 greater than 0xc0 were present in the subject string. (Detail: it assumed
577 UTF-8 input when processing these items.)
578
579 16. Added a lot of (int) casts to avoid compiler warnings in systems where
580 size_t is 64-bit (#991).
581
582 17. Added a check for running out of memory when PCRE is compiled with
583 --disable-stack-for-recursion (#990).
584
585 18. If the last data line in a file for pcretest does not have a newline on
586 the end, a newline was missing in the output.
587
588 19. The default pcre_chartables.c file recognizes only ASCII characters (values
589 less than 128) in its various bitmaps. However, there is a facility for
590 generating tables according to the current locale when PCRE is compiled. It
591 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
592 characters, are recognized by isspace() and therefore were getting set in
593 these tables, and indeed these tables seem to approximate to ISO 8859. This
594 caused a problem in UTF-8 mode when pcre_study() was used to create a list
595 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
596 which of course cannot start UTF-8 characters. I have changed the code so
597 that only real ASCII characters (less than 128) and the correct starting
598 bytes for UTF-8 encodings are set for characters greater than 127 when in
599 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
600 altogether.)
601
602 20. Added the /T option to pcretest so as to be able to run tests with non-
603 standard character tables, thus making it possible to include the tests
604 used for 19 above in the standard set of tests.
605
606 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
607 reference to a subpattern the other side of a comment that contains an
608 opening parenthesis caused either an internal compiling error, or a
609 reference to the wrong subpattern.
610
611
612 Version 8.02 19-Mar-2010
613 ------------------------
614
615 1. The Unicode data tables have been updated to Unicode 5.2.0.
616
617 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
618 configured.
619
620 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
621 original author of that file, following a query about its status.
622
623 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
624 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
625
626 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
627 quantifier applied to a forward-referencing subroutine call, could compile
628 incorrect code or give the error "internal error: previously-checked
629 referenced subpattern not found".
630
631 6. Both MS Visual Studio and Symbian OS have problems with initializing
632 variables to point to external functions. For these systems, therefore,
633 pcre_malloc etc. are now initialized to local functions that call the
634 relevant global functions.
635
636 7. There were two entries missing in the vectors called coptable and poptable
637 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
638 I've fixed the data, and added a kludgy way of testing at compile time that
639 the lengths are correct (equal to the number of opcodes).
640
641 8. Following on from 7, I added a similar kludge to check the length of the
642 eint vector in pcreposix.c.
643
644 9. Error texts for pcre_compile() are held as one long string to avoid too
645 much relocation at load time. To find a text, the string is searched,
646 counting zeros. There was no check for running off the end of the string,
647 which could happen if a new error number was added without updating the
648 string.
649
650 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
651
652 11. \K was not working if it appeared in an atomic group or in a group that
653 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
654 \K is "not well defined" if used in an assertion. PCRE now accepts it if
655 the assertion is positive, but not if it is negative.
656
657 12. Change 11 fortuitously reduced the size of the stack frame used in the
658 "match()" function of pcre_exec.c by one pointer. Forthcoming
659 implementation of support for (*MARK) will need an extra pointer on the
660 stack; I have reserved it now, so that the stack frame size does not
661 decrease.
662
663 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
664 item in branch that calls a recursion is a subroutine call - as in the
665 second branch in the above example - was incorrectly given the compile-
666 time error "recursive call could loop indefinitely" because pcre_compile()
667 was not correctly checking the subroutine for matching a non-empty string.
668
669 14. The checks for overrunning compiling workspace could trigger after an
670 overrun had occurred. This is a "should never occur" error, but it can be
671 triggered by pathological patterns such as hundreds of nested parentheses.
672 The checks now trigger 100 bytes before the end of the workspace.
673
674 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
675
676
677 Version 8.01 19-Jan-2010
678 ------------------------
679
680 1. If a pattern contained a conditional subpattern with only one branch (in
681 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
682 computed the wrong minimum data length (which is of course zero for such
683 subpatterns). This could cause incorrect "no match" results.
684
685 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
686 the pattern is reset in the first branch, pcre_compile() failed with
687 "internal error: code overflow at offset...". This happened only when
688 the reset was to the original external option setting. (An optimization
689 abstracts leading options settings into an external setting, which was the
690 cause of this.)
691
692 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
693 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
694 assertion pattern did not match (meaning that the assertion was true), it
695 was incorrectly treated as false if the SKIP had been reached during the
696 matching. This also applied to assertions used as conditions.
697
698 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
699 assertion subpattern, including such a pattern used as a condition,
700 unpredictable results occurred, instead of the error return
701 PCRE_ERROR_DFA_UITEM.
702
703 5. The C++ GlobalReplace function was not working like Perl for the special
704 situation when an empty string is matched. It now does the fancy magic
705 stuff that is necessary.
706
707 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
708 removed. (These were left over from very, very early versions of PCRE.)
709
710 7. Some cosmetic changes to the code to make life easier when compiling it
711 as part of something else:
712
713 (a) Change DEBUG to PCRE_DEBUG.
714
715 (b) In pcre_compile(), rename the member of the "branch_chain" structure
716 called "current" as "current_branch", to prevent a collision with the
717 Linux macro when compiled as a kernel module.
718
719 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
720 prevent a collision with the Linux macro when compiled as a kernel
721 module.
722
723 8. In pcre_compile() there are some checks for integer overflows that used to
724 cast potentially large values to (double). This has been changed to that
725 when building, a check for int64_t is made, and if it is found, it is used
726 instead, thus avoiding the use of floating point arithmetic. (There is no
727 other use of FP in PCRE.) If int64_t is not found, the fallback is to
728 double.
729
730 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
731 2005 (difference between two addresses compared to an unsigned value).
732
733 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
734 custom one, because of the following reported problem in Windows:
735
736 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
737 under Win32.
738 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
739 therefore missing the function definition.
740 - The compiler thus generates a "C" signature for the test function.
741 - The linker fails to find the "C" function.
742 - PCRE fails to configure if asked to do so against libbz2.
743
744 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
745 messages were output:
746
747 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
748 rerunning libtoolize, to keep the correct libtool macros in-tree.
749 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
750
751 I have done both of these things.
752
753 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
754 most of the time, it *can* run out if it is given a pattern that contains a
755 runaway infinite recursion. I updated the discussion in the pcrestack man
756 page.
757
758 13. Now that we have gone to the x.xx style of version numbers, the minor
759 version may start with zero. Using 08 or 09 is a bad idea because users
760 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
761 interpreted as invalid octal numbers. I've updated the previous comment in
762 configure.ac, and also added a check that gives an error if 08 or 09 are
763 used.
764
765 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
766 causing partial matching to fail when the end of the subject matched \W
767 in a UTF-8 pattern where \W was quantified with a minimum of 3.
768
769 15. There were some discrepancies between the declarations in pcre_internal.h
770 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
771 their definitions. The declarations used "const uschar *" and the
772 definitions used USPTR. Even though USPTR is normally defined as "const
773 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
774 reported that: "This difference in casting confuses some C++ compilers, for
775 example, SunCC recognizes above declarations as different functions and
776 generates broken code for hbpcre." I have changed the declarations to use
777 USPTR.
778
779 16. GNU libtool is named differently on some systems. The autogen.sh script now
780 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
781 (FreeBSD).
782
783 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
784 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
785 comment: "Figure out how to create a longlong from a string: strtoll and
786 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
787 instance, but it only takes 2 args instead of 3!"
788
789 18. A subtle bug concerned with back references has been fixed by a change of
790 specification, with a corresponding code fix. A pattern such as
791 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
792 refers, was giving matches when it shouldn't. For example, xa=xaaa would
793 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
794 same bug. Such groups have to be quantified to be useful, or contained
795 inside another quantified group. (If there's no repetition, the reference
796 can never match.) The problem arises because, having left the group and
797 moved on to the rest of the pattern, a later failure that backtracks into
798 the group uses the captured value from the final iteration of the group
799 rather than the correct earlier one. I have fixed this in PCRE by forcing
800 any group that contains a reference to itself to be an atomic group; that
801 is, there cannot be any backtracking into it once it has completed. This is
802 similar to recursive and subroutine calls.
803
804
805 Version 8.00 19-Oct-09
806 ----------------------
807
808 1. The table for translating pcre_compile() error codes into POSIX error codes
809 was out-of-date, and there was no check on the pcre_compile() error code
810 being within the table. This could lead to an OK return being given in
811 error.
812
813 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
814 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
815 in a Windows environment.
816
817 3. The pcregrep --count option prints the count for each file even when it is
818 zero, as does GNU grep. However, pcregrep was also printing all files when
819 --files-with-matches was added. Now, when both options are given, it prints
820 counts only for those files that have at least one match. (GNU grep just
821 prints the file name in this circumstance, but including the count seems
822 more useful - otherwise, why use --count?) Also ensured that the
823 combination -clh just lists non-zero counts, with no names.
824
825 4. The long form of the pcregrep -F option was incorrectly implemented as
826 --fixed_strings instead of --fixed-strings. This is an incompatible change,
827 but it seems right to fix it, and I didn't think it was worth preserving
828 the old behaviour.
829
830 5. The command line items --regex=pattern and --regexp=pattern were not
831 recognized by pcregrep, which required --regex pattern or --regexp pattern
832 (with a space rather than an '='). The man page documented the '=' forms,
833 which are compatible with GNU grep; these now work.
834
835 6. No libpcreposix.pc file was created for pkg-config; there was just
836 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
837
838 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
839 when UCP support is not needed, by modifying the Python script that
840 generates it from Unicode data files. This should not matter if the module
841 is correctly used as a library, but I received one complaint about 50K of
842 unwanted data. My guess is that the person linked everything into his
843 program rather than using a library. Anyway, it does no harm.
844
845 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
846 was a minimum greater than 1 for a wide character in a possessive
847 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
848 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
849 character. Chaos in the form of incorrect output or a compiling loop could
850 result.
851
852 9. The restrictions on what a pattern can contain when partial matching is
853 requested for pcre_exec() have been removed. All patterns can now be
854 partially matched by this function. In addition, if there are at least two
855 slots in the offset vector, the offset of the earliest inspected character
856 for the match and the offset of the end of the subject are set in them when
857 PCRE_ERROR_PARTIAL is returned.
858
859 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
860 synonymous with PCRE_PARTIAL, for backwards compatibility, and
861 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
862 and may be more useful for multi-segment matching.
863
864 11. Partial matching with pcre_exec() is now more intuitive. A partial match
865 used to be given if ever the end of the subject was reached; now it is
866 given only if matching could not proceed because another character was
867 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
868 string "Z", which now yields "no match" instead of "partial match". In the
869 case of pcre_dfa_exec(), "no match" is given if every matching path for the
870 final character ended with (*FAIL).
871
872 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
873 if the pattern had a "must contain" character that was already found in the
874 earlier partial match, unless partial matching was again requested. For
875 example, with the pattern /dog.(body)?/, the "must contain" character is
876 "g". If the first part-match was for the string "dog", restarting with
877 "sbody" failed. This bug has been fixed.
878
879 13. The string returned by pcre_dfa_exec() after a partial match has been
880 changed so that it starts at the first inspected character rather than the
881 first character of the match. This makes a difference only if the pattern
882 starts with a lookbehind assertion or \b or \B (\K is not supported by
883 pcre_dfa_exec()). It's an incompatible change, but it makes the two
884 matching functions compatible, and I think it's the right thing to do.
885
886 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
887 so that the demonstration program is easily available in environments where
888 PCRE has not been installed from source.
889
890 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
891 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
892 library.
893
894 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
895 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
896 is not the first non-POSIX option to be added. Clearly some people find
897 these options useful.
898
899 17. If a caller to the POSIX matching function regexec() passes a non-zero
900 value for nmatch with a NULL value for pmatch, the value of
901 nmatch is forced to zero.
902
903 18. RunGrepTest did not have a test for the availability of the -u option of
904 the diff command, as RunTest does. It now checks in the same way as
905 RunTest, and also checks for the -b option.
906
907 19. If an odd number of negated classes containing just a single character
908 interposed, within parentheses, between a forward reference to a named
909 subpattern and the definition of the subpattern, compilation crashed with
910 an internal error, complaining that it could not find the referenced
911 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
912 [The bug was that it was starting one character too far in when skipping
913 over the character class, thus treating the ] as data rather than
914 terminating the class. This meant it could skip too much.]
915
916 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
917 /g option in pcretest when the pattern contains \K, which makes it possible
918 to have an empty string match not at the start, even when the pattern is
919 anchored. Updated pcretest and pcredemo to use this option.
920
921 21. If the maximum number of capturing subpatterns in a recursion was greater
922 than the maximum at the outer level, the higher number was returned, but
923 with unset values at the outer level. The correct (outer level) value is
924 now given.
925
926 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
927 PCRE did not set those parentheses (unlike Perl). I have now found a way to
928 make it do so. The string so far is captured, making this feature
929 compatible with Perl.
930
931 23. The tests have been re-organized, adding tests 11 and 12, to make it
932 possible to check the Perl 5.10 features against Perl 5.10.
933
934 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
935 pattern matches a fixed length string. PCRE did not allow this; now it
936 does. Neither allows recursion.
937
938 25. I finally figured out how to implement a request to provide the minimum
939 length of subject string that was needed in order to match a given pattern.
940 (It was back references and recursion that I had previously got hung up
941 on.) This code has now been added to pcre_study(); it finds a lower bound
942 to the length of subject needed. It is not necessarily the greatest lower
943 bound, but using it to avoid searching strings that are too short does give
944 some useful speed-ups. The value is available to calling programs via
945 pcre_fullinfo().
946
947 26. While implementing 25, I discovered to my embarrassment that pcretest had
948 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
949 study optimizations had never been tested with that matching function.
950 Oops. What is worse, even when it was passed study data, there was a bug in
951 pcre_dfa_exec() that meant it never actually used it. Double oops. There
952 were also very few tests of studied patterns with pcre_dfa_exec().
953
954 27. If (?| is used to create subpatterns with duplicate numbers, they are now
955 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
956 on the other side of the coin, they are no longer allowed to have different
957 names, because these cannot be distinguished in PCRE, and this has caused
958 confusion. (This is a difference from Perl.)
959
960 28. When duplicate subpattern names are present (necessarily with different
961 numbers, as required by 27 above), and a test is made by name in a
962 conditional pattern, either for a subpattern having been matched, or for
963 recursion in such a pattern, all the associated numbered subpatterns are
964 tested, and the overall condition is true if the condition is true for any
965 one of them. This is the way Perl works, and is also more like the way
966 testing by number works.
967
968
969 Version 7.9 11-Apr-09
970 ---------------------
971
972 1. When building with support for bzlib/zlib (pcregrep) and/or readline
973 (pcretest), all targets were linked against these libraries. This included
974 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
975 libraries. This caused unwanted dependencies to be created. This problem
976 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
977 pcretest is linked with readline.
978
979 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
980 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
981 moved outside it again, because FALSE and TRUE are already defined in AIX,
982 but BOOL is not.
983
984 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
985 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
986
987 4. The pcregrep documentation said spaces were inserted as well as colons (or
988 hyphens) following file names and line numbers when outputting matching
989 lines. This is not true; no spaces are inserted. I have also clarified the
990 wording for the --colour (or --color) option.
991
992 5. In pcregrep, when --colour was used with -o, the list of matching strings
993 was not coloured; this is different to GNU grep, so I have changed it to be
994 the same.
995
996 6. When --colo(u)r was used in pcregrep, only the first matching substring in
997 each matching line was coloured. Now it goes on to look for further matches
998 of any of the test patterns, which is the same behaviour as GNU grep.
999
1000 7. A pattern that could match an empty string could cause pcregrep to loop; it
1001 doesn't make sense to accept an empty string match in pcregrep, so I have
1002 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1003 seems to be how GNU grep behaves.
1004
1005 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1006 start or after a newline", because the conditional assertion was not being
1007 correctly handled. The rule now is that both the assertion and what follows
1008 in the first alternative must satisfy the test.
1009
1010 9. If auto-callout was enabled in a pattern with a conditional group whose
1011 condition was an assertion, PCRE could crash during matching, both with
1012 pcre_exec() and pcre_dfa_exec().
1013
1014 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1015 used for matching.
1016
1017 11. Unicode property support in character classes was not working for
1018 characters (bytes) greater than 127 when not in UTF-8 mode.
1019
1020 12. Added the -M command line option to pcretest.
1021
1022 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1023
1024 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1025
1026 16. Added comments and documentation about mis-use of no_arg in the C++
1027 wrapper.
1028
1029 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1030 from Martin Jerabek that uses macro names for all relevant character and
1031 string constants.
1032
1033 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1034 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1035 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1036 these, but not everybody uses configure.
1037
1038 19. A conditional group that had only one branch was not being correctly
1039 recognized as an item that could match an empty string. This meant that an
1040 enclosing group might also not be so recognized, causing infinite looping
1041 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1042 with the subject "ab", where knowledge that the repeated group can match
1043 nothing is needed in order to break the loop.
1044
1045 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1046 exec(), but without supplying a callout function, matching went wrong.
1047
1048 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1049 leak if the size of the offset vector was greater than 30. When the vector
1050 is smaller, the saved offsets during recursion go onto a local stack
1051 vector, but for larger vectors malloc() is used. It was failing to free
1052 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1053 error, in fact).
1054
1055 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1056 heapframe that is used only when UTF-8 support is enabled. This caused no
1057 problem, but was untidy.
1058
1059 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1060 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1061 included within another project.
1062
1063 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1064 slightly modified by me:
1065
1066 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1067 not building pcregrep.
1068
1069 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1070 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1071
1072 25. Forward references, both numeric and by name, in patterns that made use of
1073 duplicate group numbers, could behave incorrectly or give incorrect errors,
1074 because when scanning forward to find the reference group, PCRE was not
1075 taking into account the duplicate group numbers. A pattern such as
1076 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1077
1078 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1079 the feature of a custom pointer more persuasive (as requested by a user).
1080
1081 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1082 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1083 pre-defined.
1084
1085 28. Added support for (*UTF8) at the start of a pattern.
1086
1087 29. Arrange for flags added by the "release type" setting in CMake to be shown
1088 in the configuration summary.
1089
1090
1091 Version 7.8 05-Sep-08
1092 ---------------------
1093
1094 1. Replaced UCP searching code with optimized version as implemented for Ad
1095 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1096 stage table and inline lookup instead of a function, giving speed ups of 2
1097 to 5 times on some simple patterns that I tested. Permission was given to
1098 distribute the MultiStage2.py script that generates the tables (it's not in
1099 the tarball, but is in the Subversion repository).
1100
1101 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1102 scripts.
1103
1104 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1105 a group with a zero qualifier. The result of the study could be incorrect,
1106 or the function might crash, depending on the pattern.
1107
1108 4. Caseless matching was not working for non-ASCII characters in back
1109 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1110 It now works when Unicode Property Support is available.
1111
1112 5. In pcretest, an escape such as \x{de} in the data was always generating
1113 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1114 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1115 truncation.
1116
1117 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1118
1119 7. Added two (int) casts to pcregrep when printing the difference of two
1120 pointers, in case they are 64-bit values.
1121
1122 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1123 test 2 if it fails.
1124
1125 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1126 and a #define of that name to empty if it is not externally set. This is to
1127 allow users of MSVC to set it if necessary.
1128
1129 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1130 the convenience functions in the pcre_get.c source file.
1131
1132 11. An option change at the start of a pattern that had top-level alternatives
1133 could cause overwriting and/or a crash. This command provoked a crash in
1134 some environments:
1135
1136 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1137
1138 This potential security problem was recorded as CVE-2008-2371.
1139
1140 12. For a pattern where the match had to start at the beginning or immediately
1141 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1142 pcre_dfa_exec() could read past the end of the passed subject if there was
1143 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1144 pcretest so that it places the subject at the end of its malloc-ed buffer.
1145
1146 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1147 exec() might read past the end of the data buffer in UTF-8 mode.
1148
1149 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1150 the data contained the byte 0x85 as part of a UTF-8 character within its
1151 first line. This applied both to normal and DFA matching.
1152
1153 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1154 /^[^d]*?$/8 failed to match "abc".
1155
1156 16. Added a missing copyright notice to pcrecpp_internal.h.
1157
1158 17. Make it more clear in the documentation that values returned from
1159 pcre_exec() in ovector are byte offsets, not character counts.
1160
1161 18. Tidied a few places to stop certain compilers from issuing warnings.
1162
1163 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1164 supplied by Stefan Weber. I made a further small update for 7.8 because
1165 there is a change of source arrangements: the pcre_searchfuncs.c module is
1166 replaced by pcre_ucd.c.
1167
1168
1169 Version 7.7 07-May-08
1170 ---------------------
1171
1172 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1173 a string to a long long, pretend we don't even have a long long." This is
1174 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1175
1176 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1177 pre-7.6 versions, which defined a global no_arg variable instead of putting
1178 it in the RE class. (See also #8 below.)
1179
1180 3. Remove a line of dead code, identified by coverity and reported by Nuno
1181 Lopes.
1182
1183 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1184
1185 (1) The include/exclude patterns were being applied to the whole pathnames
1186 of files, instead of just to the final components.
1187
1188 (2) If there was more than one level of directory, the subdirectories were
1189 skipped unless they satisfied the include/exclude conditions. This is
1190 inconsistent with GNU grep (and could even be seen as contrary to the
1191 pcregrep specification - which I improved to make it absolutely clear).
1192 The action now is always to scan all levels of directory, and just
1193 apply the include/exclude patterns to regular files.
1194
1195 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1196 --exclude_dir in the tests to avoid scanning .svn directories.
1197
1198 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1199 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1200 doesn't support NULs in patterns.
1201
1202 7. Added some missing "const"s to declarations of static tables in
1203 pcre_compile.c and pcre_dfa_exec.c.
1204
1205 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1206 caused by fix #2 above. (Subsequently also a second patch to fix the
1207 first patch. And a third patch - this was a messy problem.)
1208
1209 9. Applied Craig's patch to remove the use of push_back().
1210
1211 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1212 matching function regexec().
1213
1214 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1215 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1216 references. PCRE supports relative numbers with this syntax (I don't think
1217 Oniguruma does).
1218
1219 12. Previously, a group with a zero repeat such as (...){0} was completely
1220 omitted from the compiled regex. However, this means that if the group
1221 was called as a subroutine from elsewhere in the pattern, things went wrong
1222 (an internal error was given). Such groups are now left in the compiled
1223 pattern, with a new opcode that causes them to be skipped at execution
1224 time.
1225
1226 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1227 to the way PCRE behaves:
1228
1229 (a) A lone ] character is dis-allowed (Perl treats it as data).
1230
1231 (b) A back reference to an unmatched subpattern matches an empty string
1232 (Perl fails the current match path).
1233
1234 (c) A data ] in a character class must be notated as \] because if the
1235 first data character in a class is ], it defines an empty class. (In
1236 Perl it is not possible to have an empty class.) The empty class []
1237 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1238 The negative empty class [^] matches any one character, independently
1239 of the DOTALL setting.
1240
1241 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1242 non-existent subpattern following a character class starting with ']' and
1243 containing () gave an internal compiling error instead of "reference to
1244 non-existent subpattern". Fortunately, when the pattern did exist, the
1245 compiled code was correct. (When scanning forwards to check for the
1246 existencd of the subpattern, it was treating the data ']' as terminating
1247 the class, so got the count wrong. When actually compiling, the reference
1248 was subsequently set up correctly.)
1249
1250 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1251 it was being rejected as not supported by pcre_dfa_exec(), even though
1252 other assertions are supported. I have made pcre_dfa_exec() support
1253 (*FAIL).
1254
1255 16. The implementation of 13c above involved the invention of a new opcode,
1256 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1257 cannot be changed at match time, I realized I could make a small
1258 improvement to matching performance by compiling OP_ALLANY instead of
1259 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1260 on the OP_ANY path.
1261
1262 17. Compiling pcretest on Windows with readline support failed without the
1263 following two fixes: (1) Make the unistd.h include conditional on
1264 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1265
1266 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1267 ncurses library to be included for pcretest when ReadLine support is
1268 requested, but also to allow for it to be overridden. This patch came from
1269 Daniel Bergström.
1270
1271 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1272 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1273 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1274 spotting this.
1275
1276
1277 Version 7.6 28-Jan-08
1278 ---------------------
1279
1280 1. A character class containing a very large number of characters with
1281 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1282 overflow.
1283
1284 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1285 HAVE_LONG_LONG is not defined.
1286
1287 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1288 bring them up to date and include new features. This patch includes:
1289
1290 - Fixed PH's badly added libz and libbz2 support.
1291 - Fixed a problem with static linking.
1292 - Added pcredemo. [But later removed - see 7 below.]
1293 - Fixed dftables problem and added an option.
1294 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1295 HAVE_LONG_LONG.
1296 - Added readline support for pcretest.
1297 - Added an listing of the option settings after cmake has run.
1298
1299 4. A user submitted a patch to Makefile that makes it easy to create
1300 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1301 Makefile.am that cause it to include this special target, without
1302 affecting anything else. Note that the same mingw target plus all
1303 the other distribution libraries and programs are now supported
1304 when configuring with CMake (see 6 below) instead of with
1305 Configure/Make.
1306
1307 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1308 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1309 exported in the Windows port". It has not yet been confirmed that the patch
1310 solves the problem, but it does no harm.
1311
1312 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1313 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1314 with CMake, and also correct the comment about stack recursion.
1315
1316 7. Remove the automatic building of pcredemo from the ./configure system and
1317 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1318 of a program that users should build themselves after PCRE is installed, so
1319 building it automatically is not really right. What is more, it gave
1320 trouble in some build environments.
1321
1322 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1323
1324
1325 Version 7.5 10-Jan-08
1326 ---------------------
1327
1328 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1329 values in parens when parsing an RE using the C++ wrapper."
1330
1331 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1332 Characters greater than 255 were excluded from the class instead of being
1333 included.
1334
1335 3. The same bug as (2) above applied to negated POSIX classes such as
1336 [:^space:].
1337
1338 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1339 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1340 I have changed it.
1341
1342 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1343 first named subpattern) and a construct such as (?&a) would reference the
1344 first named subpattern whose name started with "a" (in other words, the
1345 length check was missing). Both these problems are fixed. "Subpattern name
1346 expected" is now given for (?&) (a zero-length name), and this patch also
1347 makes it give the same error for \k'' (previously it complained that that
1348 was a reference to a non-existent subpattern).
1349
1350 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1351 this is right because (?- can be followed by option settings as well as by
1352 digits. I have, however, made the messages clearer.
1353
1354 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1355 than the number used in the conditional) now cause a compile-time error.
1356 This is actually not compatible with Perl, which accepts such patterns, but
1357 treats the conditional as always being FALSE (as PCRE used to), but it
1358 seems to me that giving a diagnostic is better.
1359
1360 8. Change "alphameric" to the more common word "alphanumeric" in comments
1361 and messages.
1362
1363 9. Fix two occurrences of "backslash" in comments that should have been
1364 "backspace".
1365
1366 10. Remove two redundant lines of code that can never be obeyed (their function
1367 was moved elsewhere).
1368
1369 11. The program that makes PCRE's Unicode character property table had a bug
1370 which caused it to generate incorrect table entries for sequences of
1371 characters that have the same character type, but are in different scripts.
1372 It amalgamated them into a single range, with the script of the first of
1373 them. In other words, some characters were in the wrong script. There were
1374 thirteen such cases, affecting characters in the following ranges:
1375
1376 U+002b0 - U+002c1
1377 U+0060c - U+0060d
1378 U+0061e - U+00612
1379 U+0064b - U+0065e
1380 U+0074d - U+0076d
1381 U+01800 - U+01805
1382 U+01d00 - U+01d77
1383 U+01d9b - U+01dbf
1384 U+0200b - U+0200f
1385 U+030fc - U+030fe
1386 U+03260 - U+0327f
1387 U+0fb46 - U+0fbb1
1388 U+10450 - U+1049d
1389
1390 12. The -o option (show only the matching part of a line) for pcregrep was not
1391 compatible with GNU grep in that, if there was more than one match in a
1392 line, it showed only the first of them. It now behaves in the same way as
1393 GNU grep.
1394
1395 13. If the -o and -v options were combined for pcregrep, it printed a blank
1396 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1397 does the same. The return code can be used to tell if there were any
1398 non-matching lines.
1399
1400 14. Added --file-offsets and --line-offsets to pcregrep.
1401
1402 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1403 infinitely looping recursion. The bug was that positive lookaheads were not
1404 being skipped when checking for a possible empty match (negative lookaheads
1405 and both kinds of lookbehind were skipped).
1406
1407 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1408 inclusion of <windows.h> to before rather than after the definition of
1409 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1410
1411 17. Specifying a possessive quantifier with a specific limit for a Unicode
1412 character property caused pcre_compile() to compile bad code, which led at
1413 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1414 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1415 caused the error; without that there was no problem.
1416
1417 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1418
1419 19. Added --enable-pcretest-libreadline.
1420
1421 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1422 RE::GlobalReplace(). As a result, the number of replacements returned was
1423 double what it should be. I removed one of the increments, but Craig sent a
1424 later patch that removed the other one (the right fix) and added unit tests
1425 that check the return values (which was not done before).
1426
1427 21. Several CMake things:
1428
1429 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1430 the names libpcre and libpcreposix, not just pcre and pcreposix.
1431
1432 (2) The above change means that pcretest and pcregrep are now correctly
1433 linked with the newly-built libraries, not previously installed ones.
1434
1435 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1436
1437 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1438 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1439 UTF-8 newline character). The key issue is that the pattern starts .*;
1440 this means that the match must be either at the beginning, or after a
1441 newline. The bug was in the code for advancing after a failed match and
1442 checking that the new position followed a newline. It was not taking
1443 account of UTF-8 characters correctly.
1444
1445 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1446 character classes. PCRE was not treating the sequence [:...:] as a
1447 character class unless the ... were all letters. Perl, however, seems to
1448 allow any characters between [: and :], though of course it rejects as
1449 unknown any "names" that contain non-letters, because all the known class
1450 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1451 for example, whereas PCRE did not - it did not recognize a POSIX character
1452 class. This seemed a bit dangerous, so the code has been changed to be
1453 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1454 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1455 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1456 Perl does, and where it didn't before.
1457
1458 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1459 Windows environments %n is disabled by default.
1460
1461
1462 Version 7.4 21-Sep-07
1463 ---------------------
1464
1465 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1466 means that a class such as [\s] counted as "explicit reference to CR or
1467 LF". That isn't really right - the whole point of the change was to try to
1468 help when there was an actual mention of one of the two characters. So now
1469 the change happens only if \r or \n (or a literal CR or LF) character is
1470 encountered.
1471
1472 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1473 of both had grown to the point where there were only 3 bits left.
1474 Fortunately, there was spare space in the data structure, and so I have
1475 moved the internal flags into a new 16-bit field to free up more option
1476 bits.
1477
1478 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1479 but did not set the internal JCHANGED flag - either of these is enough to
1480 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1481 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1482 start sets both bits.
1483
1484 4. Added options (at build time, compile time, exec time) to change \R from
1485 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1486
1487 5. doc/pcresyntax.html was missing from the distribution.
1488
1489 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1490 compatibility, even though it is no longer used.
1491
1492 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1493 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1494 windows.h file is present (where different names are used). [This was
1495 reversed later after testing - see 16 below.]
1496
1497 8. Changed all #include <config.h> to #include "config.h". There were also
1498 some further <pcre.h> cases that I changed to "pcre.h".
1499
1500 9. When pcregrep was used with the --colour option, it missed the line ending
1501 sequence off the lines that it output.
1502
1503 10. It was pointed out to me that arrays of string pointers cause lots of
1504 relocations when a shared library is dynamically loaded. A technique of
1505 using a single long string with a table of offsets can drastically reduce
1506 these. I have refactored PCRE in four places to do this. The result is
1507 dramatic:
1508
1509 Originally: 290
1510 After changing UCP table: 187
1511 After changing error message table: 43
1512 After changing table of "verbs" 36
1513 After changing table of Posix names 22
1514
1515 Thanks to the folks working on Gregex for glib for this insight.
1516
1517 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1518 unicode-properties was also set.
1519
1520 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1521
1522 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1523 checked only for CRLF.
1524
1525 14. Added casts to pcretest.c to avoid compiler warnings.
1526
1527 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1528
1529 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1530 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1531 entirely. This removes changes made in 7 above.
1532
1533 17. The CMake files have been updated, and there is now more information about
1534 building with CMake in the NON-UNIX-USE document.
1535
1536
1537 Version 7.3 28-Aug-07
1538 ---------------------
1539
1540 1. In the rejigging of the build system that eventually resulted in 7.1, the
1541 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1542 brackets there is not right, since it causes compilers to look for an
1543 installed pcre.h, not the version that is in the source that is being
1544 compiled (which of course may be different). I have changed it back to:
1545
1546 #include "pcre.h"
1547
1548 I have a vague recollection that the change was concerned with compiling in
1549 different directories, but in the new build system, that is taken care of
1550 by the VPATH setting the Makefile.
1551
1552 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1553 when the subject happened to end in the byte 0x85 (e.g. if the last
1554 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1555 characters but of course it shouldn't be taken as a newline when it is part
1556 of another character. The bug was that, for an unlimited repeat of . in
1557 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1558 characters when looking for a newline.
1559
1560 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1561
1562 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1563 in debug output.
1564
1565 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1566 long printing in the pcrecpp unittest when running under MinGW.
1567
1568 6. ESC_K was left out of the EBCDIC table.
1569
1570 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1571 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1572 limit also applies to "virtual nesting" when a pattern is recursive, and in
1573 this case 1000 isn't so big. I have been able to remove this limit at the
1574 expense of backing off one optimization in certain circumstances. Normally,
1575 when pcre_exec() would call its internal match() function recursively and
1576 immediately return the result unconditionally, it uses a "tail recursion"
1577 feature to save stack. However, when a subpattern that can match an empty
1578 string has an unlimited repetition quantifier, it no longer makes this
1579 optimization. That gives it a stack frame in which to save the data for
1580 checking that an empty string has been matched. Previously this was taken
1581 from the 1000-entry workspace that had been reserved. So now there is no
1582 explicit limit, but more stack is used.
1583
1584 8. Applied Daniel's patches to solve problems with the import/export magic
1585 syntax that is required for Windows, and which was going wrong for the
1586 pcreposix and pcrecpp parts of the library. These were overlooked when this
1587 problem was solved for the main library.
1588
1589 9. There were some crude static tests to avoid integer overflow when computing
1590 the size of patterns that contain repeated groups with explicit upper
1591 limits. As the maximum quantifier is 65535, the maximum group length was
1592 set at 30,000 so that the product of these two numbers did not overflow a
1593 32-bit integer. However, it turns out that people want to use groups that
1594 are longer than 30,000 bytes (though not repeat them that many times).
1595 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1596 made it possible to implement the integer overflow checks in a much more
1597 dynamic way, which I have now done. The artificial limitation on group
1598 length has been removed - we now have only the limit on the total length of
1599 the compiled pattern, which depends on the LINK_SIZE setting.
1600
1601 10. Fixed a bug in the documentation for get/copy named substring when
1602 duplicate names are permitted. If none of the named substrings are set, the
1603 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1604 empty string.
1605
1606 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1607 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1608 because the ] is interpreted as the first data character and the
1609 terminating ] is not found. PCRE has been made compatible with Perl in this
1610 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1611 cause memory overwriting.
1612
1613 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1614 string has been matched (to stop an infinite loop). It was not recognizing
1615 a conditional subpattern that could match an empty string if that
1616 subpattern was within another subpattern. For example, it looped when
1617 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1618 condition was not nested. This bug has been fixed.
1619
1620 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1621 past the start of the subject in the presence of bytes with the top bit
1622 set, for example "\x8aBCD".
1623
1624 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1625 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1626
1627 14. Optimized (?!) to (*FAIL).
1628
1629 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1630 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1631 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1632 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1633 does: it's just the validity check that is more restrictive.
1634
1635 16. Inserted checks for integer overflows during escape sequence (backslash)
1636 processing, and also fixed erroneous offset values for syntax errors during
1637 backslash processing.
1638
1639 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1640 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1641
1642 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1643 caused an overrun.
1644
1645 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1646 something other than just ASCII characters) inside a group that had an
1647 unlimited repeat caused a loop at compile time (while checking to see
1648 whether the group could match an empty string).
1649
1650 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1651 [\P{Any}] did so. (Error in the code for printing property names.)
1652
1653 21. An orphan \E inside a character class could cause a crash.
1654
1655 22. A repeated capturing bracket such as (A)? could cause a wild memory
1656 reference during compilation.
1657
1658 23. There are several functions in pcre_compile() that scan along a compiled
1659 expression for various reasons (e.g. to see if it's fixed length for look
1660 behind). There were bugs in these functions when a repeated \p or \P was
1661 present in the pattern. These operators have additional parameters compared
1662 with \d, etc, and these were not being taken into account when moving along
1663 the compiled data. Specifically:
1664
1665 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1666 length.
1667
1668 (b) An item such as \pL+ within a repeated group could cause crashes or
1669 loops.
1670
1671 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1672 "reference to non-existent subpattern" error.
1673
1674 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1675
1676 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1677 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1678
1679 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1680
1681 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1682 character were causing crashes (broken optimization).
1683
1684 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1685 \p or \P) caused a compile-time loop.
1686
1687 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1688 break. For example, the unstudied pattern [\r\n]A does not match the string
1689 "\r\nA" because change 7.0/46 below moves the current point on by two
1690 characters after failing to match at the start. However, the pattern \nA
1691 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1692 the same is true. There doesn't seem any very clean way out of this, but
1693 what I have chosen to do makes the common cases work: PCRE now takes note
1694 of whether there can be an explicit match for \r or \n anywhere in the
1695 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1696 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1697 pattern has explicit CR or LF references.
1698
1699 29. Added (*CR) etc for changing newline setting at start of pattern.
1700
1701
1702 Version 7.2 19-Jun-07
1703 ---------------------
1704
1705 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1706 which is apparently normally available under Windows.
1707
1708 2. Re-jig the pcregrep tests with different newline settings in an attempt
1709 to make them independent of the local environment's newline setting.
1710
1711 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1712
1713 4. Some of the "internals" tests were previously cut out when the link size
1714 was not 2, because the output contained actual offsets. The recent new
1715 "Z" feature of pcretest means that these can be cut out, making the tests
1716 usable with all link sizes.
1717
1718 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1719 stack recursion. This gives a massive performance boost under BSD, but just
1720 a small improvement under Linux. However, it saves one field in the frame
1721 in all cases.
1722
1723 6. Added more features from the forthcoming Perl 5.10:
1724
1725 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1726 recursion call. It refers to the nth most recently opened parentheses.
1727
1728 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1729 to be opened parentheses.
1730
1731 (c) Conditions that refer to capturing parentheses can be specified
1732 relatively, for example, (?(-2)... or (?(+3)...
1733
1734 (d) \K resets the start of the current match so that everything before
1735 is not part of it.
1736
1737 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1738
1739 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1740 reference syntax.
1741
1742 (g) (?| introduces a group in which the numbering of parentheses in each
1743 alternative starts with the same number.
1744
1745 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1746
1747 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1748 PCRE_INFO_JCHANGED.
1749
1750 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1751 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1752 for detecting groups that can match an empty string.
1753
1754 9. A pattern with a very large number of alternatives (more than several
1755 hundred) was running out of internal workspace during the pre-compile
1756 phase, where pcre_compile() figures out how much memory will be needed. A
1757 bit of new cunning has reduced the workspace needed for groups with
1758 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1759 workspace instead of running out of the 4096 that are available.
1760
1761 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1762
1763 11. Applied patch from Google to remove an optimization that didn't quite work.
1764 The report of the bug said:
1765
1766 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1767 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1768 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1769
1770 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1771 it matched the wrong number of bytes.
1772
1773
1774 Version 7.1 24-Apr-07
1775 ---------------------
1776
1777 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1778 that is more "standard", making use of automake and other Autotools. There
1779 is some re-arrangement of the files and adjustment of comments consequent
1780 on this.
1781
1782 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1783 for recursive directory scanning broke on some systems because the files
1784 are not scanned in any specific order and on different systems the order
1785 was different. A call to "sort" has been inserted into RunGrepTest for the
1786 approprate test as a short-term fix. In the longer term there may be an
1787 alternative.
1788
1789 3. I had an email from Eric Raymond about problems translating some of PCRE's
1790 man pages to HTML (despite the fact that I distribute HTML pages, some
1791 people do their own conversions for various reasons). The problems
1792 concerned the use of low-level troff macros .br and .in. I have therefore
1793 removed all such uses from the man pages (some were redundant, some could
1794 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1795 HTML has been updated to handle .nf/.fi and to complain if it encounters
1796 .br or .in.
1797
1798 4. Updated comments in configure.ac that get placed in config.h.in and also
1799 arranged for config.h to be included in the distribution, with the name
1800 config.h.generic, for the benefit of those who have to compile without
1801 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1802
1803 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1804 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1805 makevp.bat for the current PCRE, using the additional files
1806 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1807
1808 6. A Windows user reported a minor discrepancy with test 2, which turned out
1809 to be caused by a trailing space on an input line that had got lost in his
1810 copy. The trailing space was an accident, so I've just removed it.
1811
1812 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1813 that is needed.
1814
1815 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1816 as "const" (a) because they are and (b) because it helps the PHP
1817 maintainers who have recently made a script to detect big data structures
1818 in the php code that should be moved to the .rodata section. I remembered
1819 to update Builducptable as well, so it won't revert if ucptable.h is ever
1820 re-created.
1821
1822 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1823 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1824 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1825 support is not required. This saves 1.5-2K of code, which is important in
1826 some applications.
1827
1828 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1829 so as not to refer to the tables, even though these functions will never be
1830 called when UTF-8 support is disabled. Otherwise there are problems with a
1831 shared library.
1832
1833 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1834
1835 (a) It was defining its arguments as char * instead of void *.
1836
1837 (b) It was assuming that all moves were upwards in memory; this was true
1838 a long time ago when I wrote it, but is no longer the case.
1839
1840 The emulated memove() is provided for those environments that have neither
1841 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1842 is clearly not the case, as these two bugs were recently reported.
1843
1844 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1845 and Detrail to create the HTML documentation, the .txt form of the man
1846 pages, and it removes trailing spaces from listed files. It also creates
1847 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1848 case, it wraps all the #defines with #ifndefs. This script should be run
1849 before "make dist".
1850
1851 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1852 with Unicode property support.
1853
1854 (a) For a maximizing quantifier, if the two different cases of the
1855 character were of different lengths in their UTF-8 codings (there are
1856 some cases like this - I found 11), and the matching function had to
1857 back up over a mixture of the two cases, it incorrectly assumed they
1858 were both the same length.
1859
1860 (b) When PCRE was configured to use the heap rather than the stack for
1861 recursion during matching, it was not correctly preserving the data for
1862 the other case of a UTF-8 character when checking ahead for a match
1863 while processing a minimizing repeat. If the check also involved
1864 matching a wide character, but failed, corruption could cause an
1865 erroneous result when trying to check for a repeat of the original
1866 character.
1867
1868 13. Some tidying changes to the testing mechanism:
1869
1870 (a) The RunTest script now detects the internal link size and whether there
1871 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1872 values substituted by "configure". (The RunGrepTest script already did
1873 this for UTF-8.) The configure.ac script no longer substitutes the
1874 relevant variables.
1875
1876 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1877 with length and offset values. This means that the output is different
1878 for different internal link sizes. Test 2 is skipped for link sizes
1879 other than 2 because of this, bypassing the problem. Unfortunately,
1880 there was also a test in test 3 (the locale tests) that used /B and
1881 failed for link sizes other than 2. Rather than cut the whole test out,
1882 I have added a new /Z option to pcretest that replaces the length and
1883 offset values with spaces. This is now used to make test 3 independent
1884 of link size. (Test 2 will be tidied up later.)
1885
1886 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1887 segmentation fault instead of returning the appropriate error message.
1888
1889 15. In multiline mode when the newline sequence was set to "any", the pattern
1890 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1891 This doesn't seem right; it now treats the CRLF combination as the line
1892 ending, and so does not match in that case. It's only a pattern such as ^$
1893 that would hit this one: something like ^ABC$ would have failed after \r
1894 and then tried again after \r\n.
1895
1896 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1897 in an attempt to make files that differ only in their line terminators
1898 compare equal. This works on Linux.
1899
1900 17. Under certain error circumstances pcregrep might try to free random memory
1901 as it exited. This is now fixed, thanks to valgrind.
1902
1903 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1904 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1905 was because its rules for how to advance for /g after matching an empty
1906 string at the end of a line did not allow for this case. They now check for
1907 it specially.
1908
1909 20. pcretest is supposed to handle patterns and data of any length, by
1910 extending its buffers when necessary. It was getting this wrong when the
1911 buffer for a data line had to be extended.
1912
1913 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1914 CRLF as a newline sequence.
1915
1916 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1917 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1918 I have nevertheless tidied it up.
1919
1920 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1921
1922 24. Added a man page for pcre-config.
1923
1924
1925 Version 7.0 19-Dec-06
1926 ---------------------
1927
1928 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1929 moving to gcc 4.1.1.
1930
1931 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1932 sys/time.h, which is documented as needed for this function. It doesn't
1933 seem to matter on Linux, but it showed up on some releases of OS X.
1934
1935 3. It seems that there are systems where bytes whose values are greater than
1936 127 match isprint() in the "C" locale. The "C" locale should be the
1937 default when a C program starts up. In most systems, only ASCII printing
1938 characters match isprint(). This difference caused the output from pcretest
1939 to vary, making some of the tests fail. I have changed pcretest so that:
1940
1941 (a) When it is outputting text in the compiled version of a pattern, bytes
1942 other than 32-126 are always shown as hex escapes.
1943
1944 (b) When it is outputting text that is a matched part of a subject string,
1945 it does the same, unless a different locale has been set for the match
1946 (using the /L modifier). In this case, it uses isprint() to decide.
1947
1948 4. Fixed a major bug that caused incorrect computation of the amount of memory
1949 required for a compiled pattern when options that changed within the
1950 pattern affected the logic of the preliminary scan that determines the
1951 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1952 that the computed length was too small. The symptoms of this bug were
1953 either the PCRE error "internal error: code overflow" from pcre_compile(),
1954 or a glibc crash with a message such as "pcretest: free(): invalid next
1955 size (fast)". Examples of patterns that provoked this bug (shown in
1956 pcretest format) are:
1957
1958 /(?-x: )/x
1959 /(?x)(?-x: \s*#\s*)/
1960 /((?i)[\x{c0}])/8
1961 /(?i:[\x{c0}])/8
1962
1963 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1964 is now done differently.
1965
1966 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1967 wrapper classes; (b) implement a new function in the C++ scanner that is
1968 more efficient than the old way of doing things because it avoids levels of
1969 recursion in the regex matching; (c) add a paragraph to the documentation
1970 for the FullMatch() function.
1971
1972 6. The escape sequence \n was being treated as whatever was defined as
1973 "newline". Not only was this contrary to the documentation, which states
1974 that \n is character 10 (hex 0A), but it also went horribly wrong when
1975 "newline" was defined as CRLF. This has been fixed.
1976
1977 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1978 was being set to -1 for the "end of line" case (supposedly a value that no
1979 character can have). Though this value is never used (the check for end of
1980 line is "zero bytes in current character"), it caused compiler complaints.
1981 I've changed it to 0xffffffff.
1982
1983 8. In pcre_version.c, the version string was being built by a sequence of
1984 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1985 string (as it is for production releases) called a macro with an empty
1986 argument. The C standard says the result of this is undefined. The gcc
1987 compiler treats it as an empty string (which was what was wanted) but it is
1988 reported that Visual C gives an error. The source has been hacked around to
1989 avoid this problem.
1990
1991 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1992 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1993 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1994 of them did).
1995
1996 10. Originally, pcretest opened its input and output without "b"; then I was
1997 told that "b" was needed in some environments, so it was added for release
1998 5.0 to both the input and output. (It makes no difference on Unix-like
1999 systems.) Later I was told that it is wrong for the input on Windows. I've
2000 now abstracted the modes into two macros, to make it easier to fiddle with
2001 them, and removed "b" from the input mode under Windows.
2002
2003 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2004
2005 12. Added -help and --help to pcretest as an official way of being reminded
2006 of the options.
2007
2008 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2009 and pcrecpp.cc because they annoy compilers at high warning levels.
2010
2011 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2012
2013 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2014 scripts are not C programs :-) and which was not noticed because it works
2015 on Linux.
2016
2017 16. pcretest is supposed to handle any length of pattern and data line (as one
2018 line or as a continued sequence of lines) by extending its input buffer if
2019 necessary. This feature was broken for very long pattern lines, leading to
2020 a string of junk being passed to pcre_compile() if the pattern was longer
2021 than about 50K.
2022
2023 17. I have done a major re-factoring of the way pcre_compile() computes the
2024 amount of memory needed for a compiled pattern. Previously, there was code
2025 that made a preliminary scan of the pattern in order to do this. That was
2026 OK when PCRE was new, but as the facilities have expanded, it has become
2027 harder and harder to keep it in step with the real compile phase, and there
2028 have been a number of bugs (see for example, 4 above). I have now found a
2029 cunning way of running the real compile function in a "fake" mode that
2030 enables it to compute how much memory it would need, while actually only
2031 ever using a few hundred bytes of working memory and without too many
2032 tests of the mode. This should make future maintenance and development
2033 easier. A side effect of this work is that the limit of 200 on the nesting
2034 depth of parentheses has been removed (though this was never a serious
2035 limitation, I suspect). However, there is a downside: pcre_compile() now
2036 runs more slowly than before (30% or more, depending on the pattern). I
2037 hope this isn't a big issue. There is no effect on runtime performance.
2038
2039 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2040 newline (only possible for the last line of a file) and it was a
2041 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2042
2043 19. Added additional timing features to pcretest. (1) The -tm option now times
2044 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2045 separate command line item, by a number that specifies the number of
2046 repeats to use when timing. The default is 50000; this gives better
2047 precision, but takes uncomfortably long for very large patterns.
2048
2049 20. Extended pcre_study() to be more clever in cases where a branch of a
2050 subpattern has no definite first character. For example, (a*|b*)[cd] would
2051 previously give no result from pcre_study(). Now it recognizes that the
2052 first character must be a, b, c, or d.
2053
2054 21. There was an incorrect error "recursive call could loop indefinitely" if
2055 a subpattern (or the entire pattern) that was being tested for matching an
2056 empty string contained only one non-empty item after a nested subpattern.
2057 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2058 incorrectly, because the \d was being skipped in the check.
2059
2060 22. The pcretest program now has a new pattern option /B and a command line
2061 option -b, which is equivalent to adding /B to every pattern. This causes
2062 it to show the compiled bytecode, without the additional information that
2063 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2064 is the same as /B/I).
2065
2066 23. A new optimization is now able automatically to treat some sequences such
2067 as a*b as a*+b. More specifically, if something simple (such as a character
2068 or a simple class like \d) has an unlimited quantifier, and is followed by
2069 something that cannot possibly match the quantified thing, the quantifier
2070 is automatically "possessified".
2071
2072 24. A recursive reference to a subpattern whose number was greater than 39
2073 went wrong under certain circumstances in UTF-8 mode. This bug could also
2074 have affected the operation of pcre_study().
2075
2076 25. Realized that a little bit of performance could be had by replacing
2077 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2078
2079 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2080
2081 27. Possessive quantifiers such as a++ were previously implemented by turning
2082 them into atomic groups such as ($>a+). Now they have their own opcodes,
2083 which improves performance. This includes the automatically created ones
2084 from 23 above.
2085
2086 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2087 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2088 the first matched character to be a colon. This applied both to named and
2089 numbered groups.
2090
2091 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2092
2093 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2094 building PCRE on Windows easier, so I have included it in the distribution.
2095
2096 31. There is now a check in pcretest against a ridiculously large number being
2097 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2098 loop, the loop is abandoned.
2099
2100 32. Forward references to subpatterns in conditions such as (?(2)...) where
2101 subpattern 2 is defined later cause pcre_compile() to search forwards in
2102 the pattern for the relevant set of parentheses. This search went wrong
2103 when there were unescaped parentheses in a character class, parentheses
2104 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2105
2106 33. "Subroutine" calls and backreferences were previously restricted to
2107 referencing subpatterns earlier in the regex. This restriction has now
2108 been removed.
2109
2110 34. Added a number of extra features that are going to be in Perl 5.10. On the
2111 whole, these are just syntactic alternatives for features that PCRE had
2112 previously implemented using the Python syntax or my own invention. The
2113 other formats are all retained for compatibility.
2114
2115 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2116 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2117 also .NET compatible.
2118
2119 (b) A recursion or subroutine call to a named group can now be defined as
2120 (?&name) as well as (?P>name).
2121
2122 (c) A backreference to a named group can now be defined as \k<name> or
2123 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2124 5.10, are also .NET compatible.
2125
2126 (d) A conditional reference to a named group can now use the syntax
2127 (?(<name>) or (?('name') as well as (?(name).
2128
2129 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2130 groups (named and numbered) that are never evaluated inline, but can be
2131 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2132 is always false. There may be only one alternative in such a group.
2133
2134 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2135 as the simple (?(R). The condition is true only if the most recent
2136 recursion is that of the given number or name. It does not search out
2137 through the entire recursion stack.
2138
2139 (g) The escape \gN or \g{N} has been added, where N is a positive or
2140 negative number, specifying an absolute or relative reference.
2141
2142 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2143 some "unreachable code" warnings.
2144
2145 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2146 things, this adds five new scripts.
2147
2148 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2149 There were also incompatibilities regarding the handling of \Q..\E inside
2150 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2151 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2152
2153 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2154 matches an empty string, and forcibly breaks the loop. There were bugs in
2155 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2156 against aaaa the result was just "a" rather than "aaaa", for example. Two
2157 separate and independent bugs (that affected different cases) have been
2158 fixed.
2159
2160 39. Refactored the code to abolish the use of different opcodes for small
2161 capturing bracket numbers. This is a tidy that I avoided doing when I
2162 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2163 The new approach is not only tidier, it makes it possible to reduce the
2164 memory needed to fix the previous bug (38).
2165
2166 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2167 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2168 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2169 mode.
2170
2171 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2172 report.
2173
2174 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2175 copy construction and assignment in the C++ wrapper.
2176
2177 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2178 couple of bugs that could have given wrong results in the "--newline=crlf"
2179 case.
2180
2181 44. Added a number of casts and did some reorganization of signed/unsigned int
2182 variables following suggestions from Dair Grant. Also renamed the variable
2183 "this" as "item" because it is a C++ keyword.
2184
2185 45. Arranged for dftables to add
2186
2187 #include "pcre_internal.h"
2188
2189 to pcre_chartables.c because without it, gcc 4.x may remove the array
2190 definition from the final binary if PCRE is built into a static library and
2191 dead code stripping is activated.
2192
2193 46. For an unanchored pattern, if a match attempt fails at the start of a
2194 newline sequence, and the newline setting is CRLF or ANY, and the next two
2195 characters are CRLF, advance by two characters instead of one.
2196
2197
2198 Version 6.7 04-Jul-06
2199 ---------------------
2200
2201 1. In order to handle tests when input lines are enormously long, pcretest has
2202 been re-factored so that it automatically extends its buffers when
2203 necessary. The code is crude, but this _is_ just a test program. The
2204 default size has been increased from 32K to 50K.
2205
2206 2. The code in pcre_study() was using the value of the re argument before
2207 testing it for NULL. (Of course, in any sensible call of the function, it
2208 won't be NULL.)
2209
2210 3. The memmove() emulation function in pcre_internal.h, which is used on
2211 systems that lack both memmove() and bcopy() - that is, hardly ever -
2212 was missing a "static" storage class specifier.
2213
2214 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2215 containing an extended class (one that cannot be represented by a bitmap
2216 because it contains high-valued characters or Unicode property items, e.g.
2217 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2218 pattern, but PCRE should not loop if you do not (it no longer does).
2219 [Detail: two cases were found: (a) a repeated subpattern containing an
2220 extended class; (b) a recursive reference to a subpattern that followed a
2221 previous extended class. It wasn't skipping over the extended class
2222 correctly when UTF-8 mode was not set.]
2223
2224 5. A negated single-character class was not being recognized as fixed-length
2225 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2226 compile error "lookbehind assertion is not fixed length".
2227
2228 6. The RunPerlTest auxiliary script was showing an unexpected difference
2229 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2230 write a Perl script that can interpret lines of an input file either as
2231 byte characters or as UTF-8, which is what "perltest" was being required to
2232 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2233 can't do is switch easily at run time between having the "use utf8;" pragma
2234 or not. In the end, I fudged it by using the RunPerlTest script to insert
2235 "use utf8;" explicitly for the UTF-8 tests.
2236
2237 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2238 the end of the subject string, contrary to the documentation and to what
2239 Perl does. This was true of both matching functions. Now it matches only at
2240 the start of the subject and immediately after *internal* newlines.
2241
2242 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2243 a pointer to an int instead of a pointer to an unsigned long int. This
2244 caused problems on 64-bit systems.
2245
2246 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2247 instance of the 'standard' template library not being so standard".
2248
2249 10. There was no check on the number of named subpatterns nor the maximum
2250 length of a subpattern name. The product of these values is used to compute
2251 the size of the memory block for a compiled pattern. By supplying a very
2252 long subpattern name and a large number of named subpatterns, the size
2253 computation could be caused to overflow. This is now prevented by limiting
2254 the length of names to 32 characters, and the number of named subpatterns
2255 to 10,000.
2256
2257 11. Subpatterns that are repeated with specific counts have to be replicated in
2258 the compiled pattern. The size of memory for this was computed from the
2259 length of the subpattern and the repeat count. The latter is limited to
2260 65535, but there was no limit on the former, meaning that integer overflow
2261 could in principle occur. The compiled length of a repeated subpattern is
2262 now limited to 30,000 bytes in order to prevent this.
2263
2264 12. Added the optional facility to have named substrings with the same name.
2265
2266 13. Added the ability to use a named substring as a condition, using the
2267 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2268 are numbers (not recommended). Forward references are permitted.
2269
2270 14. Added forward references in named backreferences (if you see what I mean).
2271
2272 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2273 pattern could run off the end of the subject. For example, the pattern
2274 "(?s)(.{1,5})"8 did this with the subject "ab".
2275
2276 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2277 PCRE_CASELESS was set when matching characters that were quantified with ?
2278 or *.
2279
2280 17. A character class other than a single negated character that had a minimum
2281 but no maximum quantifier - for example [ab]{6,} - was not handled
2282 correctly by pce_dfa_exec(). It would match only one character.
2283
2284 18. A valid (though odd) pattern that looked like a POSIX character
2285 class but used an invalid character after [ (for example [[,abc,]]) caused
2286 pcre_compile() to give the error "Failed: internal error: code overflow" or
2287 in some cases to crash with a glibc free() error. This could even happen if
2288 the pattern terminated after [[ but there just happened to be a sequence of
2289 letters, a binary zero, and a closing ] in the memory that followed.
2290
2291 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2292 over the years. Originally (before any Unicode support), just the bottom 8
2293 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2294 output from "man perlunicode" includes this:
2295
2296 The regular expression compiler produces polymorphic opcodes. That
2297 is, the pattern adapts to the data and automatically switches to
2298 the Unicode character scheme when presented with Unicode data--or
2299 instead uses a traditional byte scheme when presented with byte
2300 data.
2301
2302 Sadly, a wide octal escape does not cause a switch, and in a string with
2303 no other multibyte characters, these octal escapes are treated as before.
2304 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2305 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2306 Unicode string.
2307
2308 I have not perpetrated such confusion in PCRE. Up till now, it took just
2309 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2310 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2311 translate to the appropriate multibyte character.
2312
2313 29. Applied some refactoring to reduce the number of warnings from Microsoft
2314 and Borland compilers. This has included removing the fudge introduced
2315 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2316 a warning about an unused variable.
2317
2318 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2319 characters since release 4.0, because Perl (from release 5.004) does not.
2320 [Or at least, is documented not to: some releases seem to be in conflict
2321 with the documentation.] However, when a pattern was studied with
2322 pcre_study() and all its branches started with \s, PCRE still included VT
2323 as a possible starting character. Of course, this did no harm; it just
2324 caused an unnecessary match attempt.
2325
2326 22. Removed a now-redundant internal flag bit that recorded the fact that case
2327 dependency changed within the pattern. This was once needed for "required
2328 byte" processing, but is no longer used. This recovers a now-scarce options
2329 bit. Also moved the least significant internal flag bit to the most-
2330 significant bit of the word, which was not previously used (hangover from
2331 the days when it was an int rather than a uint) to free up another bit for
2332 the future.
2333
2334 23. Added support for CRLF line endings as well as CR and LF. As well as the
2335 default being selectable at build time, it can now be changed at runtime
2336 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2337 specify that it is scanning data with non-default line endings.
2338
2339 24. Changed the definition of CXXLINK to make it agree with the definition of
2340 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2341
2342 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2343 recursions. This makes a big different to stack usage for some patterns.
2344
2345 26. If a subpattern containing a named recursion or subroutine reference such
2346 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2347 the space required for the compiled pattern went wrong and gave too small a
2348 value. Depending on the environment, this could lead to "Failed: internal
2349 error: code overflow at offset 49" or "glibc detected double free or
2350 corruption" errors.
2351
2352 27. Applied patches from Google (a) to support the new newline modes and (b) to
2353 advance over multibyte UTF-8 characters in GlobalReplace.
2354
2355 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2356 difference for some implementation of PCRE in some Windows version.
2357
2358 29. Added some extra testing facilities to pcretest:
2359
2360 \q<number> in a data line sets the "match limit" value
2361 \Q<number> in a data line sets the "match recursion limt" value
2362 -S <number> sets the stack size, where <number> is in megabytes
2363
2364 The -S option isn't available for Windows.
2365
2366
2367 Version 6.6 06-Feb-06
2368 ---------------------
2369
2370 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2371 in pcreposix.h. I have copied the definition from pcre.h.
2372
2373 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2374 because pcre.h is no longer a built file.
2375
2376 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2377 not normally included in the compiled code.
2378
2379
2380 Version 6.5 01-Feb-06
2381 ---------------------
2382
2383 1. When using the partial match feature with pcre_dfa_exec(), it was not
2384 anchoring the second and subsequent partial matches at the new starting
2385 point. This could lead to incorrect results. For example, with the pattern
2386 /1234/, partially matching against "123" and then "a4" gave a match.
2387
2388 2. Changes to pcregrep:
2389
2390 (a) All non-match returns from pcre_exec() were being treated as failures
2391 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2392 error message is output. Some extra information is given for the
2393 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2394 probably the only errors that are likely to be caused by users (by
2395 specifying a regex that has nested indefinite repeats, for instance).
2396 If there are more than 20 of these errors, pcregrep is abandoned.
2397
2398 (b) A binary zero was treated as data while matching, but terminated the
2399 output line if it was written out. This has been fixed: binary zeroes
2400 are now no different to any other data bytes.
2401
2402 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2403 used to set a locale for matching. The --locale=xxxx long option has
2404 been added (no short equivalent) to specify a locale explicitly on the
2405 pcregrep command, overriding the environment variables.
2406
2407 (d) When -B was used with -n, some line numbers in the output were one less
2408 than they should have been.
2409
2410 (e) Added the -o (--only-matching) option.
2411
2412 (f) If -A or -C was used with -c (count only), some lines of context were
2413 accidentally printed for the final match.
2414
2415 (g) Added the -H (--with-filename) option.
2416
2417 (h) The combination of options -rh failed to suppress file names for files
2418 that were found from directory arguments.
2419
2420 (i) Added the -D (--devices) and -d (--directories) options.
2421
2422 (j) Added the -F (--fixed-strings) option.
2423
2424 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2425
2426 (l) Added the --colo(u)r option.
2427
2428 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2429 is not present by default.
2430
2431 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2432 items such as (?R) or (?1), when the recursion could match a number of
2433 alternatives. If it matched one of the alternatives, but subsequently,
2434 outside the recursion, there was a failure, the code tried to back up into
2435 the recursion. However, because of the way PCRE is implemented, this is not
2436 possible, and the result was an incorrect result from the match.
2437
2438 In order to prevent this happening, the specification of recursion has
2439 been changed so that all such subpatterns are automatically treated as
2440 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2441
2442 4. I had overlooked the fact that, in some locales, there are characters for
2443 which isalpha() is true but neither isupper() nor islower() are true. In
2444 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2445 and ordfeminine) are like this. This affected the treatment of \w and \W
2446 when they appeared in character classes, but not when they appeared outside
2447 a character class. The bit map for "word" characters is now created
2448 separately from the results of isalnum() instead of just taking it from the
2449 upper, lower, and digit maps. (Plus the underscore character, of course.)
2450
2451 5. The above bug also affected the handling of POSIX character classes such as
2452 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2453 permanent tables. Instead, the bit maps for such a class were previously
2454 created as the appropriate unions of the upper, lower, and digit bitmaps.
2455 Now they are created by subtraction from the [[:word:]] class, which has
2456 its own bitmap.
2457
2458 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2459 It is created by subtracting the vertical space characters (\x09, \x0a,
2460 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2461 subtraction was done in the overall bitmap for a character class, meaning
2462 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2463 be recognized. This bug has been fixed.
2464
2465 7. Patches from the folks at Google:
2466
2467 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2468 real life, but is still worth protecting against".
2469
2470 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2471 regular expressions".
2472
2473 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2474 have it.
2475
2476 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2477 "configure" and the latter not, in order to fix a problem somebody had
2478 with compiling the Arg class on HP-UX.
2479
2480 (e) Improve the error-handling of the C++ wrapper a little bit.
2481
2482 (f) New tests for checking recursion limiting.
2483
2484 8. The pcre_memmove() function, which is used only if the environment does not
2485 have a standard memmove() function (and is therefore rarely compiled),
2486 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2487 returning a result (though PCRE never actually uses the result).
2488
2489 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2490 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2491 returned instead of calling malloc() with an overflowing number that would
2492 most likely cause subsequent chaos.
2493
2494 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2495
2496 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2497 with this option is matched, the nmatch and pmatch options of regexec() are
2498 ignored.
2499
2500 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2501 provided in case anyone wants to the the POSIX interface with UTF-8
2502 strings.
2503
2504 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2505 C++ linking (needed for some HP-UX environments).
2506
2507 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2508 (unused parameter) and in the pcre_printint() function (omitted "default"
2509 switch label when the default is to do nothing).
2510
2511 15. Added some code to make it possible, when PCRE is compiled as a C++
2512 library, to replace subject pointers for pcre_exec() with a smart pointer
2513 class, thus making it possible to process discontinuous strings.
2514
2515 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2516 much the same function. They were added by different people who were trying
2517 to make PCRE easy to compile on non-Unix systems. It has been suggested
2518 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2519 for compiling on Windows systems. I have therefore replaced it with
2520 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2521 defaults to "extern" for C or "extern C" for C++, which works fine on
2522 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2523 SCOPE with something explicit in config.h. In addition:
2524
2525 (a) pcreposix.h still had just "extern" instead of either of these macros;
2526 I have replaced it with PCRE_DATA_SCOPE.
2527
2528 (b) Functions such as _pcre_xclass(), which are internal to the library,
2529 but external in the C sense, all had PCRE_EXPORT in their definitions.
2530 This is apparently wrong for the Windows case, so I have removed it.
2531 (It makes no difference on Unix-like systems.)
2532
2533 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2534 of recursive calls to match(). This is different to MATCH_LIMIT because
2535 that limits the total number of calls to match(), not all of which increase
2536 the depth of recursion. Limiting the recursion depth limits the amount of
2537 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2538 when PCRE is compiled, and changed at run time. A patch from Google adds
2539 this functionality to the C++ interface.
2540
2541 18. Changes to the handling of Unicode character properties:
2542
2543 (a) Updated the table to Unicode 4.1.0.
2544
2545 (b) Recognize characters that are not in the table as "Cn" (undefined).
2546
2547 (c) I revised the way the table is implemented to a much improved format
2548 which includes recognition of ranges. It now supports the ranges that
2549 are defined in UnicodeData.txt, and it also amalgamates other
2550 characters into ranges. This has reduced the number of entries in the
2551 table from around 16,000 to around 3,000, thus reducing its size
2552 considerably. I realized I did not need to use a tree structure after
2553 all - a binary chop search is just as efficient. Having reduced the
2554 number of entries, I extended their size from 6 bytes to 8 bytes to
2555 allow for more data.
2556
2557 (d) Added support for Unicode script names via properties such as \p{Han}.
2558
2559 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2560 matching that character.
2561
2562 20. When matching a repeated Unicode property with a minimum greater than zero,
2563 (for example \pL{2,}), PCRE could look past the end of the subject if it
2564 reached it while seeking the minimum number of characters. This could
2565 happen only if some of the characters were more than one byte long, because
2566 there is a check for at least the minimum number of bytes.
2567
2568 21. Refactored the implementation of \p and \P so as to be more general, to
2569 allow for more different types of property in future. This has changed the
2570 compiled form incompatibly. Anybody with saved compiled patterns that use
2571 \p or \P will have to recompile them.
2572
2573 22. Added "Any" and "L&" to the supported property types.
2574
2575 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2576 but give a compile time error if the value is greater than 0xff.
2577
2578 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2579 accidentally not being installed or uninstalled.
2580
2581 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2582 made were to insert the current release number. This seemed silly, because
2583 it made things harder for people building PCRE on systems that don't run
2584 "configure". I have turned pcre.h into a distributed file, no longer built
2585 by "configure", with the version identification directly included. There is
2586 no longer a pcre.h.in file.
2587
2588 However, this change necessitated a change to the pcre-config script as
2589 well. It is built from pcre-config.in, and one of the substitutions was the
2590 release number. I have updated configure.ac so that ./configure now finds
2591 the release number by grepping pcre.h.
2592
2593 26. Added the ability to run the tests under valgrind.
2594
2595
2596 Version 6.4 05-Sep-05
2597 ---------------------
2598
2599 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2600 "--" to be printed when multiple files were scanned, even when none of the
2601 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2602 consider it to be a bug, and have restored the previous behaviour.
2603
2604 2. A couple of code tidies to get rid of compiler warnings.
2605
2606 3. The pcretest program used to cheat by referring to symbols in the library
2607 whose names begin with _pcre_. These are internal symbols that are not
2608 really supposed to be visible externally, and in some environments it is
2609 possible to suppress them. The cheating is now confined to including
2610 certain files from the library's source, which is a bit cleaner.
2611
2612 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2613 file's purpose clearer.
2614
2615 5. Reorganized pcre_ucp_findchar().
2616
2617
2618 Version 6.3 15-Aug-05
2619 ---------------------
2620
2621 1. The file libpcre.pc.in did not have general read permission in the tarball.
2622
2623 2. There were some problems when building without C++ support:
2624
2625 (a) If C++ support was not built, "make install" and "make test" still
2626 tried to test it.
2627
2628 (b) There were problems when the value of CXX was explicitly set. Some
2629 changes have been made to try to fix these, and ...
2630
2631 (c) --disable-cpp can now be used to explicitly disable C++ support.
2632
2633 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2634 backslash in a target when C++ was disabled. This confuses some
2635 versions of "make", apparently. Using an intermediate variable solves
2636 this. (Same for CPP_LOBJ.)
2637
2638 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2639 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2640 necessary on certain architectures.
2641
2642 4. Added a setting of -export-symbols-regex to the link command to remove
2643 those symbols that are exported in the C sense, but actually are local
2644 within the library, and not documented. Their names all begin with
2645 "_pcre_". This is not a perfect job, because (a) we have to except some
2646 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2647 available (and never for static libraries). I have made a note to try to
2648 find a way round (a) in the future.
2649
2650
2651 Version 6.2 01-Aug-05
2652 ---------------------
2653
2654 1. There was no test for integer overflow of quantifier values. A construction
2655 such as {1111111111111111} would give undefined results. What is worse, if
2656 a minimum quantifier for a parenthesized subpattern overflowed and became
2657 negative, the calculation of the memory size went wrong. This could have
2658 led to memory overwriting.
2659
2660 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2661
2662 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2663 operating environments where this matters.
2664
2665 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2666 PCRE options from within the C++ wrapper.
2667
2668 5. Named capturing subpatterns were not being correctly counted when a pattern
2669 was compiled. This caused two problems: (a) If there were more than 100
2670 such subpatterns, the calculation of the memory needed for the whole
2671 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2672 back references of the form \12, where the number was greater than 9, were
2673 not recognized as back references, even though there were sufficient
2674 previous subpatterns.
2675
2676 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2677 versions of gcc, e.g. 2.95.4.
2678
2679
2680 Version 6.1 21-Jun-05
2681 ---------------------
2682
2683 1. There was one reference to the variable "posix" in pcretest.c that was not
2684 surrounded by "#if !defined NOPOSIX".
2685
2686 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2687 the cross-check on the old pcre_info() function, for the benefit of the
2688 cut-down version of PCRE that is currently imported into Exim.
2689
2690 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2691 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2692 patterns that start (?i) but I don't think that matters. The use of (?i) is
2693 just an example; this all applies to the other options as well.
2694
2695 4. Since libtool seems to echo the compile commands it is issuing, the output
2696 from "make" can be reduced a bit by putting "@" in front of each libtool
2697 compile command.
2698
2699 5. Patch from the folks at Google for configure.in to be a bit more thorough
2700 in checking for a suitable C++ installation before trying to compile the
2701 C++ stuff. This should fix a reported problem when a compiler was present,
2702 but no suitable headers.
2703
2704 6. The man pages all had just "PCRE" as their title. I have changed them to
2705 be the relevant file name. I have also arranged that these names are
2706 retained in the file doc/pcre.txt, which is a concatenation in text format
2707 of all the man pages except the little individual ones for each function.
2708
2709 7. The NON-UNIX-USE file had not been updated for the different set of source
2710 files that come with release 6. I also added a few comments about the C++
2711 wrapper.
2712
2713
2714 Version 6.0 07-Jun-05
2715 ---------------------
2716
2717 1. Some minor internal re-organization to help with my DFA experiments.
2718
2719 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2720 didn't matter for the library itself when fully configured, but did matter
2721 when compiling without UCP support, or within Exim, where the ucp files are
2722 not imported.
2723
2724 3. Refactoring of the library code to split up the various functions into
2725 different source modules. The addition of the new DFA matching code (see
2726 below) to a single monolithic source would have made it really too
2727 unwieldy, quite apart from causing all the code to be include in a
2728 statically linked application, when only some functions are used. This is
2729 relevant even without the DFA addition now that patterns can be compiled in
2730 one application and matched in another.
2731
2732 The downside of splitting up is that there have to be some external
2733 functions and data tables that are used internally in different modules of
2734 the library but which are not part of the API. These have all had their
2735 names changed to start with "_pcre_" so that they are unlikely to clash
2736 with other external names.
2737
2738 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2739 a different (DFA) algorithm. Although it is slower than the original
2740 function, it does have some advantages for certain types of matching
2741 problem.
2742
2743 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2744 including restarting after a partial match.
2745
2746 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2747 defined when compiling for Windows was sent to me. I have put it into the
2748 code, though I have no means of testing or verifying it.
2749
2750 7. Added the pcre_refcount() auxiliary function.
2751
2752 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2753 match before or at the first newline in the subject string. In pcretest,
2754 the /f option on a pattern can be used to set this.
2755
2756 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2757 would behave wrongly. This has been present in PCRE since release 4.0.
2758
2759 10. A number of changes to the pcregrep command:
2760
2761 (a) Refactored how -x works; insert ^(...)$ instead of setting
2762 PCRE_ANCHORED and checking the length, in preparation for adding
2763 something similar for -w.
2764
2765 (b) Added the -w (match as a word) option.
2766
2767 (c) Refactored the way lines are read and buffered so as to have more
2768 than one at a time available.
2769
2770 (d) Implemented a pcregrep test script.
2771
2772 (e) Added the -M (multiline match) option. This allows patterns to match
2773 over several lines of the subject. The buffering ensures that at least
2774 8K, or the rest of the document (whichever is the shorter) is available
2775 for matching (and similarly the previous 8K for lookbehind assertions).
2776
2777 (f) Changed the --help output so that it now says
2778
2779 -w, --word-regex(p)
2780
2781 instead of two lines, one with "regex" and the other with "regexp"
2782 because that confused at least one person since the short forms are the
2783 same. (This required a bit of code, as the output is generated
2784 automatically from a table. It wasn't just a text change.)
2785
2786 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2787 option but starts with a hyphen. Could be a pattern or a path name
2788 starting with a hyphen, for instance.
2789
2790 (h) "-" can be given as a file name to represent stdin.
2791
2792 (i) When file names are being printed, "(standard input)" is used for
2793 the standard input, for compatibility with GNU grep. Previously
2794 "<stdin>" was used.
2795
2796 (j) The option --label=xxx can be used to supply a name to be used for
2797 stdin when file names are being printed. There is no short form.
2798
2799 (k) Re-factored the options decoding logic because we are going to add
2800 two more options that take data. Such options can now be given in four
2801 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2802
2803 (l) Added the -A, -B, and -C options for requesting that lines of context
2804 around matches be printed.
2805
2806 (m) Added the -L option to print the names of files that do not contain
2807 any matching lines, that is, the complement of -l.
2808
2809 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2810 continue to scan other files.
2811
2812 (o) The -s option was incorrectly implemented. For compatibility with other
2813 greps, it now suppresses the error message for a non-existent or non-
2814 accessible file (but not the return code). There is a new option called
2815 -q that suppresses the output of matching lines, which was what -s was
2816 previously doing.
2817
2818 (p) Added --include and --exclude options to specify files for inclusion
2819 and exclusion when recursing.
2820
2821 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2822 Hopefully, it now does.
2823
2824 12. Missing cast in pcre_study().
2825
2826 13. Added an "uninstall" target to the makefile.
2827
2828 14. Replaced "extern" in the function prototypes in Makefile.in with
2829 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2830 world, but is set differently for Windows.
2831
2832 15. Added a second compiling function called pcre_compile2(). The only
2833 difference is that it has an extra argument, which is a pointer to an
2834 integer error code. When there is a compile-time failure, this is set
2835 non-zero, in addition to the error test pointer being set to point to an
2836 error message. The new argument may be NULL if no error number is required
2837 (but then you may as well call pcre_compile(), which is now just a
2838 wrapper). This facility is provided because some applications need a
2839 numeric error indication, but it has also enabled me to tidy up the way
2840 compile-time errors are handled in the POSIX wrapper.
2841
2842 16. Added VPATH=.libs to the makefile; this should help when building with one
2843 prefix path and installing with another. (Or so I'm told by someone who
2844 knows more about this stuff than I do.)
2845
2846 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2847 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2848 match everything, including newlines. This is not POSIX-compatible, but
2849 somebody wanted the feature. From pcretest it can be activated by using
2850 both the P and the s flags.
2851
2852 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2853
2854 19. libpcre.pc was being incorrectly installed as executable.
2855
2856 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2857 it now also looks for '\r' so that it will work unmodified on Windows.
2858
2859 21. Added Google's contributed C++ wrapper to the distribution.
2860
2861 22. Added some untidy missing memory free() calls in pcretest, to keep
2862 Electric Fence happy when testing.
2863
2864
2865
2866 Version 5.0 13-Sep-04
2867 ---------------------
2868
2869 1. Internal change: literal characters are no longer packed up into items
2870 containing multiple characters in a single byte-string. Each character
2871 is now matched using a separate opcode. However, there may be more than one
2872 byte in the character in UTF-8 mode.
2873
2874 2. The pcre_callout_block structure has two new fields: pattern_position and
2875 next_item_length. These contain the offset in the pattern to the next match
2876 item, and its length, respectively.
2877
2878 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2879 insertion of callouts before each pattern item. Added the /C option to
2880 pcretest to make use of this.
2881
2882 4. On the advice of a Windows user, the lines
2883
2884 #if defined(_WIN32) || defined(WIN32)
2885 _setmode( _fileno( stdout ), 0x8000 );
2886 #endif /* defined(_WIN32) || defined(WIN32) */
2887
2888 have been added to the source of pcretest. This apparently does useful
2889 magic in relation to line terminators.
2890
2891 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2892 for the benefit of those environments where the "b" makes a difference.
2893
2894 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2895 to know about it. I have put a hack into configure.in that adds in code
2896 to set GCC=yes if CC=icc. This seems to end up at a point in the
2897 generated configure script that is early enough to affect the setting of
2898 compiler options, which is what is needed, but I have no means of testing
2899 whether it really works. (The user who reported this had patched the
2900 generated configure script, which of course I cannot do.)
2901
2902 LATER: After change 22 below (new libtool files), the configure script
2903 seems to know about icc (and also ecc). Therefore, I have commented out
2904 this hack in configure.in.
2905
2906 7. Added support for pkg-config (2 patches were sent in).
2907
2908 8. Negated POSIX character classes that used a combination of internal tables
2909 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2910 [[:^ascii]]. Typically, they would match almost any characters. The other
2911 POSIX classes were not broken in this way.
2912
2913 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2914 to find the match, as PCRE was deluded into thinking that the match had to
2915 start at the start point or following a newline. The same bug applied to
2916 patterns with negative forward assertions or any backward assertions
2917 preceding ".*" at the start, unless the pattern required a fixed first
2918 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2919
2920 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2921 starting at the last subject character, bytes beyond the end of the subject
2922 string were read.
2923
2924 11. Renamed the variable "class" as "classbits" to make life easier for C++
2925 users. (Previously there was a macro definition, but it apparently wasn't
2926 enough.)
2927
2928 12. Added the new field "tables" to the extra data so that tables can be passed
2929 in at exec time, or the internal tables can be re-selected. This allows
2930 a compiled regex to be saved and re-used at a later time by a different
2931 program that might have everything at different addresses.
2932
2933 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2934 -R library as well as a -L library.
2935
2936 14. The debugging options of pcretest (-d on the command line or D on a
2937 pattern) showed incorrect output for anything following an extended class
2938 that contained multibyte characters and which was followed by a quantifier.
2939
2940 15. Added optional support for general category Unicode character properties
2941 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2942 support. It adds about 90K to the size of the library. The meanings of the
2943 inbuilt class escapes such as \d and \s have NOT been changed.
2944
2945 16. Updated pcredemo.c to include calls to free() to release the memory for the
2946 compiled pattern.
2947
2948 17. The generated file chartables.c was being created in the source directory
2949 instead of in the building directory. This caused the build to fail if the
2950 source directory was different from the building directory, and was
2951 read-only.
2952
2953 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2954 file. No doubt somebody will tell me if they don't make sense... Also added
2955 Dan Mooney's comments about building on OpenVMS.
2956
2957 19. Added support for partial matching via the PCRE_PARTIAL option for
2958 pcre_exec() and the \P data escape in pcretest.
2959
2960 20. Extended pcretest with 3 new pattern features:
2961
2962 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2963 write the compiled pattern to the file whose name is "rest-of-line".
2964 This is a straight binary dump of the data, with the saved pointer to
2965 the character tables forced to be NULL. The study data, if any, is
2966 written too. After writing, pcretest reads a new pattern.
2967
2968 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2969 compiled pattern from the given file. There must not be any
2970 occurrences of "<" in the file name (pretty unlikely); if there are,
2971 pcretest will instead treat the initial "<" as a pattern delimiter.
2972 After reading in the pattern, pcretest goes on to read data lines as
2973 usual.
2974
2975 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2976 and 16-bit fields in a compiled pattern, to simulate a pattern that
2977 was compiled on a host of opposite endianness.
2978
2979 21. The pcre-exec() function can now cope with patterns that were compiled on
2980 hosts of opposite endianness, with this restriction:
2981
2982 As for any compiled expression that is saved and used later, the tables
2983 pointer field cannot be preserved; the extra_data field in the arguments
2984 to pcre_exec() should be used to pass in a tables address if a value
2985 other than the default internal tables were used at compile time.
2986
2987 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2988 now diagnosed as an error. Previously, most of the time, a negative number
2989 would have been treated as zero, but if in addition "ovector" was passed as
2990 NULL, a crash could occur.
2991
2992 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2993 new versions from the libtool 1.5 distribution (the last one is a copy of
2994 a file called libtool.m4). This seems to have fixed the need to patch
2995 "configure" to support Darwin 1.3 (which I used to do). However, I still
2996 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2997 workstation).
2998
2999 24. Changed the PCRE licence to be the more standard "BSD" licence.
3000
3001
3002 Version 4.5 01-Dec-03
3003 ---------------------
3004
3005 1. There has been some re-arrangement of the code for the match() function so
3006 that it can be compiled in a version that does not call itself recursively.
3007 Instead, it keeps those local variables that need separate instances for
3008 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3009 needs to "recurse". Keeping track of where control must go is done by means
3010 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3011 hide most of the details from the main code, and operates only if
3012 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3013 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3014 operating.
3015
3016 To make it easier for callers to provide specially tailored get/free
3017 functions for this usage, two new functions, pcre_stack_malloc, and
3018 pcre_stack_free, are used. They are always called in strict stacking order,
3019 and the size of block requested is always the same.
3020
3021 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3022 PCRE has been compiled to use the stack or the heap for recursion. The
3023 -C option of pcretest uses this to show which version is compiled.
3024
3025 A new data escape \S, is added to pcretest; it causes the amounts of store
3026 obtained and freed by both kinds of malloc/free at match time to be added
3027 to the output.
3028
3029 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3030 what's available on my current Linux desktop machine.
3031
3032 3. When matching a UTF-8 string, the test for a valid string at the start has
3033 been extended. If start_offset is not zero, PCRE now checks that it points
3034 to a byte that is the start of a UTF-8 character. If not, it returns
3035 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3036 this is necessary because there may be backward assertions in the pattern.
3037 When matching the same subject several times, it may save resources to use
3038 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3039
3040 4. The code for checking the validity of UTF-8 strings has been tightened so
3041 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3042 containing "overlong sequences".
3043
3044 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3045 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3046 should have been "&", but it just so happened that all the cases this let
3047 through by mistake were picked up later in the function.
3048
3049 6. I had used a variable called "isblank" - this is a C99 function, causing
3050 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3051
3052 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3053 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3054 so the version is shown; (c) stop "make test" echoing "./RunTest".
3055
3056 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3057
3058 9. The prototype for memmove() for systems that don't have it was using
3059 size_t, but the inclusion of the header that defines size_t was later. I've
3060 moved the #includes for the C headers earlier to avoid this.
3061
3062 10. Added some adjustments to the code to make it easier to compiler on certain
3063 special systems:
3064
3065 (a) Some "const" qualifiers were missing.
3066 (b) Added the macro EXPORT before all exported functions; by default this
3067 is defined to be empty.
3068 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3069 that it reads its output file name as an argument instead of writing
3070 to the standard output and assuming this can be redirected.
3071
3072 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3073 class containing characters with values greater than 255, PCRE compilation
3074 went into a loop.
3075
3076 12. A recursive reference to a subpattern that was within another subpattern
3077 that had a minimum quantifier of zero caused PCRE to crash. For example,
3078 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3079 recursion. If the recursively-called subpattern itself had a zero repeat,
3080 that was OK.
3081
3082 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3083 buffer into which it was copied (for escape processing) was still set at
3084 1024, so long lines caused crashes.
3085
3086 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3087 "internal error: code overflow...". This applied to any character class
3088 that was followed by a possessive quantifier.
3089
3090 15. Modified the Makefile to add libpcre.la as a prerequisite for
3091 libpcreposix.la because I was told this is needed for a parallel build to
3092 work.
3093
3094 16. If a pattern that contained .* following optional items at the start was
3095 studied, the wrong optimizing data was generated, leading to matching
3096 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3097 matching string must start with a or b or c. The correct conclusion for
3098 this pattern is that a match can start with any character.
3099
3100
3101 Version 4.4 13-Aug-03
3102 ---------------------
3103
3104 1. In UTF-8 mode, a character class containing characters with values between
3105 127 and 255 was not handled correctly if the compiled pattern was studied.
3106 In fixing this, I have also improved the studying algorithm for such
3107 classes (slightly).
3108
3109 2. Three internal functions had redundant arguments passed to them. Removal
3110 might give a very teeny performance improvement.
3111
3112 3. Documentation bug: the value of the capture_top field in a callout is *one
3113 more than* the number of the hightest numbered captured substring.
3114
3115 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3116 in incorrectly linking with a previously installed version. They now link
3117 explicitly with libpcre.la.
3118
3119 5. configure.in no longer needs to recognize Cygwin specially.
3120
3121 6. A problem in pcre.in for Windows platforms is fixed.
3122
3123 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3124 pcretest, it used to include the size of the study block as part of its
3125 output. Unfortunately, the structure contains a field that has a different
3126 size on different hardware architectures. This meant that the tests that
3127 showed this size failed. As the block is currently always of a fixed size,
3128 this information isn't actually particularly useful in pcretest output, so
3129 I have just removed it.
3130
3131 8. Three pre-processor statements accidentally did not start in column 1.
3132 Sadly, there are *still* compilers around that complain, even though
3133 standard C has not required this for well over a decade. Sigh.
3134
3135 9. In pcretest, the code for checking callouts passed small integers in the
3136 callout_data field, which is a void * field. However, some picky compilers
3137 complained about the casts involved for this on 64-bit systems. Now
3138 pcretest passes the address of the small integer instead, which should get
3139 rid of the warnings.
3140
3141 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3142 both compile and run time, and gives an error if an invalid UTF-8 sequence
3143 is found. There is a option for disabling this check in cases where the
3144 string is known to be correct and/or the maximum performance is wanted.
3145
3146 11. In response to a bug report, I changed one line in Makefile.in from
3147
3148 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3149 to
3150 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3151
3152 to look similar to other lines, but I have no way of telling whether this
3153 is the right thing to do, as I do not use Windows. No doubt I'll get told
3154 if it's wrong...
3155
3156
3157 Version 4.3 21-May-03
3158 ---------------------
3159
3160 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3161 Makefile.
3162
3163 2. Some refactoring to improve the quality of the code:
3164
3165 (i) The utf8_table... variables are now declared "const".
3166
3167 (ii) The code for \cx, which used the "case flipping" table to upper case
3168 lower case letters, now just substracts 32. This is ASCII-specific,
3169 but the whole concept of \cx is ASCII-specific, so it seems
3170 reasonable.
3171
3172 (iii) PCRE was using its character types table to recognize decimal and
3173 hexadecimal digits in the pattern. This is silly, because it handles
3174 only 0-9, a-f, and A-F, but the character types table is locale-
3175 specific, which means strange things might happen. A private
3176 table is now used for this - though it costs 256 bytes, a table is
3177 much faster than multiple explicit tests. Of course, the standard
3178 character types table is still used for matching digits in subject
3179 strings against \d.
3180
3181 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3182 ending in _t are). So I've renamed it as ESC_tee.
3183
3184 3. The first argument for regexec() in the POSIX wrapper should have been
3185 defined as "const".
3186
3187 4. Changed pcretest to use malloc() for its buffers so that they can be
3188 Electric Fenced for debugging.
3189
3190 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3191 to read one or more bytes before the start of the subject string. Often this
3192 had no effect on PCRE's behaviour, but in some circumstances it could
3193 provoke a segmentation fault.
3194
3195 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3196 to try to read one or more bytes before the start of the subject string.
3197
3198 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3199 UTF-8 support could misbehave in various ways if the subject string
3200 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3201 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3202 back over UTF-8 characters.)
3203
3204
3205 Version 4.2 14-Apr-03
3206 ---------------------
3207
3208 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3209
3210 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3211 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3212 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3213 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3214 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3215 and BUILD_EXEEXT
3216 Note: automatic setting of the BUILD variables is not yet working
3217 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3218 compile-time but not at link-time
3219 [LINK]: use for linking executables only
3220 make different versions for Windows and non-Windows
3221 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3222 libraries
3223 [LINK_FOR_BUILD]: new variable
3224 [OBJEXT]: use throughout
3225 [EXEEXT]: use throughout
3226 <winshared>: new target
3227 <wininstall>: new target
3228 <dftables.o>: use native compiler
3229 <dftables>: use native linker
3230 <install>: handle Windows platform correctly
3231 <clean>: ditto
3232 <check>: ditto
3233 copy DLL to top builddir before testing
3234
3235 As part of these changes, -no-undefined was removed again. This was reported
3236 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3237 in any case.
3238
3239 3. Some tidies to get rid of compiler warnings:
3240
3241 . In the match_data structure, match_limit was an unsigned long int, whereas
3242 match_call_count was an int. I've made them both unsigned long ints.
3243
3244 . In pcretest the fact that a const uschar * doesn't automatically cast to
3245 a void * provoked a warning.
3246
3247 . Turning on some more compiler warnings threw up some "shadow" variables
3248 and a few more missing casts.
3249
3250 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3251 option, a class that contained a single character with a value between 128
3252 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3253
3254 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3255 option, a class that contained several characters, but with at least one
3256 whose value was between 128 and 255 caused PCRE to crash.
3257
3258
3259 Version 4.1 12-Mar-03
3260 ---------------------
3261
3262 1. Compiling with gcc -pedantic found a couple of places where casts were
3263 needed, and a string in dftables.c that was longer than standard compilers are
3264 required to support.
3265
3266 2. Compiling with Sun's compiler found a few more places where the code could
3267 be tidied up in order to avoid warnings.
3268
3269 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3270 first of these names is deprecated in the latest Autoconf in favour of the name
3271 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3272 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3273 analogy I have changed it to CFLAGS_FOR_BUILD.
3274
3275 4. Added -no-undefined to the linking command in the Makefile, because this is
3276 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3277 linking step for the pcreposix library.
3278
3279 5. PCRE was failing to diagnose the case of two named groups with the same
3280 name.
3281
3282 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3283 literal character that is needed in the subject for a match, and scans along to
3284 ensure that it is present before embarking on the full matching process. This
3285 saves time in cases of nested unlimited repeats that are never going to match.
3286 Problem: the scan can take a lot of time if the subject is very long (e.g.
3287 megabytes), thus penalizing straightforward matches. It is now done only if the
3288 amount of subject to be scanned is less than 1000 bytes.
3289
3290 7. A lesser problem with the same optimization is that it was recording the
3291 first character of an anchored pattern as "needed", thus provoking a search
3292 right along the subject, even when the first match of the pattern was going to
3293 fail. The "needed" character is now not set for anchored patterns, unless it
3294 follows something in the pattern that is of non-fixed length. Thus, it still
3295 fulfils its original purpose of finding quick non-matches in cases of nested
3296 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3297
3298
3299 Version 4.0 17-Feb-03
3300 ---------------------
3301
3302 1. If a comment in an extended regex that started immediately after a meta-item
3303 extended to the end of string, PCRE compiled incorrect data. This could lead to
3304 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3305
3306 2. Moved to autoconf 2.53 and libtool 1.4.2.
3307
3308 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3309 the special perltest8 script is no longer needed - all the tests can be run
3310 from a single perltest script.
3311
3312 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3313 by \s. It has now been removed in PCRE. This means it isn't recognized as
3314 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3315 class [:space:] *does* include VT, thereby creating a mess.
3316
3317 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3318 space and tab.
3319
3320 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3321 its new features into the main test script, reducing the number of scripts.
3322
3323 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3324 were backward compatible, and made the (?i) apply to the whole pattern, as if
3325 /i were given. Now it behaves more logically, and applies the option setting
3326 only to what follows. PCRE has been changed to follow suit. However, if it
3327 finds options settings right at the start of the pattern, it extracts them into
3328 the global options, as before. Thus, they show up in the info data.
3329
3330 8. Added support for the \Q...\E escape sequence. Characters in between are
3331 treated as literals. This is slightly different from Perl in that $ and @ are
3332 also handled as literals inside the quotes. In Perl, they will cause variable
3333 interpolation. Note the following examples:
3334
3335 Pattern PCRE matches Perl matches
3336
3337 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3338 \Qabc\$xyz\E abc\$xyz abc\$xyz
3339 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3340
3341 For compatibility with Perl, \Q...\E sequences are recognized inside character
3342 classes as well as outside them.
3343
3344 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3345 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3346 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3347 signed/unsigned warnings.
3348
3349 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3350 option for pcretest, so I've replaced it by a simple function that does just
3351 that job.
3352
3353 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3354 "pcregrep -".
3355
3356 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3357 Java package. This provides some syntactic sugar for simple cases of what my
3358 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3359 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3360 item, you can use this simplified notation. Note that only makes sense with
3361 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3362 greediness, whatever the setting of the PCRE_UNGREEDY option.
3363
3364 13. A change of greediness default within a pattern was not taking effect at
3365 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3366 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3367 was abstracted outside.
3368
3369 14. PCRE now supports the \G assertion. It is true when the current matching
3370 position is at the start point of the match. This differs from \A when the
3371 starting offset is non-zero. Used with the /g option of pcretest (or similar
3372 code), it works in the same way as it does for Perl's /g option. If all
3373 alternatives of a regex begin with \G, the expression is anchored to the start
3374 match position, and the "anchored" flag is set in the compiled expression.
3375
3376 15. Some bugs concerning the handling of certain option changes within patterns
3377 have been fixed. These applied to options other than (?ims). For example,
3378 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3379 been the other way round. Some of this was related to change 7 above.
3380
3381 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3382 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3383 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3384 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3385
3386 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3387 mode. Unlike ".", it always matches newline, whatever the setting of
3388 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3389 assertions. Perl allows it, but it doesn't (in general) work because it can't
3390 calculate the length of the lookbehind. At least, that's the case for Perl
3391 5.8.0 - I've been told they are going to document that it doesn't work in
3392 future.
3393
3394 18. Added an error diagnosis for escapes that PCRE does not support: these are
3395 \L, \l, \N, \P, \p, \U, \u, and \X.
3396
3397 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3398 reading past the end of the pattern in cases such as /[abcd/.
3399
3400 20. PCRE was getting more memory than necessary for patterns with classes that
3401 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3402
3403 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3404 compiling PCRE for use with Virtual Pascal.
3405
3406 22. Small fix to the Makefile to make it work properly if the build is done
3407 outside the source tree.
3408
3409 23. Added a new extension: a condition to go with recursion. If a conditional
3410 subpattern starts with (?(R) the "true" branch is used if recursion has
3411 happened, whereas the "false" branch is used only at the top level.
3412
3413 24. When there was a very long string of literal characters (over 255 bytes
3414 without UTF support, over 250 bytes with UTF support), the computation of how
3415 much memory was required could be incorrect, leading to segfaults or other
3416 strange effects.
3417
3418 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3419 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3420 there was a subsequent back reference to those brackets. This meant that, for
3421 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3422 possible to check for precisely this case. All we can do is abandon the
3423 optimization if .* occurs inside capturing brackets when there are any back
3424 references whatsoever. (See below for a better fix that came later.)
3425
3426 26. The handling of the optimization for finding the first character of a
3427 non-anchored pattern, and for finding a character that is required later in the
3428 match were failing in some cases. This didn't break the matching; it just
3429 failed to optimize when it could. The way this is done has been re-implemented.
3430
3431 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3432
3433 28. Added a new feature that provides some of the functionality that Perl
3434 provides with (?{...}). The facility is termed a "callout". The way it is done
3435 in PCRE is for the caller to provide an optional function, by setting
3436 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3437 global variable. By default it is unset, which disables all calling out. To get
3438 the function called, the regex must include (?C) at appropriate points. This
3439 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3440 This provides a means of identifying different callout points. When PCRE
3441 reaches such a point in the regex, if pcre_callout has been set, the external
3442 function is called. It is provided with data in a structure called
3443 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3444 matching continues; if it returns a non-zero value, the match at the current
3445 point fails. However, backtracking will occur if possible. [This was changed
3446 later and other features added - see item 49 below.]
3447
3448 29. pcretest is upgraded to test the callout functionality. It provides a
3449 callout function that displays information. By default, it shows the start of
3450 the match and the current position in the text. There are some new data escapes
3451 to vary what happens:
3452
3453 \C+ in addition, show current contents of captured substrings
3454 \C- do not supply a callout function
3455 \C!n return 1 when callout number n is reached
3456 \C!n!m return 1 when callout number n is reached for the mth time
3457
3458 30. If pcregrep was called with the -l option and just a single file name, it
3459 output "<stdin>" if a match was found, instead of the file name.
3460
3461 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3462 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3463 pcre_exec(). This saves a malloc/free per call. The default value of
3464 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3465 when configuring.
3466
3467 32. The default maximum size of a compiled pattern is 64K. There have been a
3468 few cases of people hitting this limit. The code now uses macros to handle the
3469 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3470 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3471 configuring. Tests 2 and 5 work only with 2-byte links because they output
3472 debugging information about compiled patterns.
3473
3474 33. Internal code re-arrangements:
3475
3476 (a) Moved the debugging function for printing out a compiled regex into
3477 its own source file (printint.c) and used #include to pull it into
3478 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3479 separate copies.
3480
3481 (b) Defined the list of op-code names for debugging as a macro in
3482 internal.h so that it is next to the definition of the opcodes.
3483
3484 (c) Defined a table of op-code lengths for simpler skipping along compiled
3485 code. This is again a macro in internal.h so that it is next to the
3486 definition of the opcodes.
3487
3488 34. Added support for recursive calls to individual subpatterns, along the
3489 lines of Robin Houston's patch (but implemented somewhat differently).
3490
3491 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3492 allow it to read and process whole directories in Win32. This code was
3493 contributed by Lionel Fourquaux; it has not been tested by me.
3494
3495 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3496 used to name a group. Names consist of alphanumerics and underscores, and must
3497 be unique. Back references use the syntax (?P=name) and recursive calls use
3498 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3499 numbers. The function pcre_fullinfo() can be used after compilation to extract
3500 a name/number map. There are three relevant calls:
3501
3502 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3503 PCRE_INFO_NAMECOUNT yields the number of entries
3504 PCRE_INFO_NAMETABLE yields a pointer to the map.
3505
3506 The map is a vector of fixed-size entries. The size of each entry depends on
3507 the length of the longest name used. The first two bytes of each entry are the
3508 group number, most significant byte first. There follows the corresponding
3509 name, zero terminated. The names are in alphabetical order.
3510
3511 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3512 case instead of 255. Making it the same both with and without UTF-8 support
3513 means that the same test output works with both.
3514
3515 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3516 calling malloc() with a zero argument.
3517
3518 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3519 optimization. I've improved things by keeping a bitmap of backreferences with
3520 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3521 fact referenced, the optimization can be applied. It is unlikely that a
3522 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3523 the match to follow \n) will appear inside brackets with a number greater than
3524 31, but if it does, any back reference > 31 suppresses the optimization.
3525
3526 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3527 of disabling numbered capturing parentheses. Any opening parenthesis that is
3528 not followed by ? behaves as if it were followed by ?: but named parentheses
3529 can still be used for capturing (and they will acquire numbers in the usual
3530 way).
3531
3532 41. Redesigned the return codes from the match() function into yes/no/error so
3533 that errors can be passed back from deep inside the nested calls. A malloc
3534 failure while inside a recursive subpattern call now causes the
3535 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3536
3537 42. It is now possible to set a limit on the number of times the match()
3538 function is called in a call to pcre_exec(). This facility makes it possible to
3539 limit the amount of recursion and backtracking, though not in a directly
3540 obvious way, because the match() function is used in a number of different
3541 circumstances. The count starts from zero for each position in the subject
3542 string (for non-anchored patterns). The default limit is, for compatibility, a
3543 large number, namely 10 000 000. You can change this in two ways:
3544
3545 (a) When configuring PCRE before making, you can use --with-match-limit=n
3546 to set a default value for the compiled library.
3547
3548 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3549 a different value is set. See 45 below.
3550
3551 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3552
3553 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3554 of things that can be changed at compile time. The first argument specifies
3555 what is wanted and the second points to where the information is to be placed.
3556 The current list of available information is:
3557
3558 PCRE_CONFIG_UTF8
3559
3560 The output is an integer that is set to one if UTF-8 support is available;
3561 otherwise it is set to zero.
3562
3563 PCRE_CONFIG_NEWLINE
3564
3565 The output is an integer that it set to the value of the code that is used for
3566 newline. It is either LF (10) or CR (13).
3567
3568 PCRE_CONFIG_LINK_SIZE
3569
3570 The output is an integer that contains the number of bytes used for internal
3571 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3572
3573 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3574
3575 The output is an integer that contains the threshold above which the POSIX
3576 interface uses malloc() for output vectors. See item 31 above.
3577
3578 PCRE_CONFIG_MATCH_LIMIT
3579
3580 The output is an unsigned integer that contains the default limit of the number
3581 of match() calls in a pcre_exec() execution. See 42 above.
3582
3583 44. pcretest has been upgraded by the addition of the -C option. This causes it
3584 to extract all the available output from the new pcre_config() function, and to
3585 output it. The program then exits immediately.
3586
3587 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3588 order to support additional features. One way would have been to define
3589 pcre_exec2() (for example) with extra arguments, but this would not have been
3590 extensible, and would also have required all calls to the original function to
3591 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3592 is used for passing in "extra" data from pcre_study().
3593
3594 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3595 contains the following fields:
3596
3597 flags a bitmap indicating which of the following fields are set
3598 study_data opaque data from pcre_study()
3599 match_limit a way of specifying a limit on match() calls for a specific
3600 call to pcre_exec()
3601 callout_data data for callouts (see 49 below)
3602
3603 The flag bits are also defined in pcre.h, and are
3604
3605 PCRE_EXTRA_STUDY_DATA
3606 PCRE_EXTRA_MATCH_LIMIT
3607 PCRE_EXTRA_CALLOUT_DATA
3608
3609 The pcre_study() function now returns one of these new pcre_extra blocks, with
3610 the actual study data pointed to by the study_data field, and the
3611 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3612 before. That is, this change is entirely upwards-compatible and requires no
3613 change to existing code.
3614
3615 If you want to pass in additional data to pcre_exec(), you can either place it
3616 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3617 block.
3618
3619 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3620 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3621 times with different match limits, until it finds the minimum value needed for
3622 pcre_exec() to complete. The value is then output. This can be instructive; for
3623 most simple matches the number is quite small, but for pathological cases it
3624 gets very large very quickly.
3625
3626 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3627 returns the size of the data block pointed to by the study_data field in a
3628 pcre_extra block, that is, the value that was passed as the argument to
3629 pcre_malloc() when PCRE was getting memory in which to place the information
3630 created by pcre_study(). The fourth argument should point to a size_t variable.
3631 pcretest has been extended so that this information is shown after a successful
3632 pcre_study() call when information about the compiled regex is being displayed.
3633
3634 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3635 because what follows is always an absolute path. (Later: it turns out that this
3636 is more than cosmetic for MinGW, because it doesn't like empty path
3637 components.)
3638
3639 49. Some changes have been made to the callout feature (see 28 above):
3640
3641 (i) A callout function now has three choices for what it returns:
3642
3643 0 => success, carry on matching
3644 > 0 => failure at this point, but backtrack if possible
3645 < 0 => serious error, return this value from pcre_exec()
3646
3647 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3648 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3649 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3650 use by callout functions. It will never be used by PCRE itself.
3651
3652 (ii) The pcre_extra structure (see 45 above) has a void * field called
3653 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3654 pcre_callout_block structure has a field of the same name. The contents of
3655 the field passed in the pcre_extra structure are passed to the callout
3656 function in the corresponding field in the callout block. This makes it
3657 easier to use the same callout-containing regex from multiple threads. For
3658 testing, the pcretest program has a new data escape
3659
3660 \C*n pass the number n (may be negative) as callout_data
3661
3662 If the callout function in pcretest receives a non-zero value as
3663 callout_data, it returns that value.
3664
3665 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3666 there were some redundant $(CFLAGS) in commands that are now specified as
3667 $(LINK), which already includes $(CFLAGS).
3668
3669 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3670 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3671 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3672 one-byte characters throughout. Note that case-insensitive matching applies
3673 only to characters whose values are less than 256. PCRE doesn't support the
3674 notion of cases for higher-valued characters.
3675
3676 (i) A character class whose characters are all within 0-255 is handled as
3677 a bit map, and the map is inverted for negative classes. Previously, a
3678 character > 255 always failed to match such a class; however it should
3679 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3680
3681 (ii) A negated character class with a single character < 255 is coded as
3682 "not this character" (OP_NOT). This wasn't working properly when the test
3683 character was multibyte, either singly or repeated.
3684
3685 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3686 mode, for example: \x{100}{2,3}.
3687
3688 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3689 singly or repeated) now correctly test multibyte characters. However,
3690 PCRE doesn't recognize any characters with values greater than 255 as
3691 digits, spaces, or word characters. Such characters always match \D, \S,
3692 and \W, and never match \d, \s, or \w.
3693
3694 (v) Classes may now contain characters and character ranges with values
3695 greater than 255. For example: [ab\x{100}-\x{400}].
3696
3697 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3698 PCRE in UTF-8 mode.
3699
3700 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3701 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3702 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3703 value.)
3704
3705 53. The single man page has become too large. I have therefore split it up into
3706 a number of separate man pages. These also give rise to individual HTML pages;
3707 these are now put in a separate directory, and there is an index.html page that
3708 lists them all. Some hyperlinking between the pages has been installed.
3709
3710 54. Added convenience functions for handling named capturing parentheses.
3711
3712 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3713 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3714 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3715 are faulted.
3716
3717 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3718 calling configure. These values are used when compiling the dftables.c program
3719 which is run to generate the source of the default character tables. They
3720 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3721 you will need to set these values.
3722
3723 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3724
3725
3726 Version 3.9 02-Jan-02
3727 ---------------------
3728
3729 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3730
3731 2. If --disable-static was given, the building process failed when trying to
3732 build pcretest and pcregrep. (For some reason it was using libtool to compile
3733 them, which is not right, as they aren't part of the library.)
3734
3735
3736 Version 3.8 18-Dec-01
3737 ---------------------
3738
3739 1. The experimental UTF-8 code was completely screwed up. It was packing the
3740 bytes in the wrong order. How dumb can you get?
3741
3742
3743 Version 3.7 29-Oct-01
3744 ---------------------
3745
3746 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3747 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3748 this didn't happen under Solaris 8, where I normally test things.
3749
3750 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3751 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3752 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3753 this fix an hour or so after the initial 3.7 release.)
3754
3755
3756 Version 3.6 23-Oct-01
3757 ---------------------
3758
3759 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3760 offsets passed as NULL with zero offset count.
3761
3762 2. The config.guess and config.sub files had not been updated when I moved to
3763 the latest autoconf.
3764
3765
3766 Version 3.5 15-Aug-01
3767 ---------------------
3768
3769 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3770 had been forgotten.
3771
3772 2. By using declared but undefined structures, we can avoid using "void"
3773 definitions in pcre.h while keeping the internal definitions of the structures
3774 private.
3775
3776 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3777 user point of view, this means that both static and shared libraries are built
3778 by default, but this can be individually controlled. More of the work of
3779 handling this static/shared cases is now inside libtool instead of PCRE's make
3780 file.
3781
3782 4. The pcretest utility is now installed along with pcregrep because it is
3783 useful for users (to test regexs) and by doing this, it automatically gets
3784 relinked by libtool. The documentation has been turned into a man page, so
3785 there are now .1, .txt, and .html versions in /doc.
3786
3787 5. Upgrades to pcregrep:
3788 (i) Added long-form option names like gnu grep.
3789 (ii) Added --help to list all options with an explanatory phrase.
3790 (iii) Added -r, --recursive to recurse into sub-directories.
3791 (iv) Added -f, --file to read patterns from a file.
3792
3793 6. pcre_exec() was referring to its "code" argument before testing that
3794 argument for NULL (and giving an error if it was NULL).
3795
3796 7. Upgraded Makefile.in to allow for compiling in a different directory from
3797 the source directory.
3798
3799 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3800 options bits, the pointer it was passed was to an int instead of to an unsigned
3801 long int. This mattered only on 64-bit systems.
3802
3803 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3804 generated) instead of pcre.in, which it its source. Also made the same change
3805 in several of the .c files.
3806
3807 10. A new release of gcc defines printf() as a macro, which broke pcretest
3808 because it had an ifdef in the middle of a string argument for printf(). Fixed
3809 by using separate calls to printf().
3810
3811 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3812 script, to force use of CR or LF instead of \n in the source. On non-Unix
3813 systems, the value can be set in config.h.
3814
3815 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3816 absolute limit. Changed the text of the error message to make this clear, and
3817 likewise updated the man page.
3818
3819 13. The limit of 99 on the number of capturing subpatterns has been removed.
3820 The new limit is 65535, which I hope will not be a "real" limit.
3821
3822
3823 Version 3.4 22-Aug-00
3824 ---------------------
3825
3826 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3827
3828 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3829
3830
3831 Version 3.3 01-Aug-00
3832 ---------------------
3833
3834 1. If an octal character was given, but the value was greater than \377, it
3835 was not getting masked to the least significant bits, as documented. This could
3836 lead to crashes in some systems.
3837
3838 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3839 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3840
3841 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3842 These just pass their arguments on to (pcre_free)(), but they are provided
3843 because some uses of PCRE bind it to non-C systems that can call its functions,
3844 but cannot call free() or pcre_free() directly.
3845
3846 4. Add "make test" as a synonym for "make check". Corrected some comments in
3847 the Makefile.
3848
3849 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3850 Makefile.
3851
3852 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3853 command called pgrep for grepping around the active processes.
3854
3855 7. Added the beginnings of support for UTF-8 character strings.
3856
3857 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3858 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3859 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3860 out for the ar command.)
3861
3862
3863 Version 3.2 12-May-00
3864 ---------------------
3865
3866 This is purely a bug fixing release.
3867
3868 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3869 of ZA. This was just one example of several cases that could provoke this bug,
3870 which was introduced by change 9 of version 2.00. The code for breaking
3871 infinite loops after an iteration that matches an empty string was't working
3872 correctly.
3873
3874 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3875 when matched against abbab (for example). After matching an empty string, it
3876 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3877 caused it to match further down the string than it should.
3878
3879 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3880 was there because it doesn't seem to be needed, and it causes trouble on some
3881 systems, as it is not a Standard C header. It has been removed.
3882
3883 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3884 were reported on the Macintosh. The changes were from
3885
3886 while ((c = *(++ptr)) != 0 && c != '\n');
3887 to
3888 while ((c = *(++ptr)) != 0 && c != '\n') ;
3889
3890 Totally extraordinary, but if that's what it takes...
3891
3892 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3893 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3894 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3895 assumes the way PCRE uses memmove() (always moving upwards).
3896
3897 6. PCRE is being used in one environment where strchr() is not available. There
3898 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3899 faster code anyway.
3900
3901
3902 Version 3.1 09-Feb-00
3903 ---------------------
3904
3905 The only change in this release is the fixing of some bugs in Makefile.in for
3906 the "install" target:
3907
3908 (1) It was failing to install pcreposix.h.
3909
3910 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3911
3912
3913 Version 3.0 01-Feb-00
3914 ---------------------
3915
3916 1. Add support for the /+ modifier to perltest (to output $` like it does in
3917 pcretest).
3918
3919 2. Add support for the /g modifier to perltest.
3920
3921 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3922 matches null strings.
3923
3924 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3925 pattern. Perl treats empty patterns specially - it reuses the most recent
3926 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3927 effect.
3928
3929 5. The POSIX interface was broken in that it was just handing over the POSIX
3930 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3931 required a bigger vector, with some working space on the end. This means that
3932 the POSIX wrapper now has to get and free some memory, and copy the results.
3933
3934 6. Added some simple autoconf support, placing the test data and the
3935 documentation in separate directories, re-organizing some of the
3936 information files, and making it build pcre-config (a GNU standard). Also added
3937 libtool support for building PCRE as a shared library, which is now the
3938 default.
3939
3940 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3941 09 are not valid octal constants. Single digits will be used for minor values
3942 less than 10.
3943
3944 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3945 existing programs that set these in the POSIX interface can use PCRE without
3946 modification.
3947
3948 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3949 return all that pcre_info() returns, plus additional data. The pcre_info()
3950 function is retained for compatibility, but is considered to be obsolete.
3951
3952 10. Added experimental recursion feature (?R) to handle one common case that
3953 Perl 5.6 will be able to do with (?p{...}).
3954
3955 11. Added support for POSIX character classes like [:alpha:], which Perl is
3956 adopting.
3957
3958
3959 Version 2.08 31-Aug-99
3960 ----------------------
3961
3962 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3963 trying to match at the startoffset position, but instead was moving forward to
3964 the next newline as if a previous match had failed.
3965
3966 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3967 and could get into a loop if a null string was matched other than at the start
3968 of the subject.
3969
3970 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3971 be distinguished at compile time, and for completeness also added PCRE_DATE.
3972
3973 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3974 in GnuWin32 environments.
3975
3976
3977 Version 2.07 29-Jul-99
3978 ----------------------
3979
3980 1. The documentation is now supplied in plain text form and HTML as well as in
3981 the form of man page sources.
3982
3983 2. C++ compilers don't like assigning (void *) values to other pointer types.
3984 In particular this affects malloc(). Although there is no problem in Standard
3985 C, I've put in casts to keep C++ compilers happy.
3986
3987 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3988 should be (const char *).
3989
3990 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3991 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3992 However, I haven't made this a standard facility. The documentation doesn't
3993 mention it, and the Makefile doesn't support it.
3994
3995 5. The Makefile now contains an "install" target, with editable destinations at
3996 the top of the file. The pcretest program is not installed.
3997
3998 6. pgrep -V now gives the PCRE version number and date.
3999
4000 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4001 causing the entire string to be ignored, instead of just the last character.
4002
4003 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4004 non-matching string, it can take a very, very long time, even for strings of
4005 quite modest length, because of the nested recursion. PCRE now does better in
4006 some of these cases. It does this by remembering the last required literal
4007 character in the pattern, and pre-searching the subject to ensure it is present
4008 before running the real match. In other words, it applies a heuristic to detect
4009 some types of certain failure quickly, and in the above example, if presented
4010 with a string that has no trailing " it gives "no match" very quickly.
4011
4012 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4013 other alternatives are tried instead.
4014
4015
4016 Version 2.06 09-Jun-99
4017 ----------------------
4018
4019 1. Change pcretest's output for amount of store used to show just the code
4020 space, because the remainder (the data block) varies in size between 32-bit and
4021 64-bit systems.
4022
4023 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4024 start matching at. This allows lookbehinds to work when searching for multiple
4025 occurrences in a string.
4026
4027 3. Added additional options to pcretest for testing multiple occurrences:
4028
4029 /+ outputs the rest of the string that follows a match
4030 /g loops for multiple occurrences, using the new startoffset argument
4031 /G loops for multiple occurrences by passing an incremented pointer
4032
4033 4. PCRE wasn't doing the "first character" optimization for patterns starting
4034 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4035 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4036 the letter 'x'. On long subject strings, this gives a significant speed-up.
4037
4038
4039 Version 2.05 21-Apr-99
4040 ----------------------
4041
4042 1. Changed the type of magic_number from int to long int so that it works
4043 properly on 16-bit systems.
4044
4045 2. Fixed a bug which caused patterns starting with .* not to work correctly
4046 when the subject string contained newline characters. PCRE was assuming
4047 anchoring for such patterns in all cases, which is not correct because .* will
4048 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4049 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4050 must be retried after every newline in the subject.
4051
4052
4053 Version 2.04 18-Feb-99
4054 ----------------------
4055
4056 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4057 computation of the store needed to hold the pattern was incorrect (too large).
4058 If such patterns were nested a few deep, this could multiply and become a real
4059 problem.
4060
4061 2. Added /M option to pcretest to show the memory requirement of a specific
4062 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4063
4064 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4065 compiled in such a way that the backtracking after subsequent failure was
4066 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4067 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4068
4069
4070 Version 2.03 02-Feb-99
4071 ----------------------
4072
4073 1. Fixed typo and small mistake in man page.
4074
4075 2. Added 4th condition (GPL supersedes if conflict) and created separate
4076 LICENCE file containing the conditions.
4077
4078 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4079 Perl, that is the internal \ allows the delimiter to be included in the
4080 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4081 the final delimiter, add \ to the end of the pattern (to test the error).
4082
4083 4. Added the convenience functions for extracting substrings after a successful
4084 match. Updated pcretest to make it able to test these functions.
4085
4086
4087 Version 2.02 14-Jan-99
4088 ----------------------
4089
4090 1. Initialized the working variables associated with each extraction so that
4091 their saving and restoring doesn't refer to uninitialized store.
4092
4093 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4094 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4095 fix the problem.
4096
4097 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4098 calls, and wasn't printing the correct value for compiling calls. Increased the
4099 default value of LOOPREPEAT, and the number of significant figures in the
4100 times.
4101
4102 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4103
4104 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4105 a building problem on Windows NT with a FAT file system.
4106
4107
4108 Version 2.01 21-Oct-98
4109 ----------------------
4110
4111 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4112 to character tables built by pcre_maketables() in the current locale. If NULL
4113 is passed, the default tables are used.
4114
4115
4116 Version 2.00 24-Sep-98
4117 ----------------------
4118
4119 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4120 it any more.
4121
4122 2. Allow quantification of (?>) groups, and make it work correctly.
4123
4124 3. The first character computation wasn't working for (?>) groups.
4125
4126 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4127 end of the subject) and add 5.005's \z, which really does match only at the
4128 very end of the subject.
4129
4130 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4131
4132 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4133 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4134 localized options. All options to pcre_study() were also removed.
4135
4136 7. Add other new features from 5.005:
4137
4138 $(?<= positive lookbehind
4139 $(?<! negative lookbehind
4140 (?imsx-imsx) added the unsetting capability
4141 such a setting is global if at outer level; local otherwise
4142 (?imsx-imsx:) non-capturing groups with option setting
4143 (?(cond)re|re) conditional pattern matching
4144
4145 A backreference to itself in a repeated group matches the previous
4146 captured string.
4147
4148 8. General tidying up of studying (both automatic and via "study")
4149 consequential on the addition of new assertions.
4150
4151 9. As in 5.005, unlimited repeated groups that could match an empty substring
4152 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4153 runtime if any iteration does actually match an empty substring.
4154
4155 10. Include the RunTest script in the distribution.
4156
4157 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4158 discrepancies, some of which were old and were also with respect to 5.004. They
4159 have now been fixed.
4160
4161
4162 Version 1.09 28-Apr-98
4163 ----------------------
4164
4165 1. A negated single character class followed by a quantifier with a minimum
4166 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4167 program crashes, or just wrong answers. This did not apply to negated classes
4168 containing more than one character, or to minima other than one.
4169
4170
4171 Version 1.08 27-Mar-98
4172 ----------------------
4173
4174 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4175
4176 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4177 latter must appear before anything that relies on it in the pattern.
4178
4179
4180 Version 1.07 16-Feb-98
4181 ----------------------
4182
4183 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4184 repeat of a potentially empty string).
4185
4186
4187 Version 1.06 23-Jan-98
4188 ----------------------
4189
4190 1. Added Markus Oberhumer's little patches for C++.
4191
4192 2. Literal strings longer than 255 characters were broken.
4193
4194
4195 Version 1.05 23-Dec-97
4196 ----------------------
4197
4198 1. Negated character classes containing more than one character were failing if
4199 PCRE_CASELESS was set at run time.
4200
4201
4202 Version 1.04 19-Dec-97
4203 ----------------------
4204
4205 1. Corrected the man page, where some "const" qualifiers had been omitted.
4206
4207 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4208 input syntax.
4209
4210 3. Fixed memory leak which occurred when a regex with back references was
4211 matched with an offsets vector that wasn't big enough. The temporary memory
4212 that is used in this case wasn't being freed if the match failed.
4213
4214 4. Tidied pcretest to ensure it frees memory that it gets.
4215
4216 5. Temporary memory was being obtained in the case where the passed offsets
4217 vector was exactly big enough.
4218
4219 6. Corrected definition of offsetof() from change 5 below.
4220
4221 7. I had screwed up change 6 below and broken the rules for the use of
4222 setjmp(). Now fixed.
4223
4224
4225 Version 1.03 18-Dec-97
4226 ----------------------
4227
4228 1. A erroneous regex with a missing opening parenthesis was correctly
4229 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4230 on some systems.
4231
4232 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4233 it was reported that one broken compiler failed on the former because "code" is
4234 also an independent variable.
4235
4236 3. The erroneous regex a[]b caused an array overrun reference.
4237
4238 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4239 fail on data ending with that character. (It was going on too far, and checking
4240 the next character, typically a binary zero.) This was specific to the
4241 optimized code for single-character negative classes.
4242
4243 5. Added a contributed patch from the TIN world which does the following:
4244
4245 + Add an undef for memmove, in case the the system defines a macro for it.
4246
4247 + Add a definition of offsetof(), in case there isn't one. (I don't know
4248 the reason behind this - offsetof() is part of the ANSI standard - but
4249 it does no harm).
4250
4251 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4252 most of the places where whitespace preceded '#'. I have given up and
4253 allowed the remaining 2 cases to be at the margin.
4254
4255 + Rename some variables in pcre to eliminate shadowing. This seems very
4256 pedantic, but does no harm, of course.
4257
4258 6. Moved the call to setjmp() into its own function, to get rid of warnings
4259 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4260
4261 7. Constructs such as \d{8,} were compiling into the equivalent of
4262 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4263 outcome, but in this particular case used more store than had been allocated,
4264 which caused the bug to be discovered because it threw up an internal error.
4265
4266 8. The debugging code in both pcre and pcretest for outputting the compiled
4267 form of a regex was going wrong in the case of back references followed by
4268 curly-bracketed repeats.
4269
4270
4271 Version 1.02 12-Dec-97
4272 ----------------------
4273
4274 1. Typos in pcre.3 and comments in the source fixed.
4275
4276 2. Applied a contributed patch to get rid of places where it used to remove
4277 'const' from variables, and fixed some signed/unsigned and uninitialized
4278 variable warnings.
4279
4280 3. Added the "runtest" target to Makefile.
4281
4282 4. Set default compiler flag to -O2 rather than just -O.
4283
4284
4285 Version 1.01 19-Nov-97
4286 ----------------------
4287
4288 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4289 like /([ab]*)*/, that is, for classes with more than one character in them.
4290
4291 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4292 as /((?>a*))*/ (a PCRE_EXTRA facility).
4293
4294
4295 Version 1.00 18-Nov-97
4296 ----------------------
4297
4298 1. Added compile-time macros to support systems such as SunOS4 which don't have
4299 memmove() or strerror() but have other things that can be used instead.
4300
4301 2. Arranged that "make clean" removes the executables.
4302
4303
4304 Version 0.99 27-Oct-97
4305 ----------------------
4306
4307 1. Fixed bug in code for optimizing classes with only one character. It was
4308 initializing a 32-byte map regardless, which could cause it to run off the end
4309 of the memory it had got.
4310
4311 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4312
4313
4314 Version 0.98 22-Oct-97
4315 ----------------------
4316
4317 1. Fixed bug in code for handling temporary memory usage when there are more
4318 back references than supplied space in the ovector. This could cause segfaults.
4319
4320
4321 Version 0.97 21-Oct-97
4322 ----------------------
4323
4324 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4325
4326 2. Optimized negated single characters not to use a bit map.
4327
4328 3. Brought error texts together as macro definitions; clarified some of them;
4329 fixed one that was wrong - it said "range out of order" when it meant "invalid
4330 escape sequence".
4331
4332 4. Changed some char * arguments to const char *.
4333
4334 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4335
4336 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4337 pcretest.
4338
4339
4340 Version 0.96 16-Oct-97
4341 ----------------------
4342
4343 1. Added a simple "pgrep" utility to the distribution.
4344
4345 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4346 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4347 where "ddd" means "one or more decimal digits".
4348
4349 3. Fixed serious bug. If a pattern had a back reference, but the call to
4350 pcre_exec() didn't supply a large enough ovector to record the related
4351 identifying subpattern, the match always failed. PCRE now remembers the number
4352 of the largest back reference, and gets some temporary memory in which to save
4353 the offsets during matching if necessary, in order to ensure that
4354 backreferences always work.
4355
4356 4. Increased the compatibility with Perl in a number of ways:
4357
4358 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4359 to request this handling. The option can be set at compile or exec time.
4360
4361 (b) $ matches before a terminating newline by default; an option
4362 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4363 mode). The option can be set at compile or exec time.
4364
4365 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4366 the same as Perl's. If the decimal number it represents is less than 10
4367 or there aren't that many previous left capturing parentheses, an octal
4368 escape is read. Inside a character class, it's always an octal escape,
4369 even if it is a single digit.
4370
4371 (d) An escaped but undefined alphabetic character is taken as a literal,
4372 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4373 escapes.
4374
4375 (e) {0} is now permitted. (The previous item is removed from the compiled
4376 pattern).
4377
4378 5. Changed all the names of code files so that the basic parts are no longer
4379 than 10 characters, and abolished the teeny "globals.c" file.
4380
4381 6. Changed the handling of character classes; they are now done with a 32-byte
4382 bit map always.
4383
4384 7. Added the -d and /D options to pcretest to make it possible to look at the
4385 internals of compilation without having to recompile pcre.
4386
4387
4388 Version 0.95 23-Sep-97
4389 ----------------------
4390
4391 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4392 \x20 at the start of a run of normal characters. These were being treated as
4393 real characters, instead of the source characters being re-checked.
4394
4395
4396 Version 0.94 18-Sep-97
4397 ----------------------
4398
4399 1. The functions are now thread-safe, with the caveat that the global variables
4400 containing pointers to malloc() and free() or alternative functions are the
4401 same for all threads.
4402
4403 2. Get pcre_study() to generate a bitmap of initial characters for non-
4404 anchored patterns when this is possible, and use it if passed to pcre_exec().
4405
4406
4407 Version 0.93 15-Sep-97
4408 ----------------------
4409
4410 1. /(b)|(:+)/ was computing an incorrect first character.
4411
4412 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4413 but not actually doing anything yet.
4414
4415 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4416 as Perl does (e.g. [-az] or [az-]).
4417
4418 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4419 all possible positions.
4420
4421 5. Split up into different modules to avoid including unneeded functions in a
4422 compiled binary. However, compile and exec are still in one module. The "study"
4423 function is split off.
4424
4425 6. The character tables are now in a separate module whose source is generated
4426 by an auxiliary program - but can then be edited by hand if required. There are
4427 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4428 toupper() in the code.
4429
4430 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4431 make them global. Abolish the function for setting them, as the caller can now
4432 set them directly.
4433
4434
4435 Version 0.92 11-Sep-97
4436 ----------------------
4437
4438 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4439 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4440
4441 2. Caseless matching was not working in character classes if the characters in
4442 the pattern were in upper case.
4443
4444 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4445
4446 4. Make PCRE_ANCHORED public and accept as a compile option.
4447
4448 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4449 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4450 pass them.
4451
4452 6. Give an error if bad option bits passed at compile or run time.
4453
4454 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4455 pcretest to cause it to pass that flag.
4456
4457 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4458 options, and the first character, if set.
4459
4460 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4461
4462
4463 Version 0.91 10-Sep-97
4464 ----------------------
4465
4466 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4467 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4468
4469 2. PCRE was looping on encountering an indefinitely repeated back reference to
4470 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4471 Perl does - treats the match as successful.
4472
4473 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5