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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 701 - (show annotations) (download)
Tue Sep 20 11:30:56 2011 UTC (3 years, 1 month ago) by ph10
File size: 214515 byte(s)
Fix miscompile of /(*ACCEPT)a/, which thought a match had to start with "a".

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12