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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 681 - (show annotations)
Tue Sep 6 09:16:32 2011 UTC (8 years, 2 months ago) by ph10
File size: 213463 byte(s)
Error occurred while calculating annotation data.
Fix loop/bad error when recursed group contains (*PRUNE) etc.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.20
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.
11
12 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no captured
13 substring, but while checking the failing first alternative, substring 1 is
14 temporarily captured. If the output vector supplied to pcre_exec() was not
15 big enough for this capture, the yield of the function was still zero
16 ("insufficient space for captured substrings"). This cannot be totally fixed
17 without adding another stack variable, which seems a lot of expense for a
18 edge case. However, I have improved the situation in cases such as
19 /(a)(b)x|abc/ matched against "abc", where the return code indicates that
20 fewer than the maximum number of slots in the ovector have been set.
21
22 3. Related to (2) above: when there are more back references in a pattern than
23 slots in the output vector, pcre_exec() uses temporary memory during
24 matching, and copies in the captures as far as possible afterwards. It was
25 using the entire output vector, but this conflicts with the specification
26 that only 2/3 is used for passing back captured substrings. Now it uses only
27 the first 2/3, for compatibility. This is, of course, another edge case.
28
29 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
30 main code base, and can be used by building with --enable-jit.
31
32 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
33 ovector, the return from the function was zero, implying that there were
34 other matches that did not fit. The correct "exactly full" value is now
35 returned.
36
37 6. If a subpattern that was called recursively or as a subroutine contained
38 (*PRUNE) or any other control that caused it to give a non-standard return,
39 invalid errors such as "Error -26 (nested recursion at the same subject
40 position)" or even infinite loops could occur.
41
42
43 Version 8.13 16-Aug-2011
44 ------------------------
45
46 1. The Unicode data tables have been updated to Unicode 6.0.0.
47
48 2. Two minor typos in pcre_internal.h have been fixed.
49
50 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
51 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
52 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
53
54 4. There were a number of related bugs in the code for matching backrefences
55 caselessly in UTF-8 mode when codes for the characters concerned were
56 different numbers of bytes. For example, U+023A and U+2C65 are an upper
57 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
58 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
59 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
60 2-byte code at the end of the subject (it thought there wasn't enough data
61 left).
62
63 5. Comprehensive information about what went wrong is now returned by
64 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
65 as the output vector has at least 2 elements. The offset of the start of
66 the failing character and a reason code are placed in the vector.
67
68 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
69 now returned is for the first byte of the failing character, instead of the
70 last byte inspected. This is an incompatible change, but I hope it is small
71 enough not to be a problem. It makes the returned offset consistent with
72 pcre_exec() and pcre_dfa_exec().
73
74 7. pcretest now gives a text phrase as well as the error number when
75 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
76 failure, the offset and reason code are output.
77
78 8. When \R was used with a maximizing quantifier it failed to skip backwards
79 over a \r\n pair if the subsequent match failed. Instead, it just skipped
80 back over a single character (\n). This seems wrong (because it treated the
81 two characters as a single entity when going forwards), conflicts with the
82 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
83 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
84 has been changed.
85
86 9. Some internal refactoring has changed the processing so that the handling
87 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
88 time (the PCRE_DOTALL option was changed this way some time ago: version
89 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
90 which was always a bit of a fudge. It also means that there is one less
91 argument for the match() function, which reduces its stack requirements
92 slightly. This change also fixes an incompatibility with Perl: the pattern
93 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
94
95 10. More internal refactoring has drastically reduced the number of recursive
96 calls to match() for possessively repeated groups such as (abc)++ when
97 using pcre_exec().
98
99 11. While implementing 10, a number of bugs in the handling of groups were
100 discovered and fixed:
101
102 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
103 (a|)*(?1) gave a compile-time internal error.
104 ((a|)+)+ did not notice that the outer group could match an empty string.
105 (^a|^)+ was not marked as anchored.
106 (.*a|.*)+ was not marked as matching at start or after a newline.
107
108 12. Yet more internal refactoring has removed another argument from the match()
109 function. Special calls to this function are now indicated by setting a
110 value in a variable in the "match data" data block.
111
112 13. Be more explicit in pcre_study() instead of relying on "default" for
113 opcodes that mean there is no starting character; this means that when new
114 ones are added and accidentally left out of pcre_study(), testing should
115 pick them up.
116
117 14. The -s option of pcretest has been documented for ages as being an old
118 synonym of -m (show memory usage). I have changed it to mean "force study
119 for every regex", that is, assume /S for every regex. This is similar to -i
120 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
121 using it. It makes it easier to run collections of tests with and without
122 study enabled, and thereby test pcre_study() more easily. All the standard
123 tests are now run with and without -s (but some patterns can be marked as
124 "never study" - see 20 below).
125
126 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
127 restoration of the capturing data to the outer values was not happening
128 correctly.
129
130 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
131 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
132 pattern had matched an empty string, and so incorrectly returned a no
133 match.
134
135 17. There was optimizing code for the last branch of non-capturing parentheses,
136 and also for the obeyed branch of a conditional subexpression, which used
137 tail recursion to cut down on stack usage. Unfortunately, now that there is
138 the possibility of (*THEN) occurring in these branches, tail recursion is
139 no longer possible because the return has to be checked for (*THEN). These
140 two optimizations have therefore been removed.
141
142 18. If a pattern containing \R was studied, it was assumed that \R always
143 matched two bytes, thus causing the minimum subject length to be
144 incorrectly computed because \R can also match just one byte.
145
146 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
147 was incorrectly computed.
148
149 20. If /S is present twice on a test pattern in pcretest input, it now
150 *disables* studying, thereby overriding the use of -s on the command line
151 (see 14 above). This is necessary for one or two tests to keep the output
152 identical in both cases.
153
154 21. When (*ACCEPT) was used in an assertion that matched an empty string and
155 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
156
157 22. When an atomic group that contained a capturing parenthesis was
158 successfully matched, but the branch in which it appeared failed, the
159 capturing was not being forgotten if a higher numbered group was later
160 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
161 group 1 to "a", when in fact it should be unset. This applied to multi-
162 branched capturing and non-capturing groups, repeated or not, and also to
163 positive assertions (capturing in negative assertions does not happen
164 in PCRE) and also to nested atomic groups.
165
166 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
167 subject after a captured substring, to make it easier to tell which of a
168 number of identical substrings has been captured.
169
170 24. The way atomic groups are processed by pcre_exec() has been changed so that
171 if they are repeated, backtracking one repetition now resets captured
172 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
173 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
174 "aaa". Previously, it would have been "a". As part of this code
175 refactoring, the way recursive calls are handled has also been changed.
176
177 25. If an assertion condition captured any substrings, they were not passed
178 back unless some other capturing happened later. For example, if
179 (?(?=(a))a) was matched against "a", no capturing was returned.
180
181 26. When studying a pattern that contained subroutine calls or assertions,
182 the code for finding the minimum length of a possible match was handling
183 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
184 group 1 called group 2 while simultaneously a separate group 2 called group
185 1). A stack overflow occurred in this case. I have fixed this by limiting
186 the recursion depth to 10.
187
188 27. Updated RunTest.bat in the distribution to the version supplied by Tom
189 Fortmann. This supports explicit test numbers on the command line, and has
190 argument validation and error reporting.
191
192 28. An instance of \X with an unlimited repeat could fail if at any point the
193 first character it looked at was a mark character.
194
195 29. Some minor code refactoring concerning Unicode properties and scripts
196 should reduce the stack requirement of match() slightly.
197
198 30. Added the '=' option to pcretest to check the setting of unused capturing
199 slots at the end of the pattern, which are documented as being -1, but are
200 not included in the return count.
201
202 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
203 compiled something random. Now it gives a compile-time error (as does
204 Perl).
205
206 32. A *MARK encountered during the processing of a positive assertion is now
207 recorded and passed back (compatible with Perl).
208
209 33. If --only-matching or --colour was set on a pcregrep call whose pattern
210 had alternative anchored branches, the search for a second match in a line
211 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
212 matched the line "0102" twice. The same bug affected patterns that started
213 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
214 twice.
215
216 34. Previously, PCRE did not allow quantification of assertions. However, Perl
217 does, and because of capturing effects, quantifying parenthesized
218 assertions may at times be useful. Quantifiers are now allowed for
219 parenthesized assertions.
220
221 35. A minor code tidy in pcre_compile() when checking options for \R usage.
222
223 36. \g was being checked for fancy things in a character class, when it should
224 just be a literal "g".
225
226 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
227 appearance of a nested POSIX class supersedes an apparent external class.
228 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
229 unescaped square brackets may also appear as part of class names. For
230 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
231 more like Perl. (But see 8.20/1 above.)
232
233 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
234 was because it thought it was \N{name}, which is not supported).
235
236 39. Add minix to OS list not supporting the -S option in pcretest.
237
238 40. PCRE tries to detect cases of infinite recursion at compile time, but it
239 cannot analyze patterns in sufficient detail to catch mutual recursions
240 such as ((?1))((?2)). There is now a runtime test that gives an error if a
241 subgroup is called recursively as a subpattern for a second time at the
242 same position in the subject string. In previous releases this might have
243 been caught by the recursion limit, or it might have run out of stack.
244
245 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
246 happen only once. PCRE was, however incorrectly giving a compile time error
247 "recursive call could loop indefinitely" because it cannot analyze the
248 pattern in sufficient detail. The compile time test no longer happens when
249 PCRE is compiling a conditional subpattern, but actual runaway loops are
250 now caught at runtime (see 40 above).
251
252 42. It seems that Perl allows any characters other than a closing parenthesis
253 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
254 has been changed to be the same.
255
256 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
257 as not to get warnings when autogen.sh is called. Also changed
258 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
259
260 44. To help people who use pcregrep to scan files containing exceedingly long
261 lines, the following changes have been made:
262
263 (a) The default value of the buffer size parameter has been increased from
264 8K to 20K. (The actual buffer used is three times this size.)
265
266 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
267 PCRE is built.
268
269 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
270 to be set at run time.
271
272 (d) Numerical values in pcregrep options can be followed by K or M, for
273 example --buffer-size=50K.
274
275 (e) If a line being scanned overflows pcregrep's buffer, an error is now
276 given and the return code is set to 2.
277
278 45. Add a pointer to the latest mark to the callout data block.
279
280 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
281 partial match of an empty string instead of no match. This was specific to
282 the use of ".".
283
284 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
285 complete match instead of a partial match. This bug was dependent on both
286 the PCRE_UTF8 and PCRE_DOTALL options being set.
287
288 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
289 starting byte set, because \b was not being ignored.
290
291
292 Version 8.12 15-Jan-2011
293 ------------------------
294
295 1. Fixed some typos in the markup of the man pages, and wrote a script that
296 checks for such things as part of the documentation building process.
297
298 2. On a big-endian 64-bit system, pcregrep did not correctly process the
299 --match-limit and --recursion-limit options (added for 8.11). In
300 particular, this made one of the standard tests fail. (The integer value
301 went into the wrong half of a long int.)
302
303 3. If the --colour option was given to pcregrep with -v (invert match), it
304 did strange things, either producing crazy output, or crashing. It should,
305 of course, ignore a request for colour when reporting lines that do not
306 match.
307
308 4. Another pcregrep bug caused similar problems if --colour was specified with
309 -M (multiline) and the pattern match finished with a line ending.
310
311 5. In pcregrep, when a pattern that ended with a literal newline sequence was
312 matched in multiline mode, the following line was shown as part of the
313 match. This seems wrong, so I have changed it.
314
315 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
316 the check for further matches in the same line (so they could be coloured)
317 to overrun the end of the current line. If another match was found, it was
318 incorrectly shown (and then shown again when found in the next line).
319
320 7. If pcregrep was compiled under Windows, there was a reference to the
321 function pcregrep_exit() before it was defined. I am assuming this was
322 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
323 reported by a user. I've moved the definition above the reference.
324
325
326 Version 8.11 10-Dec-2010
327 ------------------------
328
329 1. (*THEN) was not working properly if there were untried alternatives prior
330 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
331 backtracked to try for "b" instead of moving to the next alternative branch
332 at the same level (in this case, to look for "c"). The Perl documentation
333 is clear that when (*THEN) is backtracked onto, it goes to the "next
334 alternative in the innermost enclosing group".
335
336 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
337 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
338 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
339 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
340 (*THEN).
341
342 3. If \s appeared in a character class, it removed the VT character from
343 the class, even if it had been included by some previous item, for example
344 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
345 of \s, but is part of the POSIX "space" class.)
346
347 4. A partial match never returns an empty string (because you can always
348 match an empty string at the end of the subject); however the checking for
349 an empty string was starting at the "start of match" point. This has been
350 changed to the "earliest inspected character" point, because the returned
351 data for a partial match starts at this character. This means that, for
352 example, /(?<=abc)def/ gives a partial match for the subject "abc"
353 (previously it gave "no match").
354
355 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
356 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
357 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
358 has an implication that the given string is incomplete (because a partial
359 match is preferred over a full match). For this reason, these items now
360 give a partial match in this situation. [Aside: previously, the one case
361 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
362 match rather than a full match, which was wrong by the old rules, but is
363 now correct.]
364
365 6. There was a bug in the handling of #-introduced comments, recognized when
366 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
367 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
368 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
369 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
370 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
371 places in pcre_compile().
372
373 7. Related to (6) above, when pcre_compile() was skipping #-introduced
374 comments when looking ahead for named forward references to subpatterns,
375 the only newline sequence it recognized was NL. It now handles newlines
376 according to the set newline convention.
377
378 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
379 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
380 cater for a lack of strerror(). These oversights have been fixed.
381
382 9. Added --match-limit and --recursion-limit to pcregrep.
383
384 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
385
386 11. When the -o option was used, pcregrep was setting a return code of 1, even
387 when matches were found, and --line-buffered was not being honoured.
388
389 12. Added an optional parentheses number to the -o and --only-matching options
390 of pcregrep.
391
392 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
393 can match an empty string. The code to do it in pcretest and pcredemo
394 needed fixing:
395
396 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
397 only one byte after an empty string match just before CRLF (this case
398 just got forgotten; "any" and "anycrlf" were OK).
399
400 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
401 mode when an empty string match preceded an ASCII character followed by
402 a non-ASCII character. (The code for advancing by one character rather
403 than one byte was nonsense.)
404
405 (c) The pcredemo.c sample program did not have any code at all to handle
406 the cases when CRLF is a valid newline sequence.
407
408 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
409 as a starting offset was within the subject string. There is now a new
410 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
411 negative or greater than the length of the string. In order to test this,
412 pcretest is extended to allow the setting of negative starting offsets.
413
414 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
415 starting offset points to the beginning of a UTF-8 character was
416 unnecessarily clumsy. I tidied it up.
417
418 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
419 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
420
421 17. Nobody had reported that the --include_dir option, which was added in
422 release 7.7 should have been called --include-dir (hyphen, not underscore)
423 for compatibility with GNU grep. I have changed it to --include-dir, but
424 left --include_dir as an undocumented synonym, and the same for
425 --exclude-dir, though that is not available in GNU grep, at least as of
426 release 2.5.4.
427
428 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
429 characters from a string of bytes) have been redefined so as not to use
430 loops, in order to improve performance in some environments. At the same
431 time, I abstracted some of the common code into auxiliary macros to save
432 repetition (this should not affect the compiled code).
433
434 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
435 compile-time error is now given if \c is not followed by an ASCII
436 character, that is, a byte less than 128. (In EBCDIC mode, the code is
437 different, and any byte value is allowed.)
438
439 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
440 START_OPTIMIZE option, which is now allowed at compile time - but just
441 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
442 to pcregrep and other applications that have no direct access to PCRE
443 options. The new /Y option in pcretest sets this option when calling
444 pcre_compile().
445
446 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
447 back references. Groups containing recursive back references were forced to
448 be atomic by that change, but in the case of named groups, the amount of
449 memory required was incorrectly computed, leading to "Failed: internal
450 error: code overflow". This has been fixed.
451
452 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
453 pcretest.c, to avoid build problems in some Borland environments.
454
455
456 Version 8.10 25-Jun-2010
457 ------------------------
458
459 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
460 THEN.
461
462 2. (*ACCEPT) was not working when inside an atomic group.
463
464 3. Inside a character class, \B is treated as a literal by default, but
465 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
466 causes the error). The code is unchanged, but I tidied the documentation.
467
468 4. Inside a character class, PCRE always treated \R and \X as literals,
469 whereas Perl faults them if its -w option is set. I have changed PCRE so
470 that it faults them when PCRE_EXTRA is set.
471
472 5. Added support for \N, which always matches any character other than
473 newline. (It is the same as "." when PCRE_DOTALL is not set.)
474
475 6. When compiling pcregrep with newer versions of gcc which may have
476 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
477 declared with attribute warn_unused_result" were given. Just casting the
478 result to (void) does not stop the warnings; a more elaborate fudge is
479 needed. I've used a macro to implement this.
480
481 7. Minor change to pcretest.c to avoid a compiler warning.
482
483 8. Added four artifical Unicode properties to help with an option to make
484 \s etc use properties (see next item). The new properties are: Xan
485 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
486
487 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
488 use Unicode properties. (*UCP) at the start of a pattern can be used to set
489 this option. Modified pcretest to add /W to test this facility. Added
490 REG_UCP to make it available via the POSIX interface.
491
492 10. Added --line-buffered to pcregrep.
493
494 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
495 studied, and the match started with a letter with a code point greater than
496 127 whose first byte was different to the first byte of the other case of
497 the letter, the other case of this starting letter was not recognized
498 (#976).
499
500 12. If a pattern that was studied started with a repeated Unicode property
501 test, for example, \p{Nd}+, there was the theoretical possibility of
502 setting up an incorrect bitmap of starting bytes, but fortunately it could
503 not have actually happened in practice until change 8 above was made (it
504 added property types that matched character-matching opcodes).
505
506 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
507 possible starting bytes for non-anchored patterns.
508
509 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
510 \R, and also a number of cases that involve Unicode properties, both
511 explicit and implicit when PCRE_UCP is set.
512
513 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
514 input, it could crash or give wrong results if characters with values
515 greater than 0xc0 were present in the subject string. (Detail: it assumed
516 UTF-8 input when processing these items.)
517
518 16. Added a lot of (int) casts to avoid compiler warnings in systems where
519 size_t is 64-bit (#991).
520
521 17. Added a check for running out of memory when PCRE is compiled with
522 --disable-stack-for-recursion (#990).
523
524 18. If the last data line in a file for pcretest does not have a newline on
525 the end, a newline was missing in the output.
526
527 19. The default pcre_chartables.c file recognizes only ASCII characters (values
528 less than 128) in its various bitmaps. However, there is a facility for
529 generating tables according to the current locale when PCRE is compiled. It
530 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
531 characters, are recognized by isspace() and therefore were getting set in
532 these tables, and indeed these tables seem to approximate to ISO 8859. This
533 caused a problem in UTF-8 mode when pcre_study() was used to create a list
534 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
535 which of course cannot start UTF-8 characters. I have changed the code so
536 that only real ASCII characters (less than 128) and the correct starting
537 bytes for UTF-8 encodings are set for characters greater than 127 when in
538 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
539 altogether.)
540
541 20. Added the /T option to pcretest so as to be able to run tests with non-
542 standard character tables, thus making it possible to include the tests
543 used for 19 above in the standard set of tests.
544
545 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
546 reference to a subpattern the other side of a comment that contains an
547 opening parenthesis caused either an internal compiling error, or a
548 reference to the wrong subpattern.
549
550
551 Version 8.02 19-Mar-2010
552 ------------------------
553
554 1. The Unicode data tables have been updated to Unicode 5.2.0.
555
556 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
557 configured.
558
559 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
560 original author of that file, following a query about its status.
561
562 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
563 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
564
565 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
566 quantifier applied to a forward-referencing subroutine call, could compile
567 incorrect code or give the error "internal error: previously-checked
568 referenced subpattern not found".
569
570 6. Both MS Visual Studio and Symbian OS have problems with initializing
571 variables to point to external functions. For these systems, therefore,
572 pcre_malloc etc. are now initialized to local functions that call the
573 relevant global functions.
574
575 7. There were two entries missing in the vectors called coptable and poptable
576 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
577 I've fixed the data, and added a kludgy way of testing at compile time that
578 the lengths are correct (equal to the number of opcodes).
579
580 8. Following on from 7, I added a similar kludge to check the length of the
581 eint vector in pcreposix.c.
582
583 9. Error texts for pcre_compile() are held as one long string to avoid too
584 much relocation at load time. To find a text, the string is searched,
585 counting zeros. There was no check for running off the end of the string,
586 which could happen if a new error number was added without updating the
587 string.
588
589 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
590
591 11. \K was not working if it appeared in an atomic group or in a group that
592 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
593 \K is "not well defined" if used in an assertion. PCRE now accepts it if
594 the assertion is positive, but not if it is negative.
595
596 12. Change 11 fortuitously reduced the size of the stack frame used in the
597 "match()" function of pcre_exec.c by one pointer. Forthcoming
598 implementation of support for (*MARK) will need an extra pointer on the
599 stack; I have reserved it now, so that the stack frame size does not
600 decrease.
601
602 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
603 item in branch that calls a recursion is a subroutine call - as in the
604 second branch in the above example - was incorrectly given the compile-
605 time error "recursive call could loop indefinitely" because pcre_compile()
606 was not correctly checking the subroutine for matching a non-empty string.
607
608 14. The checks for overrunning compiling workspace could trigger after an
609 overrun had occurred. This is a "should never occur" error, but it can be
610 triggered by pathological patterns such as hundreds of nested parentheses.
611 The checks now trigger 100 bytes before the end of the workspace.
612
613 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
614
615
616 Version 8.01 19-Jan-2010
617 ------------------------
618
619 1. If a pattern contained a conditional subpattern with only one branch (in
620 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
621 computed the wrong minimum data length (which is of course zero for such
622 subpatterns). This could cause incorrect "no match" results.
623
624 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
625 the pattern is reset in the first branch, pcre_compile() failed with
626 "internal error: code overflow at offset...". This happened only when
627 the reset was to the original external option setting. (An optimization
628 abstracts leading options settings into an external setting, which was the
629 cause of this.)
630
631 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
632 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
633 assertion pattern did not match (meaning that the assertion was true), it
634 was incorrectly treated as false if the SKIP had been reached during the
635 matching. This also applied to assertions used as conditions.
636
637 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
638 assertion subpattern, including such a pattern used as a condition,
639 unpredictable results occurred, instead of the error return
640 PCRE_ERROR_DFA_UITEM.
641
642 5. The C++ GlobalReplace function was not working like Perl for the special
643 situation when an empty string is matched. It now does the fancy magic
644 stuff that is necessary.
645
646 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
647 removed. (These were left over from very, very early versions of PCRE.)
648
649 7. Some cosmetic changes to the code to make life easier when compiling it
650 as part of something else:
651
652 (a) Change DEBUG to PCRE_DEBUG.
653
654 (b) In pcre_compile(), rename the member of the "branch_chain" structure
655 called "current" as "current_branch", to prevent a collision with the
656 Linux macro when compiled as a kernel module.
657
658 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
659 prevent a collision with the Linux macro when compiled as a kernel
660 module.
661
662 8. In pcre_compile() there are some checks for integer overflows that used to
663 cast potentially large values to (double). This has been changed to that
664 when building, a check for int64_t is made, and if it is found, it is used
665 instead, thus avoiding the use of floating point arithmetic. (There is no
666 other use of FP in PCRE.) If int64_t is not found, the fallback is to
667 double.
668
669 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
670 2005 (difference between two addresses compared to an unsigned value).
671
672 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
673 custom one, because of the following reported problem in Windows:
674
675 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
676 under Win32.
677 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
678 therefore missing the function definition.
679 - The compiler thus generates a "C" signature for the test function.
680 - The linker fails to find the "C" function.
681 - PCRE fails to configure if asked to do so against libbz2.
682
683 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
684 messages were output:
685
686 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
687 rerunning libtoolize, to keep the correct libtool macros in-tree.
688 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
689
690 I have done both of these things.
691
692 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
693 most of the time, it *can* run out if it is given a pattern that contains a
694 runaway infinite recursion. I updated the discussion in the pcrestack man
695 page.
696
697 13. Now that we have gone to the x.xx style of version numbers, the minor
698 version may start with zero. Using 08 or 09 is a bad idea because users
699 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
700 interpreted as invalid octal numbers. I've updated the previous comment in
701 configure.ac, and also added a check that gives an error if 08 or 09 are
702 used.
703
704 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
705 causing partial matching to fail when the end of the subject matched \W
706 in a UTF-8 pattern where \W was quantified with a minimum of 3.
707
708 15. There were some discrepancies between the declarations in pcre_internal.h
709 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
710 their definitions. The declarations used "const uschar *" and the
711 definitions used USPTR. Even though USPTR is normally defined as "const
712 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
713 reported that: "This difference in casting confuses some C++ compilers, for
714 example, SunCC recognizes above declarations as different functions and
715 generates broken code for hbpcre." I have changed the declarations to use
716 USPTR.
717
718 16. GNU libtool is named differently on some systems. The autogen.sh script now
719 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
720 (FreeBSD).
721
722 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
723 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
724 comment: "Figure out how to create a longlong from a string: strtoll and
725 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
726 instance, but it only takes 2 args instead of 3!"
727
728 18. A subtle bug concerned with back references has been fixed by a change of
729 specification, with a corresponding code fix. A pattern such as
730 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
731 refers, was giving matches when it shouldn't. For example, xa=xaaa would
732 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
733 same bug. Such groups have to be quantified to be useful, or contained
734 inside another quantified group. (If there's no repetition, the reference
735 can never match.) The problem arises because, having left the group and
736 moved on to the rest of the pattern, a later failure that backtracks into
737 the group uses the captured value from the final iteration of the group
738 rather than the correct earlier one. I have fixed this in PCRE by forcing
739 any group that contains a reference to itself to be an atomic group; that
740 is, there cannot be any backtracking into it once it has completed. This is
741 similar to recursive and subroutine calls.
742
743
744 Version 8.00 19-Oct-09
745 ----------------------
746
747 1. The table for translating pcre_compile() error codes into POSIX error codes
748 was out-of-date, and there was no check on the pcre_compile() error code
749 being within the table. This could lead to an OK return being given in
750 error.
751
752 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
753 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
754 in a Windows environment.
755
756 3. The pcregrep --count option prints the count for each file even when it is
757 zero, as does GNU grep. However, pcregrep was also printing all files when
758 --files-with-matches was added. Now, when both options are given, it prints
759 counts only for those files that have at least one match. (GNU grep just
760 prints the file name in this circumstance, but including the count seems
761 more useful - otherwise, why use --count?) Also ensured that the
762 combination -clh just lists non-zero counts, with no names.
763
764 4. The long form of the pcregrep -F option was incorrectly implemented as
765 --fixed_strings instead of --fixed-strings. This is an incompatible change,
766 but it seems right to fix it, and I didn't think it was worth preserving
767 the old behaviour.
768
769 5. The command line items --regex=pattern and --regexp=pattern were not
770 recognized by pcregrep, which required --regex pattern or --regexp pattern
771 (with a space rather than an '='). The man page documented the '=' forms,
772 which are compatible with GNU grep; these now work.
773
774 6. No libpcreposix.pc file was created for pkg-config; there was just
775 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
776
777 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
778 when UCP support is not needed, by modifying the Python script that
779 generates it from Unicode data files. This should not matter if the module
780 is correctly used as a library, but I received one complaint about 50K of
781 unwanted data. My guess is that the person linked everything into his
782 program rather than using a library. Anyway, it does no harm.
783
784 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
785 was a minimum greater than 1 for a wide character in a possessive
786 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
787 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
788 character. Chaos in the form of incorrect output or a compiling loop could
789 result.
790
791 9. The restrictions on what a pattern can contain when partial matching is
792 requested for pcre_exec() have been removed. All patterns can now be
793 partially matched by this function. In addition, if there are at least two
794 slots in the offset vector, the offset of the earliest inspected character
795 for the match and the offset of the end of the subject are set in them when
796 PCRE_ERROR_PARTIAL is returned.
797
798 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
799 synonymous with PCRE_PARTIAL, for backwards compatibility, and
800 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
801 and may be more useful for multi-segment matching.
802
803 11. Partial matching with pcre_exec() is now more intuitive. A partial match
804 used to be given if ever the end of the subject was reached; now it is
805 given only if matching could not proceed because another character was
806 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
807 string "Z", which now yields "no match" instead of "partial match". In the
808 case of pcre_dfa_exec(), "no match" is given if every matching path for the
809 final character ended with (*FAIL).
810
811 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
812 if the pattern had a "must contain" character that was already found in the
813 earlier partial match, unless partial matching was again requested. For
814 example, with the pattern /dog.(body)?/, the "must contain" character is
815 "g". If the first part-match was for the string "dog", restarting with
816 "sbody" failed. This bug has been fixed.
817
818 13. The string returned by pcre_dfa_exec() after a partial match has been
819 changed so that it starts at the first inspected character rather than the
820 first character of the match. This makes a difference only if the pattern
821 starts with a lookbehind assertion or \b or \B (\K is not supported by
822 pcre_dfa_exec()). It's an incompatible change, but it makes the two
823 matching functions compatible, and I think it's the right thing to do.
824
825 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
826 so that the demonstration program is easily available in environments where
827 PCRE has not been installed from source.
828
829 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
830 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
831 library.
832
833 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
834 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
835 is not the first non-POSIX option to be added. Clearly some people find
836 these options useful.
837
838 17. If a caller to the POSIX matching function regexec() passes a non-zero
839 value for nmatch with a NULL value for pmatch, the value of
840 nmatch is forced to zero.
841
842 18. RunGrepTest did not have a test for the availability of the -u option of
843 the diff command, as RunTest does. It now checks in the same way as
844 RunTest, and also checks for the -b option.
845
846 19. If an odd number of negated classes containing just a single character
847 interposed, within parentheses, between a forward reference to a named
848 subpattern and the definition of the subpattern, compilation crashed with
849 an internal error, complaining that it could not find the referenced
850 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
851 [The bug was that it was starting one character too far in when skipping
852 over the character class, thus treating the ] as data rather than
853 terminating the class. This meant it could skip too much.]
854
855 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
856 /g option in pcretest when the pattern contains \K, which makes it possible
857 to have an empty string match not at the start, even when the pattern is
858 anchored. Updated pcretest and pcredemo to use this option.
859
860 21. If the maximum number of capturing subpatterns in a recursion was greater
861 than the maximum at the outer level, the higher number was returned, but
862 with unset values at the outer level. The correct (outer level) value is
863 now given.
864
865 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
866 PCRE did not set those parentheses (unlike Perl). I have now found a way to
867 make it do so. The string so far is captured, making this feature
868 compatible with Perl.
869
870 23. The tests have been re-organized, adding tests 11 and 12, to make it
871 possible to check the Perl 5.10 features against Perl 5.10.
872
873 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
874 pattern matches a fixed length string. PCRE did not allow this; now it
875 does. Neither allows recursion.
876
877 25. I finally figured out how to implement a request to provide the minimum
878 length of subject string that was needed in order to match a given pattern.
879 (It was back references and recursion that I had previously got hung up
880 on.) This code has now been added to pcre_study(); it finds a lower bound
881 to the length of subject needed. It is not necessarily the greatest lower
882 bound, but using it to avoid searching strings that are too short does give
883 some useful speed-ups. The value is available to calling programs via
884 pcre_fullinfo().
885
886 26. While implementing 25, I discovered to my embarrassment that pcretest had
887 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
888 study optimizations had never been tested with that matching function.
889 Oops. What is worse, even when it was passed study data, there was a bug in
890 pcre_dfa_exec() that meant it never actually used it. Double oops. There
891 were also very few tests of studied patterns with pcre_dfa_exec().
892
893 27. If (?| is used to create subpatterns with duplicate numbers, they are now
894 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
895 on the other side of the coin, they are no longer allowed to have different
896 names, because these cannot be distinguished in PCRE, and this has caused
897 confusion. (This is a difference from Perl.)
898
899 28. When duplicate subpattern names are present (necessarily with different
900 numbers, as required by 27 above), and a test is made by name in a
901 conditional pattern, either for a subpattern having been matched, or for
902 recursion in such a pattern, all the associated numbered subpatterns are
903 tested, and the overall condition is true if the condition is true for any
904 one of them. This is the way Perl works, and is also more like the way
905 testing by number works.
906
907
908 Version 7.9 11-Apr-09
909 ---------------------
910
911 1. When building with support for bzlib/zlib (pcregrep) and/or readline
912 (pcretest), all targets were linked against these libraries. This included
913 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
914 libraries. This caused unwanted dependencies to be created. This problem
915 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
916 pcretest is linked with readline.
917
918 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
919 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
920 moved outside it again, because FALSE and TRUE are already defined in AIX,
921 but BOOL is not.
922
923 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
924 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
925
926 4. The pcregrep documentation said spaces were inserted as well as colons (or
927 hyphens) following file names and line numbers when outputting matching
928 lines. This is not true; no spaces are inserted. I have also clarified the
929 wording for the --colour (or --color) option.
930
931 5. In pcregrep, when --colour was used with -o, the list of matching strings
932 was not coloured; this is different to GNU grep, so I have changed it to be
933 the same.
934
935 6. When --colo(u)r was used in pcregrep, only the first matching substring in
936 each matching line was coloured. Now it goes on to look for further matches
937 of any of the test patterns, which is the same behaviour as GNU grep.
938
939 7. A pattern that could match an empty string could cause pcregrep to loop; it
940 doesn't make sense to accept an empty string match in pcregrep, so I have
941 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
942 seems to be how GNU grep behaves.
943
944 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
945 start or after a newline", because the conditional assertion was not being
946 correctly handled. The rule now is that both the assertion and what follows
947 in the first alternative must satisfy the test.
948
949 9. If auto-callout was enabled in a pattern with a conditional group whose
950 condition was an assertion, PCRE could crash during matching, both with
951 pcre_exec() and pcre_dfa_exec().
952
953 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
954 used for matching.
955
956 11. Unicode property support in character classes was not working for
957 characters (bytes) greater than 127 when not in UTF-8 mode.
958
959 12. Added the -M command line option to pcretest.
960
961 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
962
963 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
964
965 16. Added comments and documentation about mis-use of no_arg in the C++
966 wrapper.
967
968 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
969 from Martin Jerabek that uses macro names for all relevant character and
970 string constants.
971
972 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
973 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
974 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
975 these, but not everybody uses configure.
976
977 19. A conditional group that had only one branch was not being correctly
978 recognized as an item that could match an empty string. This meant that an
979 enclosing group might also not be so recognized, causing infinite looping
980 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
981 with the subject "ab", where knowledge that the repeated group can match
982 nothing is needed in order to break the loop.
983
984 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
985 exec(), but without supplying a callout function, matching went wrong.
986
987 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
988 leak if the size of the offset vector was greater than 30. When the vector
989 is smaller, the saved offsets during recursion go onto a local stack
990 vector, but for larger vectors malloc() is used. It was failing to free
991 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
992 error, in fact).
993
994 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
995 heapframe that is used only when UTF-8 support is enabled. This caused no
996 problem, but was untidy.
997
998 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
999 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1000 included within another project.
1001
1002 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1003 slightly modified by me:
1004
1005 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1006 not building pcregrep.
1007
1008 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1009 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1010
1011 25. Forward references, both numeric and by name, in patterns that made use of
1012 duplicate group numbers, could behave incorrectly or give incorrect errors,
1013 because when scanning forward to find the reference group, PCRE was not
1014 taking into account the duplicate group numbers. A pattern such as
1015 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1016
1017 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1018 the feature of a custom pointer more persuasive (as requested by a user).
1019
1020 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1021 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1022 pre-defined.
1023
1024 28. Added support for (*UTF8) at the start of a pattern.
1025
1026 29. Arrange for flags added by the "release type" setting in CMake to be shown
1027 in the configuration summary.
1028
1029
1030 Version 7.8 05-Sep-08
1031 ---------------------
1032
1033 1. Replaced UCP searching code with optimized version as implemented for Ad
1034 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1035 stage table and inline lookup instead of a function, giving speed ups of 2
1036 to 5 times on some simple patterns that I tested. Permission was given to
1037 distribute the MultiStage2.py script that generates the tables (it's not in
1038 the tarball, but is in the Subversion repository).
1039
1040 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1041 scripts.
1042
1043 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1044 a group with a zero qualifier. The result of the study could be incorrect,
1045 or the function might crash, depending on the pattern.
1046
1047 4. Caseless matching was not working for non-ASCII characters in back
1048 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1049 It now works when Unicode Property Support is available.
1050
1051 5. In pcretest, an escape such as \x{de} in the data was always generating
1052 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1053 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1054 truncation.
1055
1056 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1057
1058 7. Added two (int) casts to pcregrep when printing the difference of two
1059 pointers, in case they are 64-bit values.
1060
1061 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1062 test 2 if it fails.
1063
1064 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1065 and a #define of that name to empty if it is not externally set. This is to
1066 allow users of MSVC to set it if necessary.
1067
1068 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1069 the convenience functions in the pcre_get.c source file.
1070
1071 11. An option change at the start of a pattern that had top-level alternatives
1072 could cause overwriting and/or a crash. This command provoked a crash in
1073 some environments:
1074
1075 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1076
1077 This potential security problem was recorded as CVE-2008-2371.
1078
1079 12. For a pattern where the match had to start at the beginning or immediately
1080 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1081 pcre_dfa_exec() could read past the end of the passed subject if there was
1082 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1083 pcretest so that it places the subject at the end of its malloc-ed buffer.
1084
1085 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1086 exec() might read past the end of the data buffer in UTF-8 mode.
1087
1088 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1089 the data contained the byte 0x85 as part of a UTF-8 character within its
1090 first line. This applied both to normal and DFA matching.
1091
1092 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1093 /^[^d]*?$/8 failed to match "abc".
1094
1095 16. Added a missing copyright notice to pcrecpp_internal.h.
1096
1097 17. Make it more clear in the documentation that values returned from
1098 pcre_exec() in ovector are byte offsets, not character counts.
1099
1100 18. Tidied a few places to stop certain compilers from issuing warnings.
1101
1102 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1103 supplied by Stefan Weber. I made a further small update for 7.8 because
1104 there is a change of source arrangements: the pcre_searchfuncs.c module is
1105 replaced by pcre_ucd.c.
1106
1107
1108 Version 7.7 07-May-08
1109 ---------------------
1110
1111 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1112 a string to a long long, pretend we don't even have a long long." This is
1113 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1114
1115 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1116 pre-7.6 versions, which defined a global no_arg variable instead of putting
1117 it in the RE class. (See also #8 below.)
1118
1119 3. Remove a line of dead code, identified by coverity and reported by Nuno
1120 Lopes.
1121
1122 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1123
1124 (1) The include/exclude patterns were being applied to the whole pathnames
1125 of files, instead of just to the final components.
1126
1127 (2) If there was more than one level of directory, the subdirectories were
1128 skipped unless they satisfied the include/exclude conditions. This is
1129 inconsistent with GNU grep (and could even be seen as contrary to the
1130 pcregrep specification - which I improved to make it absolutely clear).
1131 The action now is always to scan all levels of directory, and just
1132 apply the include/exclude patterns to regular files.
1133
1134 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1135 --exclude_dir in the tests to avoid scanning .svn directories.
1136
1137 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1138 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1139 doesn't support NULs in patterns.
1140
1141 7. Added some missing "const"s to declarations of static tables in
1142 pcre_compile.c and pcre_dfa_exec.c.
1143
1144 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1145 caused by fix #2 above. (Subsequently also a second patch to fix the
1146 first patch. And a third patch - this was a messy problem.)
1147
1148 9. Applied Craig's patch to remove the use of push_back().
1149
1150 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1151 matching function regexec().
1152
1153 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1154 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1155 references. PCRE supports relative numbers with this syntax (I don't think
1156 Oniguruma does).
1157
1158 12. Previously, a group with a zero repeat such as (...){0} was completely
1159 omitted from the compiled regex. However, this means that if the group
1160 was called as a subroutine from elsewhere in the pattern, things went wrong
1161 (an internal error was given). Such groups are now left in the compiled
1162 pattern, with a new opcode that causes them to be skipped at execution
1163 time.
1164
1165 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1166 to the way PCRE behaves:
1167
1168 (a) A lone ] character is dis-allowed (Perl treats it as data).
1169
1170 (b) A back reference to an unmatched subpattern matches an empty string
1171 (Perl fails the current match path).
1172
1173 (c) A data ] in a character class must be notated as \] because if the
1174 first data character in a class is ], it defines an empty class. (In
1175 Perl it is not possible to have an empty class.) The empty class []
1176 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1177 The negative empty class [^] matches any one character, independently
1178 of the DOTALL setting.
1179
1180 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1181 non-existent subpattern following a character class starting with ']' and
1182 containing () gave an internal compiling error instead of "reference to
1183 non-existent subpattern". Fortunately, when the pattern did exist, the
1184 compiled code was correct. (When scanning forwards to check for the
1185 existencd of the subpattern, it was treating the data ']' as terminating
1186 the class, so got the count wrong. When actually compiling, the reference
1187 was subsequently set up correctly.)
1188
1189 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1190 it was being rejected as not supported by pcre_dfa_exec(), even though
1191 other assertions are supported. I have made pcre_dfa_exec() support
1192 (*FAIL).
1193
1194 16. The implementation of 13c above involved the invention of a new opcode,
1195 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1196 cannot be changed at match time, I realized I could make a small
1197 improvement to matching performance by compiling OP_ALLANY instead of
1198 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1199 on the OP_ANY path.
1200
1201 17. Compiling pcretest on Windows with readline support failed without the
1202 following two fixes: (1) Make the unistd.h include conditional on
1203 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1204
1205 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1206 ncurses library to be included for pcretest when ReadLine support is
1207 requested, but also to allow for it to be overridden. This patch came from
1208 Daniel Bergström.
1209
1210 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1211 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1212 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1213 spotting this.
1214
1215
1216 Version 7.6 28-Jan-08
1217 ---------------------
1218
1219 1. A character class containing a very large number of characters with
1220 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1221 overflow.
1222
1223 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1224 HAVE_LONG_LONG is not defined.
1225
1226 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1227 bring them up to date and include new features. This patch includes:
1228
1229 - Fixed PH's badly added libz and libbz2 support.
1230 - Fixed a problem with static linking.
1231 - Added pcredemo. [But later removed - see 7 below.]
1232 - Fixed dftables problem and added an option.
1233 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1234 HAVE_LONG_LONG.
1235 - Added readline support for pcretest.
1236 - Added an listing of the option settings after cmake has run.
1237
1238 4. A user submitted a patch to Makefile that makes it easy to create
1239 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1240 Makefile.am that cause it to include this special target, without
1241 affecting anything else. Note that the same mingw target plus all
1242 the other distribution libraries and programs are now supported
1243 when configuring with CMake (see 6 below) instead of with
1244 Configure/Make.
1245
1246 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1247 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1248 exported in the Windows port". It has not yet been confirmed that the patch
1249 solves the problem, but it does no harm.
1250
1251 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1252 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1253 with CMake, and also correct the comment about stack recursion.
1254
1255 7. Remove the automatic building of pcredemo from the ./configure system and
1256 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1257 of a program that users should build themselves after PCRE is installed, so
1258 building it automatically is not really right. What is more, it gave
1259 trouble in some build environments.
1260
1261 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1262
1263
1264 Version 7.5 10-Jan-08
1265 ---------------------
1266
1267 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1268 values in parens when parsing an RE using the C++ wrapper."
1269
1270 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1271 Characters greater than 255 were excluded from the class instead of being
1272 included.
1273
1274 3. The same bug as (2) above applied to negated POSIX classes such as
1275 [:^space:].
1276
1277 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1278 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1279 I have changed it.
1280
1281 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1282 first named subpattern) and a construct such as (?&a) would reference the
1283 first named subpattern whose name started with "a" (in other words, the
1284 length check was missing). Both these problems are fixed. "Subpattern name
1285 expected" is now given for (?&) (a zero-length name), and this patch also
1286 makes it give the same error for \k'' (previously it complained that that
1287 was a reference to a non-existent subpattern).
1288
1289 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1290 this is right because (?- can be followed by option settings as well as by
1291 digits. I have, however, made the messages clearer.
1292
1293 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1294 than the number used in the conditional) now cause a compile-time error.
1295 This is actually not compatible with Perl, which accepts such patterns, but
1296 treats the conditional as always being FALSE (as PCRE used to), but it
1297 seems to me that giving a diagnostic is better.
1298
1299 8. Change "alphameric" to the more common word "alphanumeric" in comments
1300 and messages.
1301
1302 9. Fix two occurrences of "backslash" in comments that should have been
1303 "backspace".
1304
1305 10. Remove two redundant lines of code that can never be obeyed (their function
1306 was moved elsewhere).
1307
1308 11. The program that makes PCRE's Unicode character property table had a bug
1309 which caused it to generate incorrect table entries for sequences of
1310 characters that have the same character type, but are in different scripts.
1311 It amalgamated them into a single range, with the script of the first of
1312 them. In other words, some characters were in the wrong script. There were
1313 thirteen such cases, affecting characters in the following ranges:
1314
1315 U+002b0 - U+002c1
1316 U+0060c - U+0060d
1317 U+0061e - U+00612
1318 U+0064b - U+0065e
1319 U+0074d - U+0076d
1320 U+01800 - U+01805
1321 U+01d00 - U+01d77
1322 U+01d9b - U+01dbf
1323 U+0200b - U+0200f
1324 U+030fc - U+030fe
1325 U+03260 - U+0327f
1326 U+0fb46 - U+0fbb1
1327 U+10450 - U+1049d
1328
1329 12. The -o option (show only the matching part of a line) for pcregrep was not
1330 compatible with GNU grep in that, if there was more than one match in a
1331 line, it showed only the first of them. It now behaves in the same way as
1332 GNU grep.
1333
1334 13. If the -o and -v options were combined for pcregrep, it printed a blank
1335 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1336 does the same. The return code can be used to tell if there were any
1337 non-matching lines.
1338
1339 14. Added --file-offsets and --line-offsets to pcregrep.
1340
1341 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1342 infinitely looping recursion. The bug was that positive lookaheads were not
1343 being skipped when checking for a possible empty match (negative lookaheads
1344 and both kinds of lookbehind were skipped).
1345
1346 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1347 inclusion of <windows.h> to before rather than after the definition of
1348 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1349
1350 17. Specifying a possessive quantifier with a specific limit for a Unicode
1351 character property caused pcre_compile() to compile bad code, which led at
1352 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1353 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1354 caused the error; without that there was no problem.
1355
1356 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1357
1358 19. Added --enable-pcretest-libreadline.
1359
1360 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1361 RE::GlobalReplace(). As a result, the number of replacements returned was
1362 double what it should be. I removed one of the increments, but Craig sent a
1363 later patch that removed the other one (the right fix) and added unit tests
1364 that check the return values (which was not done before).
1365
1366 21. Several CMake things:
1367
1368 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1369 the names libpcre and libpcreposix, not just pcre and pcreposix.
1370
1371 (2) The above change means that pcretest and pcregrep are now correctly
1372 linked with the newly-built libraries, not previously installed ones.
1373
1374 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1375
1376 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1377 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1378 UTF-8 newline character). The key issue is that the pattern starts .*;
1379 this means that the match must be either at the beginning, or after a
1380 newline. The bug was in the code for advancing after a failed match and
1381 checking that the new position followed a newline. It was not taking
1382 account of UTF-8 characters correctly.
1383
1384 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1385 character classes. PCRE was not treating the sequence [:...:] as a
1386 character class unless the ... were all letters. Perl, however, seems to
1387 allow any characters between [: and :], though of course it rejects as
1388 unknown any "names" that contain non-letters, because all the known class
1389 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1390 for example, whereas PCRE did not - it did not recognize a POSIX character
1391 class. This seemed a bit dangerous, so the code has been changed to be
1392 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1393 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1394 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1395 Perl does, and where it didn't before.
1396
1397 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1398 Windows environments %n is disabled by default.
1399
1400
1401 Version 7.4 21-Sep-07
1402 ---------------------
1403
1404 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1405 means that a class such as [\s] counted as "explicit reference to CR or
1406 LF". That isn't really right - the whole point of the change was to try to
1407 help when there was an actual mention of one of the two characters. So now
1408 the change happens only if \r or \n (or a literal CR or LF) character is
1409 encountered.
1410
1411 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1412 of both had grown to the point where there were only 3 bits left.
1413 Fortunately, there was spare space in the data structure, and so I have
1414 moved the internal flags into a new 16-bit field to free up more option
1415 bits.
1416
1417 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1418 but did not set the internal JCHANGED flag - either of these is enough to
1419 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1420 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1421 start sets both bits.
1422
1423 4. Added options (at build time, compile time, exec time) to change \R from
1424 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1425
1426 5. doc/pcresyntax.html was missing from the distribution.
1427
1428 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1429 compatibility, even though it is no longer used.
1430
1431 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1432 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1433 windows.h file is present (where different names are used). [This was
1434 reversed later after testing - see 16 below.]
1435
1436 8. Changed all #include <config.h> to #include "config.h". There were also
1437 some further <pcre.h> cases that I changed to "pcre.h".
1438
1439 9. When pcregrep was used with the --colour option, it missed the line ending
1440 sequence off the lines that it output.
1441
1442 10. It was pointed out to me that arrays of string pointers cause lots of
1443 relocations when a shared library is dynamically loaded. A technique of
1444 using a single long string with a table of offsets can drastically reduce
1445 these. I have refactored PCRE in four places to do this. The result is
1446 dramatic:
1447
1448 Originally: 290
1449 After changing UCP table: 187
1450 After changing error message table: 43
1451 After changing table of "verbs" 36
1452 After changing table of Posix names 22
1453
1454 Thanks to the folks working on Gregex for glib for this insight.
1455
1456 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1457 unicode-properties was also set.
1458
1459 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1460
1461 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1462 checked only for CRLF.
1463
1464 14. Added casts to pcretest.c to avoid compiler warnings.
1465
1466 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1467
1468 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1469 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1470 entirely. This removes changes made in 7 above.
1471
1472 17. The CMake files have been updated, and there is now more information about
1473 building with CMake in the NON-UNIX-USE document.
1474
1475
1476 Version 7.3 28-Aug-07
1477 ---------------------
1478
1479 1. In the rejigging of the build system that eventually resulted in 7.1, the
1480 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1481 brackets there is not right, since it causes compilers to look for an
1482 installed pcre.h, not the version that is in the source that is being
1483 compiled (which of course may be different). I have changed it back to:
1484
1485 #include "pcre.h"
1486
1487 I have a vague recollection that the change was concerned with compiling in
1488 different directories, but in the new build system, that is taken care of
1489 by the VPATH setting the Makefile.
1490
1491 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1492 when the subject happened to end in the byte 0x85 (e.g. if the last
1493 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1494 characters but of course it shouldn't be taken as a newline when it is part
1495 of another character. The bug was that, for an unlimited repeat of . in
1496 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1497 characters when looking for a newline.
1498
1499 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1500
1501 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1502 in debug output.
1503
1504 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1505 long printing in the pcrecpp unittest when running under MinGW.
1506
1507 6. ESC_K was left out of the EBCDIC table.
1508
1509 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1510 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1511 limit also applies to "virtual nesting" when a pattern is recursive, and in
1512 this case 1000 isn't so big. I have been able to remove this limit at the
1513 expense of backing off one optimization in certain circumstances. Normally,
1514 when pcre_exec() would call its internal match() function recursively and
1515 immediately return the result unconditionally, it uses a "tail recursion"
1516 feature to save stack. However, when a subpattern that can match an empty
1517 string has an unlimited repetition quantifier, it no longer makes this
1518 optimization. That gives it a stack frame in which to save the data for
1519 checking that an empty string has been matched. Previously this was taken
1520 from the 1000-entry workspace that had been reserved. So now there is no
1521 explicit limit, but more stack is used.
1522
1523 8. Applied Daniel's patches to solve problems with the import/export magic
1524 syntax that is required for Windows, and which was going wrong for the
1525 pcreposix and pcrecpp parts of the library. These were overlooked when this
1526 problem was solved for the main library.
1527
1528 9. There were some crude static tests to avoid integer overflow when computing
1529 the size of patterns that contain repeated groups with explicit upper
1530 limits. As the maximum quantifier is 65535, the maximum group length was
1531 set at 30,000 so that the product of these two numbers did not overflow a
1532 32-bit integer. However, it turns out that people want to use groups that
1533 are longer than 30,000 bytes (though not repeat them that many times).
1534 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1535 made it possible to implement the integer overflow checks in a much more
1536 dynamic way, which I have now done. The artificial limitation on group
1537 length has been removed - we now have only the limit on the total length of
1538 the compiled pattern, which depends on the LINK_SIZE setting.
1539
1540 10. Fixed a bug in the documentation for get/copy named substring when
1541 duplicate names are permitted. If none of the named substrings are set, the
1542 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1543 empty string.
1544
1545 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1546 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1547 because the ] is interpreted as the first data character and the
1548 terminating ] is not found. PCRE has been made compatible with Perl in this
1549 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1550 cause memory overwriting.
1551
1552 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1553 string has been matched (to stop an infinite loop). It was not recognizing
1554 a conditional subpattern that could match an empty string if that
1555 subpattern was within another subpattern. For example, it looped when
1556 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1557 condition was not nested. This bug has been fixed.
1558
1559 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1560 past the start of the subject in the presence of bytes with the top bit
1561 set, for example "\x8aBCD".
1562
1563 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1564 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1565
1566 14. Optimized (?!) to (*FAIL).
1567
1568 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1569 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1570 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1571 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1572 does: it's just the validity check that is more restrictive.
1573
1574 16. Inserted checks for integer overflows during escape sequence (backslash)
1575 processing, and also fixed erroneous offset values for syntax errors during
1576 backslash processing.
1577
1578 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1579 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1580
1581 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1582 caused an overrun.
1583
1584 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1585 something other than just ASCII characters) inside a group that had an
1586 unlimited repeat caused a loop at compile time (while checking to see
1587 whether the group could match an empty string).
1588
1589 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1590 [\P{Any}] did so. (Error in the code for printing property names.)
1591
1592 21. An orphan \E inside a character class could cause a crash.
1593
1594 22. A repeated capturing bracket such as (A)? could cause a wild memory
1595 reference during compilation.
1596
1597 23. There are several functions in pcre_compile() that scan along a compiled
1598 expression for various reasons (e.g. to see if it's fixed length for look
1599 behind). There were bugs in these functions when a repeated \p or \P was
1600 present in the pattern. These operators have additional parameters compared
1601 with \d, etc, and these were not being taken into account when moving along
1602 the compiled data. Specifically:
1603
1604 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1605 length.
1606
1607 (b) An item such as \pL+ within a repeated group could cause crashes or
1608 loops.
1609
1610 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1611 "reference to non-existent subpattern" error.
1612
1613 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1614
1615 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1616 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1617
1618 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1619
1620 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1621 character were causing crashes (broken optimization).
1622
1623 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1624 \p or \P) caused a compile-time loop.
1625
1626 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1627 break. For example, the unstudied pattern [\r\n]A does not match the string
1628 "\r\nA" because change 7.0/46 below moves the current point on by two
1629 characters after failing to match at the start. However, the pattern \nA
1630 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1631 the same is true. There doesn't seem any very clean way out of this, but
1632 what I have chosen to do makes the common cases work: PCRE now takes note
1633 of whether there can be an explicit match for \r or \n anywhere in the
1634 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1635 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1636 pattern has explicit CR or LF references.
1637
1638 29. Added (*CR) etc for changing newline setting at start of pattern.
1639
1640
1641 Version 7.2 19-Jun-07
1642 ---------------------
1643
1644 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1645 which is apparently normally available under Windows.
1646
1647 2. Re-jig the pcregrep tests with different newline settings in an attempt
1648 to make them independent of the local environment's newline setting.
1649
1650 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1651
1652 4. Some of the "internals" tests were previously cut out when the link size
1653 was not 2, because the output contained actual offsets. The recent new
1654 "Z" feature of pcretest means that these can be cut out, making the tests
1655 usable with all link sizes.
1656
1657 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1658 stack recursion. This gives a massive performance boost under BSD, but just
1659 a small improvement under Linux. However, it saves one field in the frame
1660 in all cases.
1661
1662 6. Added more features from the forthcoming Perl 5.10:
1663
1664 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1665 recursion call. It refers to the nth most recently opened parentheses.
1666
1667 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1668 to be opened parentheses.
1669
1670 (c) Conditions that refer to capturing parentheses can be specified
1671 relatively, for example, (?(-2)... or (?(+3)...
1672
1673 (d) \K resets the start of the current match so that everything before
1674 is not part of it.
1675
1676 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1677
1678 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1679 reference syntax.
1680
1681 (g) (?| introduces a group in which the numbering of parentheses in each
1682 alternative starts with the same number.
1683
1684 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1685
1686 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1687 PCRE_INFO_JCHANGED.
1688
1689 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1690 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1691 for detecting groups that can match an empty string.
1692
1693 9. A pattern with a very large number of alternatives (more than several
1694 hundred) was running out of internal workspace during the pre-compile
1695 phase, where pcre_compile() figures out how much memory will be needed. A
1696 bit of new cunning has reduced the workspace needed for groups with
1697 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1698 workspace instead of running out of the 4096 that are available.
1699
1700 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1701
1702 11. Applied patch from Google to remove an optimization that didn't quite work.
1703 The report of the bug said:
1704
1705 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1706 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1707 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1708
1709 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1710 it matched the wrong number of bytes.
1711
1712
1713 Version 7.1 24-Apr-07
1714 ---------------------
1715
1716 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1717 that is more "standard", making use of automake and other Autotools. There
1718 is some re-arrangement of the files and adjustment of comments consequent
1719 on this.
1720
1721 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1722 for recursive directory scanning broke on some systems because the files
1723 are not scanned in any specific order and on different systems the order
1724 was different. A call to "sort" has been inserted into RunGrepTest for the
1725 approprate test as a short-term fix. In the longer term there may be an
1726 alternative.
1727
1728 3. I had an email from Eric Raymond about problems translating some of PCRE's
1729 man pages to HTML (despite the fact that I distribute HTML pages, some
1730 people do their own conversions for various reasons). The problems
1731 concerned the use of low-level troff macros .br and .in. I have therefore
1732 removed all such uses from the man pages (some were redundant, some could
1733 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1734 HTML has been updated to handle .nf/.fi and to complain if it encounters
1735 .br or .in.
1736
1737 4. Updated comments in configure.ac that get placed in config.h.in and also
1738 arranged for config.h to be included in the distribution, with the name
1739 config.h.generic, for the benefit of those who have to compile without
1740 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1741
1742 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1743 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1744 makevp.bat for the current PCRE, using the additional files
1745 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1746
1747 6. A Windows user reported a minor discrepancy with test 2, which turned out
1748 to be caused by a trailing space on an input line that had got lost in his
1749 copy. The trailing space was an accident, so I've just removed it.
1750
1751 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1752 that is needed.
1753
1754 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1755 as "const" (a) because they are and (b) because it helps the PHP
1756 maintainers who have recently made a script to detect big data structures
1757 in the php code that should be moved to the .rodata section. I remembered
1758 to update Builducptable as well, so it won't revert if ucptable.h is ever
1759 re-created.
1760
1761 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1762 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1763 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1764 support is not required. This saves 1.5-2K of code, which is important in
1765 some applications.
1766
1767 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1768 so as not to refer to the tables, even though these functions will never be
1769 called when UTF-8 support is disabled. Otherwise there are problems with a
1770 shared library.
1771
1772 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1773
1774 (a) It was defining its arguments as char * instead of void *.
1775
1776 (b) It was assuming that all moves were upwards in memory; this was true
1777 a long time ago when I wrote it, but is no longer the case.
1778
1779 The emulated memove() is provided for those environments that have neither
1780 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1781 is clearly not the case, as these two bugs were recently reported.
1782
1783 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1784 and Detrail to create the HTML documentation, the .txt form of the man
1785 pages, and it removes trailing spaces from listed files. It also creates
1786 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1787 case, it wraps all the #defines with #ifndefs. This script should be run
1788 before "make dist".
1789
1790 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1791 with Unicode property support.
1792
1793 (a) For a maximizing quantifier, if the two different cases of the
1794 character were of different lengths in their UTF-8 codings (there are
1795 some cases like this - I found 11), and the matching function had to
1796 back up over a mixture of the two cases, it incorrectly assumed they
1797 were both the same length.
1798
1799 (b) When PCRE was configured to use the heap rather than the stack for
1800 recursion during matching, it was not correctly preserving the data for
1801 the other case of a UTF-8 character when checking ahead for a match
1802 while processing a minimizing repeat. If the check also involved
1803 matching a wide character, but failed, corruption could cause an
1804 erroneous result when trying to check for a repeat of the original
1805 character.
1806
1807 13. Some tidying changes to the testing mechanism:
1808
1809 (a) The RunTest script now detects the internal link size and whether there
1810 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1811 values substituted by "configure". (The RunGrepTest script already did
1812 this for UTF-8.) The configure.ac script no longer substitutes the
1813 relevant variables.
1814
1815 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1816 with length and offset values. This means that the output is different
1817 for different internal link sizes. Test 2 is skipped for link sizes
1818 other than 2 because of this, bypassing the problem. Unfortunately,
1819 there was also a test in test 3 (the locale tests) that used /B and
1820 failed for link sizes other than 2. Rather than cut the whole test out,
1821 I have added a new /Z option to pcretest that replaces the length and
1822 offset values with spaces. This is now used to make test 3 independent
1823 of link size. (Test 2 will be tidied up later.)
1824
1825 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1826 segmentation fault instead of returning the appropriate error message.
1827
1828 15. In multiline mode when the newline sequence was set to "any", the pattern
1829 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1830 This doesn't seem right; it now treats the CRLF combination as the line
1831 ending, and so does not match in that case. It's only a pattern such as ^$
1832 that would hit this one: something like ^ABC$ would have failed after \r
1833 and then tried again after \r\n.
1834
1835 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1836 in an attempt to make files that differ only in their line terminators
1837 compare equal. This works on Linux.
1838
1839 17. Under certain error circumstances pcregrep might try to free random memory
1840 as it exited. This is now fixed, thanks to valgrind.
1841
1842 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1843 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1844 was because its rules for how to advance for /g after matching an empty
1845 string at the end of a line did not allow for this case. They now check for
1846 it specially.
1847
1848 20. pcretest is supposed to handle patterns and data of any length, by
1849 extending its buffers when necessary. It was getting this wrong when the
1850 buffer for a data line had to be extended.
1851
1852 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1853 CRLF as a newline sequence.
1854
1855 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1856 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1857 I have nevertheless tidied it up.
1858
1859 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1860
1861 24. Added a man page for pcre-config.
1862
1863
1864 Version 7.0 19-Dec-06
1865 ---------------------
1866
1867 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1868 moving to gcc 4.1.1.
1869
1870 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1871 sys/time.h, which is documented as needed for this function. It doesn't
1872 seem to matter on Linux, but it showed up on some releases of OS X.
1873
1874 3. It seems that there are systems where bytes whose values are greater than
1875 127 match isprint() in the "C" locale. The "C" locale should be the
1876 default when a C program starts up. In most systems, only ASCII printing
1877 characters match isprint(). This difference caused the output from pcretest
1878 to vary, making some of the tests fail. I have changed pcretest so that:
1879
1880 (a) When it is outputting text in the compiled version of a pattern, bytes
1881 other than 32-126 are always shown as hex escapes.
1882
1883 (b) When it is outputting text that is a matched part of a subject string,
1884 it does the same, unless a different locale has been set for the match
1885 (using the /L modifier). In this case, it uses isprint() to decide.
1886
1887 4. Fixed a major bug that caused incorrect computation of the amount of memory
1888 required for a compiled pattern when options that changed within the
1889 pattern affected the logic of the preliminary scan that determines the
1890 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1891 that the computed length was too small. The symptoms of this bug were
1892 either the PCRE error "internal error: code overflow" from pcre_compile(),
1893 or a glibc crash with a message such as "pcretest: free(): invalid next
1894 size (fast)". Examples of patterns that provoked this bug (shown in
1895 pcretest format) are:
1896
1897 /(?-x: )/x
1898 /(?x)(?-x: \s*#\s*)/
1899 /((?i)[\x{c0}])/8
1900 /(?i:[\x{c0}])/8
1901
1902 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1903 is now done differently.
1904
1905 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1906 wrapper classes; (b) implement a new function in the C++ scanner that is
1907 more efficient than the old way of doing things because it avoids levels of
1908 recursion in the regex matching; (c) add a paragraph to the documentation
1909 for the FullMatch() function.
1910
1911 6. The escape sequence \n was being treated as whatever was defined as
1912 "newline". Not only was this contrary to the documentation, which states
1913 that \n is character 10 (hex 0A), but it also went horribly wrong when
1914 "newline" was defined as CRLF. This has been fixed.
1915
1916 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1917 was being set to -1 for the "end of line" case (supposedly a value that no
1918 character can have). Though this value is never used (the check for end of
1919 line is "zero bytes in current character"), it caused compiler complaints.
1920 I've changed it to 0xffffffff.
1921
1922 8. In pcre_version.c, the version string was being built by a sequence of
1923 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1924 string (as it is for production releases) called a macro with an empty
1925 argument. The C standard says the result of this is undefined. The gcc
1926 compiler treats it as an empty string (which was what was wanted) but it is
1927 reported that Visual C gives an error. The source has been hacked around to
1928 avoid this problem.
1929
1930 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1931 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1932 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1933 of them did).
1934
1935 10. Originally, pcretest opened its input and output without "b"; then I was
1936 told that "b" was needed in some environments, so it was added for release
1937 5.0 to both the input and output. (It makes no difference on Unix-like
1938 systems.) Later I was told that it is wrong for the input on Windows. I've
1939 now abstracted the modes into two macros, to make it easier to fiddle with
1940 them, and removed "b" from the input mode under Windows.
1941
1942 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1943
1944 12. Added -help and --help to pcretest as an official way of being reminded
1945 of the options.
1946
1947 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1948 and pcrecpp.cc because they annoy compilers at high warning levels.
1949
1950 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1951
1952 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1953 scripts are not C programs :-) and which was not noticed because it works
1954 on Linux.
1955
1956 16. pcretest is supposed to handle any length of pattern and data line (as one
1957 line or as a continued sequence of lines) by extending its input buffer if
1958 necessary. This feature was broken for very long pattern lines, leading to
1959 a string of junk being passed to pcre_compile() if the pattern was longer
1960 than about 50K.
1961
1962 17. I have done a major re-factoring of the way pcre_compile() computes the
1963 amount of memory needed for a compiled pattern. Previously, there was code
1964 that made a preliminary scan of the pattern in order to do this. That was
1965 OK when PCRE was new, but as the facilities have expanded, it has become
1966 harder and harder to keep it in step with the real compile phase, and there
1967 have been a number of bugs (see for example, 4 above). I have now found a
1968 cunning way of running the real compile function in a "fake" mode that
1969 enables it to compute how much memory it would need, while actually only
1970 ever using a few hundred bytes of working memory and without too many
1971 tests of the mode. This should make future maintenance and development
1972 easier. A side effect of this work is that the limit of 200 on the nesting
1973 depth of parentheses has been removed (though this was never a serious
1974 limitation, I suspect). However, there is a downside: pcre_compile() now
1975 runs more slowly than before (30% or more, depending on the pattern). I
1976 hope this isn't a big issue. There is no effect on runtime performance.
1977
1978 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1979 newline (only possible for the last line of a file) and it was a
1980 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1981
1982 19. Added additional timing features to pcretest. (1) The -tm option now times
1983 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1984 separate command line item, by a number that specifies the number of
1985 repeats to use when timing. The default is 50000; this gives better
1986 precision, but takes uncomfortably long for very large patterns.
1987
1988 20. Extended pcre_study() to be more clever in cases where a branch of a
1989 subpattern has no definite first character. For example, (a*|b*)[cd] would
1990 previously give no result from pcre_study(). Now it recognizes that the
1991 first character must be a, b, c, or d.
1992
1993 21. There was an incorrect error "recursive call could loop indefinitely" if
1994 a subpattern (or the entire pattern) that was being tested for matching an
1995 empty string contained only one non-empty item after a nested subpattern.
1996 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1997 incorrectly, because the \d was being skipped in the check.
1998
1999 22. The pcretest program now has a new pattern option /B and a command line
2000 option -b, which is equivalent to adding /B to every pattern. This causes
2001 it to show the compiled bytecode, without the additional information that
2002 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2003 is the same as /B/I).
2004
2005 23. A new optimization is now able automatically to treat some sequences such
2006 as a*b as a*+b. More specifically, if something simple (such as a character
2007 or a simple class like \d) has an unlimited quantifier, and is followed by
2008 something that cannot possibly match the quantified thing, the quantifier
2009 is automatically "possessified".
2010
2011 24. A recursive reference to a subpattern whose number was greater than 39
2012 went wrong under certain circumstances in UTF-8 mode. This bug could also
2013 have affected the operation of pcre_study().
2014
2015 25. Realized that a little bit of performance could be had by replacing
2016 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2017
2018 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2019
2020 27. Possessive quantifiers such as a++ were previously implemented by turning
2021 them into atomic groups such as ($>a+). Now they have their own opcodes,
2022 which improves performance. This includes the automatically created ones
2023 from 23 above.
2024
2025 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2026 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2027 the first matched character to be a colon. This applied both to named and
2028 numbered groups.
2029
2030 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2031
2032 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2033 building PCRE on Windows easier, so I have included it in the distribution.
2034
2035 31. There is now a check in pcretest against a ridiculously large number being
2036 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2037 loop, the loop is abandoned.
2038
2039 32. Forward references to subpatterns in conditions such as (?(2)...) where
2040 subpattern 2 is defined later cause pcre_compile() to search forwards in
2041 the pattern for the relevant set of parentheses. This search went wrong
2042 when there were unescaped parentheses in a character class, parentheses
2043 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2044
2045 33. "Subroutine" calls and backreferences were previously restricted to
2046 referencing subpatterns earlier in the regex. This restriction has now
2047 been removed.
2048
2049 34. Added a number of extra features that are going to be in Perl 5.10. On the
2050 whole, these are just syntactic alternatives for features that PCRE had
2051 previously implemented using the Python syntax or my own invention. The
2052 other formats are all retained for compatibility.
2053
2054 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2055 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2056 also .NET compatible.
2057
2058 (b) A recursion or subroutine call to a named group can now be defined as
2059 (?&name) as well as (?P>name).
2060
2061 (c) A backreference to a named group can now be defined as \k<name> or
2062 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2063 5.10, are also .NET compatible.
2064
2065 (d) A conditional reference to a named group can now use the syntax
2066 (?(<name>) or (?('name') as well as (?(name).
2067
2068 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2069 groups (named and numbered) that are never evaluated inline, but can be
2070 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2071 is always false. There may be only one alternative in such a group.
2072
2073 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2074 as the simple (?(R). The condition is true only if the most recent
2075 recursion is that of the given number or name. It does not search out
2076 through the entire recursion stack.
2077
2078 (g) The escape \gN or \g{N} has been added, where N is a positive or
2079 negative number, specifying an absolute or relative reference.
2080
2081 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2082 some "unreachable code" warnings.
2083
2084 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2085 things, this adds five new scripts.
2086
2087 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2088 There were also incompatibilities regarding the handling of \Q..\E inside
2089 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2090 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2091
2092 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2093 matches an empty string, and forcibly breaks the loop. There were bugs in
2094 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2095 against aaaa the result was just "a" rather than "aaaa", for example. Two
2096 separate and independent bugs (that affected different cases) have been
2097 fixed.
2098
2099 39. Refactored the code to abolish the use of different opcodes for small
2100 capturing bracket numbers. This is a tidy that I avoided doing when I
2101 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2102 The new approach is not only tidier, it makes it possible to reduce the
2103 memory needed to fix the previous bug (38).
2104
2105 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2106 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2107 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2108 mode.
2109
2110 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2111 report.
2112
2113 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2114 copy construction and assignment in the C++ wrapper.
2115
2116 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2117 couple of bugs that could have given wrong results in the "--newline=crlf"
2118 case.
2119
2120 44. Added a number of casts and did some reorganization of signed/unsigned int
2121 variables following suggestions from Dair Grant. Also renamed the variable
2122 "this" as "item" because it is a C++ keyword.
2123
2124 45. Arranged for dftables to add
2125
2126 #include "pcre_internal.h"
2127
2128 to pcre_chartables.c because without it, gcc 4.x may remove the array
2129 definition from the final binary if PCRE is built into a static library and
2130 dead code stripping is activated.
2131
2132 46. For an unanchored pattern, if a match attempt fails at the start of a
2133 newline sequence, and the newline setting is CRLF or ANY, and the next two
2134 characters are CRLF, advance by two characters instead of one.
2135
2136
2137 Version 6.7 04-Jul-06
2138 ---------------------
2139
2140 1. In order to handle tests when input lines are enormously long, pcretest has
2141 been re-factored so that it automatically extends its buffers when
2142 necessary. The code is crude, but this _is_ just a test program. The
2143 default size has been increased from 32K to 50K.
2144
2145 2. The code in pcre_study() was using the value of the re argument before
2146 testing it for NULL. (Of course, in any sensible call of the function, it
2147 won't be NULL.)
2148
2149 3. The memmove() emulation function in pcre_internal.h, which is used on
2150 systems that lack both memmove() and bcopy() - that is, hardly ever -
2151 was missing a "static" storage class specifier.
2152
2153 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2154 containing an extended class (one that cannot be represented by a bitmap
2155 because it contains high-valued characters or Unicode property items, e.g.
2156 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2157 pattern, but PCRE should not loop if you do not (it no longer does).
2158 [Detail: two cases were found: (a) a repeated subpattern containing an
2159 extended class; (b) a recursive reference to a subpattern that followed a
2160 previous extended class. It wasn't skipping over the extended class
2161 correctly when UTF-8 mode was not set.]
2162
2163 5. A negated single-character class was not being recognized as fixed-length
2164 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2165 compile error "lookbehind assertion is not fixed length".
2166
2167 6. The RunPerlTest auxiliary script was showing an unexpected difference
2168 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2169 write a Perl script that can interpret lines of an input file either as
2170 byte characters or as UTF-8, which is what "perltest" was being required to
2171 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2172 can't do is switch easily at run time between having the "use utf8;" pragma
2173 or not. In the end, I fudged it by using the RunPerlTest script to insert
2174 "use utf8;" explicitly for the UTF-8 tests.
2175
2176 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2177 the end of the subject string, contrary to the documentation and to what
2178 Perl does. This was true of both matching functions. Now it matches only at
2179 the start of the subject and immediately after *internal* newlines.
2180
2181 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2182 a pointer to an int instead of a pointer to an unsigned long int. This
2183 caused problems on 64-bit systems.
2184
2185 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2186 instance of the 'standard' template library not being so standard".
2187
2188 10. There was no check on the number of named subpatterns nor the maximum
2189 length of a subpattern name. The product of these values is used to compute
2190 the size of the memory block for a compiled pattern. By supplying a very
2191 long subpattern name and a large number of named subpatterns, the size
2192 computation could be caused to overflow. This is now prevented by limiting
2193 the length of names to 32 characters, and the number of named subpatterns
2194 to 10,000.
2195
2196 11. Subpatterns that are repeated with specific counts have to be replicated in
2197 the compiled pattern. The size of memory for this was computed from the
2198 length of the subpattern and the repeat count. The latter is limited to
2199 65535, but there was no limit on the former, meaning that integer overflow
2200 could in principle occur. The compiled length of a repeated subpattern is
2201 now limited to 30,000 bytes in order to prevent this.
2202
2203 12. Added the optional facility to have named substrings with the same name.
2204
2205 13. Added the ability to use a named substring as a condition, using the
2206 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2207 are numbers (not recommended). Forward references are permitted.
2208
2209 14. Added forward references in named backreferences (if you see what I mean).
2210
2211 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2212 pattern could run off the end of the subject. For example, the pattern
2213 "(?s)(.{1,5})"8 did this with the subject "ab".
2214
2215 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2216 PCRE_CASELESS was set when matching characters that were quantified with ?
2217 or *.
2218
2219 17. A character class other than a single negated character that had a minimum
2220 but no maximum quantifier - for example [ab]{6,} - was not handled
2221 correctly by pce_dfa_exec(). It would match only one character.
2222
2223 18. A valid (though odd) pattern that looked like a POSIX character
2224 class but used an invalid character after [ (for example [[,abc,]]) caused
2225 pcre_compile() to give the error "Failed: internal error: code overflow" or
2226 in some cases to crash with a glibc free() error. This could even happen if
2227 the pattern terminated after [[ but there just happened to be a sequence of
2228 letters, a binary zero, and a closing ] in the memory that followed.
2229
2230 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2231 over the years. Originally (before any Unicode support), just the bottom 8
2232 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2233 output from "man perlunicode" includes this:
2234
2235 The regular expression compiler produces polymorphic opcodes. That
2236 is, the pattern adapts to the data and automatically switches to
2237 the Unicode character scheme when presented with Unicode data--or
2238 instead uses a traditional byte scheme when presented with byte
2239 data.
2240
2241 Sadly, a wide octal escape does not cause a switch, and in a string with
2242 no other multibyte characters, these octal escapes are treated as before.
2243 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2244 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2245 Unicode string.
2246
2247 I have not perpetrated such confusion in PCRE. Up till now, it took just
2248 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2249 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2250 translate to the appropriate multibyte character.
2251
2252 29. Applied some refactoring to reduce the number of warnings from Microsoft
2253 and Borland compilers. This has included removing the fudge introduced
2254 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2255 a warning about an unused variable.
2256
2257 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2258 characters since release 4.0, because Perl (from release 5.004) does not.
2259 [Or at least, is documented not to: some releases seem to be in conflict
2260 with the documentation.] However, when a pattern was studied with
2261 pcre_study() and all its branches started with \s, PCRE still included VT
2262 as a possible starting character. Of course, this did no harm; it just
2263 caused an unnecessary match attempt.
2264
2265 22. Removed a now-redundant internal flag bit that recorded the fact that case
2266 dependency changed within the pattern. This was once needed for "required
2267 byte" processing, but is no longer used. This recovers a now-scarce options
2268 bit. Also moved the least significant internal flag bit to the most-
2269 significant bit of the word, which was not previously used (hangover from
2270 the days when it was an int rather than a uint) to free up another bit for
2271 the future.
2272
2273 23. Added support for CRLF line endings as well as CR and LF. As well as the
2274 default being selectable at build time, it can now be changed at runtime
2275 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2276 specify that it is scanning data with non-default line endings.
2277
2278 24. Changed the definition of CXXLINK to make it agree with the definition of
2279 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2280
2281 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2282 recursions. This makes a big different to stack usage for some patterns.
2283
2284 26. If a subpattern containing a named recursion or subroutine reference such
2285 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2286 the space required for the compiled pattern went wrong and gave too small a
2287 value. Depending on the environment, this could lead to "Failed: internal
2288 error: code overflow at offset 49" or "glibc detected double free or
2289 corruption" errors.
2290
2291 27. Applied patches from Google (a) to support the new newline modes and (b) to
2292 advance over multibyte UTF-8 characters in GlobalReplace.
2293
2294 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2295 difference for some implementation of PCRE in some Windows version.
2296
2297 29. Added some extra testing facilities to pcretest:
2298
2299 \q<number> in a data line sets the "match limit" value
2300 \Q<number> in a data line sets the "match recursion limt" value
2301 -S <number> sets the stack size, where <number> is in megabytes
2302
2303 The -S option isn't available for Windows.
2304
2305
2306 Version 6.6 06-Feb-06
2307 ---------------------
2308
2309 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2310 in pcreposix.h. I have copied the definition from pcre.h.
2311
2312 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2313 because pcre.h is no longer a built file.
2314
2315 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2316 not normally included in the compiled code.
2317
2318
2319 Version 6.5 01-Feb-06
2320 ---------------------
2321
2322 1. When using the partial match feature with pcre_dfa_exec(), it was not
2323 anchoring the second and subsequent partial matches at the new starting
2324 point. This could lead to incorrect results. For example, with the pattern
2325 /1234/, partially matching against "123" and then "a4" gave a match.
2326
2327 2. Changes to pcregrep:
2328
2329 (a) All non-match returns from pcre_exec() were being treated as failures
2330 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2331 error message is output. Some extra information is given for the
2332 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2333 probably the only errors that are likely to be caused by users (by
2334 specifying a regex that has nested indefinite repeats, for instance).
2335 If there are more than 20 of these errors, pcregrep is abandoned.
2336
2337 (b) A binary zero was treated as data while matching, but terminated the
2338 output line if it was written out. This has been fixed: binary zeroes
2339 are now no different to any other data bytes.
2340
2341 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2342 used to set a locale for matching. The --locale=xxxx long option has
2343 been added (no short equivalent) to specify a locale explicitly on the
2344 pcregrep command, overriding the environment variables.
2345
2346 (d) When -B was used with -n, some line numbers in the output were one less
2347 than they should have been.
2348
2349 (e) Added the -o (--only-matching) option.
2350
2351 (f) If -A or -C was used with -c (count only), some lines of context were
2352 accidentally printed for the final match.
2353
2354 (g) Added the -H (--with-filename) option.
2355
2356 (h) The combination of options -rh failed to suppress file names for files
2357 that were found from directory arguments.
2358
2359 (i) Added the -D (--devices) and -d (--directories) options.
2360
2361 (j) Added the -F (--fixed-strings) option.
2362
2363 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2364
2365 (l) Added the --colo(u)r option.
2366
2367 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2368 is not present by default.
2369
2370 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2371 items such as (?R) or (?1), when the recursion could match a number of
2372 alternatives. If it matched one of the alternatives, but subsequently,
2373 outside the recursion, there was a failure, the code tried to back up into
2374 the recursion. However, because of the way PCRE is implemented, this is not
2375 possible, and the result was an incorrect result from the match.
2376
2377 In order to prevent this happening, the specification of recursion has
2378 been changed so that all such subpatterns are automatically treated as
2379 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2380
2381 4. I had overlooked the fact that, in some locales, there are characters for
2382 which isalpha() is true but neither isupper() nor islower() are true. In
2383 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2384 and ordfeminine) are like this. This affected the treatment of \w and \W
2385 when they appeared in character classes, but not when they appeared outside
2386 a character class. The bit map for "word" characters is now created
2387 separately from the results of isalnum() instead of just taking it from the
2388 upper, lower, and digit maps. (Plus the underscore character, of course.)
2389
2390 5. The above bug also affected the handling of POSIX character classes such as
2391 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2392 permanent tables. Instead, the bit maps for such a class were previously
2393 created as the appropriate unions of the upper, lower, and digit bitmaps.
2394 Now they are created by subtraction from the [[:word:]] class, which has
2395 its own bitmap.
2396
2397 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2398 It is created by subtracting the vertical space characters (\x09, \x0a,
2399 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2400 subtraction was done in the overall bitmap for a character class, meaning
2401 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2402 be recognized. This bug has been fixed.
2403
2404 7. Patches from the folks at Google:
2405
2406 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2407 real life, but is still worth protecting against".
2408
2409 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2410 regular expressions".
2411
2412 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2413 have it.
2414
2415 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2416 "configure" and the latter not, in order to fix a problem somebody had
2417 with compiling the Arg class on HP-UX.
2418
2419 (e) Improve the error-handling of the C++ wrapper a little bit.
2420
2421 (f) New tests for checking recursion limiting.
2422
2423 8. The pcre_memmove() function, which is used only if the environment does not
2424 have a standard memmove() function (and is therefore rarely compiled),
2425 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2426 returning a result (though PCRE never actually uses the result).
2427
2428 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2429 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2430 returned instead of calling malloc() with an overflowing number that would
2431 most likely cause subsequent chaos.
2432
2433 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2434
2435 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2436 with this option is matched, the nmatch and pmatch options of regexec() are
2437 ignored.
2438
2439 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2440 provided in case anyone wants to the the POSIX interface with UTF-8
2441 strings.
2442
2443 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2444 C++ linking (needed for some HP-UX environments).
2445
2446 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2447 (unused parameter) and in the pcre_printint() function (omitted "default"
2448 switch label when the default is to do nothing).
2449
2450 15. Added some code to make it possible, when PCRE is compiled as a C++
2451 library, to replace subject pointers for pcre_exec() with a smart pointer
2452 class, thus making it possible to process discontinuous strings.
2453
2454 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2455 much the same function. They were added by different people who were trying
2456 to make PCRE easy to compile on non-Unix systems. It has been suggested
2457 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2458 for compiling on Windows systems. I have therefore replaced it with
2459 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2460 defaults to "extern" for C or "extern C" for C++, which works fine on
2461 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2462 SCOPE with something explicit in config.h. In addition:
2463
2464 (a) pcreposix.h still had just "extern" instead of either of these macros;
2465 I have replaced it with PCRE_DATA_SCOPE.
2466
2467 (b) Functions such as _pcre_xclass(), which are internal to the library,
2468 but external in the C sense, all had PCRE_EXPORT in their definitions.
2469 This is apparently wrong for the Windows case, so I have removed it.
2470 (It makes no difference on Unix-like systems.)
2471
2472 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2473 of recursive calls to match(). This is different to MATCH_LIMIT because
2474 that limits the total number of calls to match(), not all of which increase
2475 the depth of recursion. Limiting the recursion depth limits the amount of
2476 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2477 when PCRE is compiled, and changed at run time. A patch from Google adds
2478 this functionality to the C++ interface.
2479
2480 18. Changes to the handling of Unicode character properties:
2481
2482 (a) Updated the table to Unicode 4.1.0.
2483
2484 (b) Recognize characters that are not in the table as "Cn" (undefined).
2485
2486 (c) I revised the way the table is implemented to a much improved format
2487 which includes recognition of ranges. It now supports the ranges that
2488 are defined in UnicodeData.txt, and it also amalgamates other
2489 characters into ranges. This has reduced the number of entries in the
2490 table from around 16,000 to around 3,000, thus reducing its size
2491 considerably. I realized I did not need to use a tree structure after
2492 all - a binary chop search is just as efficient. Having reduced the
2493 number of entries, I extended their size from 6 bytes to 8 bytes to
2494 allow for more data.
2495
2496 (d) Added support for Unicode script names via properties such as \p{Han}.
2497
2498 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2499 matching that character.
2500
2501 20. When matching a repeated Unicode property with a minimum greater than zero,
2502 (for example \pL{2,}), PCRE could look past the end of the subject if it
2503 reached it while seeking the minimum number of characters. This could
2504 happen only if some of the characters were more than one byte long, because
2505 there is a check for at least the minimum number of bytes.
2506
2507 21. Refactored the implementation of \p and \P so as to be more general, to
2508 allow for more different types of property in future. This has changed the
2509 compiled form incompatibly. Anybody with saved compiled patterns that use
2510 \p or \P will have to recompile them.
2511
2512 22. Added "Any" and "L&" to the supported property types.
2513
2514 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2515 but give a compile time error if the value is greater than 0xff.
2516
2517 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2518 accidentally not being installed or uninstalled.
2519
2520 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2521 made were to insert the current release number. This seemed silly, because
2522 it made things harder for people building PCRE on systems that don't run
2523 "configure". I have turned pcre.h into a distributed file, no longer built
2524 by "configure", with the version identification directly included. There is
2525 no longer a pcre.h.in file.
2526
2527 However, this change necessitated a change to the pcre-config script as
2528 well. It is built from pcre-config.in, and one of the substitutions was the
2529 release number. I have updated configure.ac so that ./configure now finds
2530 the release number by grepping pcre.h.
2531
2532 26. Added the ability to run the tests under valgrind.
2533
2534
2535 Version 6.4 05-Sep-05
2536 ---------------------
2537
2538 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2539 "--" to be printed when multiple files were scanned, even when none of the
2540 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2541 consider it to be a bug, and have restored the previous behaviour.
2542
2543 2. A couple of code tidies to get rid of compiler warnings.
2544
2545 3. The pcretest program used to cheat by referring to symbols in the library
2546 whose names begin with _pcre_. These are internal symbols that are not
2547 really supposed to be visible externally, and in some environments it is
2548 possible to suppress them. The cheating is now confined to including
2549 certain files from the library's source, which is a bit cleaner.
2550
2551 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2552 file's purpose clearer.
2553
2554 5. Reorganized pcre_ucp_findchar().
2555
2556
2557 Version 6.3 15-Aug-05
2558 ---------------------
2559
2560 1. The file libpcre.pc.in did not have general read permission in the tarball.
2561
2562 2. There were some problems when building without C++ support:
2563
2564 (a) If C++ support was not built, "make install" and "make test" still
2565 tried to test it.
2566
2567 (b) There were problems when the value of CXX was explicitly set. Some
2568 changes have been made to try to fix these, and ...
2569
2570 (c) --disable-cpp can now be used to explicitly disable C++ support.
2571
2572 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2573 backslash in a target when C++ was disabled. This confuses some
2574 versions of "make", apparently. Using an intermediate variable solves
2575 this. (Same for CPP_LOBJ.)
2576
2577 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2578 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2579 necessary on certain architectures.
2580
2581 4. Added a setting of -export-symbols-regex to the link command to remove
2582 those symbols that are exported in the C sense, but actually are local
2583 within the library, and not documented. Their names all begin with
2584 "_pcre_". This is not a perfect job, because (a) we have to except some
2585 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2586 available (and never for static libraries). I have made a note to try to
2587 find a way round (a) in the future.
2588
2589
2590 Version 6.2 01-Aug-05
2591 ---------------------
2592
2593 1. There was no test for integer overflow of quantifier values. A construction
2594 such as {1111111111111111} would give undefined results. What is worse, if
2595 a minimum quantifier for a parenthesized subpattern overflowed and became
2596 negative, the calculation of the memory size went wrong. This could have
2597 led to memory overwriting.
2598
2599 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2600
2601 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2602 operating environments where this matters.
2603
2604 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2605 PCRE options from within the C++ wrapper.
2606
2607 5. Named capturing subpatterns were not being correctly counted when a pattern
2608 was compiled. This caused two problems: (a) If there were more than 100
2609 such subpatterns, the calculation of the memory needed for the whole
2610 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2611 back references of the form \12, where the number was greater than 9, were
2612 not recognized as back references, even though there were sufficient
2613 previous subpatterns.
2614
2615 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2616 versions of gcc, e.g. 2.95.4.
2617
2618
2619 Version 6.1 21-Jun-05
2620 ---------------------
2621
2622 1. There was one reference to the variable "posix" in pcretest.c that was not
2623 surrounded by "#if !defined NOPOSIX".
2624
2625 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2626 the cross-check on the old pcre_info() function, for the benefit of the
2627 cut-down version of PCRE that is currently imported into Exim.
2628
2629 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2630 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2631 patterns that start (?i) but I don't think that matters. The use of (?i) is
2632 just an example; this all applies to the other options as well.
2633
2634 4. Since libtool seems to echo the compile commands it is issuing, the output
2635 from "make" can be reduced a bit by putting "@" in front of each libtool
2636 compile command.
2637
2638 5. Patch from the folks at Google for configure.in to be a bit more thorough
2639 in checking for a suitable C++ installation before trying to compile the
2640 C++ stuff. This should fix a reported problem when a compiler was present,
2641 but no suitable headers.
2642
2643 6. The man pages all had just "PCRE" as their title. I have changed them to
2644 be the relevant file name. I have also arranged that these names are
2645 retained in the file doc/pcre.txt, which is a concatenation in text format
2646 of all the man pages except the little individual ones for each function.
2647
2648 7. The NON-UNIX-USE file had not been updated for the different set of source
2649 files that come with release 6. I also added a few comments about the C++
2650 wrapper.
2651
2652
2653 Version 6.0 07-Jun-05
2654 ---------------------
2655
2656 1. Some minor internal re-organization to help with my DFA experiments.
2657
2658 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2659 didn't matter for the library itself when fully configured, but did matter
2660 when compiling without UCP support, or within Exim, where the ucp files are
2661 not imported.
2662
2663 3. Refactoring of the library code to split up the various functions into
2664 different source modules. The addition of the new DFA matching code (see
2665 below) to a single monolithic source would have made it really too
2666 unwieldy, quite apart from causing all the code to be include in a
2667 statically linked application, when only some functions are used. This is
2668 relevant even without the DFA addition now that patterns can be compiled in
2669 one application and matched in another.
2670
2671 The downside of splitting up is that there have to be some external
2672 functions and data tables that are used internally in different modules of
2673 the library but which are not part of the API. These have all had their
2674 names changed to start with "_pcre_" so that they are unlikely to clash
2675 with other external names.
2676
2677 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2678 a different (DFA) algorithm. Although it is slower than the original
2679 function, it does have some advantages for certain types of matching
2680 problem.
2681
2682 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2683 including restarting after a partial match.
2684
2685 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2686 defined when compiling for Windows was sent to me. I have put it into the
2687 code, though I have no means of testing or verifying it.
2688
2689 7. Added the pcre_refcount() auxiliary function.
2690
2691 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2692 match before or at the first newline in the subject string. In pcretest,
2693 the /f option on a pattern can be used to set this.
2694
2695 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2696 would behave wrongly. This has been present in PCRE since release 4.0.
2697
2698 10. A number of changes to the pcregrep command:
2699
2700 (a) Refactored how -x works; insert ^(...)$ instead of setting
2701 PCRE_ANCHORED and checking the length, in preparation for adding
2702 something similar for -w.
2703
2704 (b) Added the -w (match as a word) option.
2705
2706 (c) Refactored the way lines are read and buffered so as to have more
2707 than one at a time available.
2708
2709 (d) Implemented a pcregrep test script.
2710
2711 (e) Added the -M (multiline match) option. This allows patterns to match
2712 over several lines of the subject. The buffering ensures that at least
2713 8K, or the rest of the document (whichever is the shorter) is available
2714 for matching (and similarly the previous 8K for lookbehind assertions).
2715
2716 (f) Changed the --help output so that it now says
2717
2718 -w, --word-regex(p)
2719
2720 instead of two lines, one with "regex" and the other with "regexp"
2721 because that confused at least one person since the short forms are the
2722 same. (This required a bit of code, as the output is generated
2723 automatically from a table. It wasn't just a text change.)
2724
2725 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2726 option but starts with a hyphen. Could be a pattern or a path name
2727 starting with a hyphen, for instance.
2728
2729 (h) "-" can be given as a file name to represent stdin.
2730
2731 (i) When file names are being printed, "(standard input)" is used for
2732 the standard input, for compatibility with GNU grep. Previously
2733 "<stdin>" was used.
2734
2735 (j) The option --label=xxx can be used to supply a name to be used for
2736 stdin when file names are being printed. There is no short form.
2737
2738 (k) Re-factored the options decoding logic because we are going to add
2739 two more options that take data. Such options can now be given in four
2740 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2741
2742 (l) Added the -A, -B, and -C options for requesting that lines of context
2743 around matches be printed.
2744
2745 (m) Added the -L option to print the names of files that do not contain
2746 any matching lines, that is, the complement of -l.
2747
2748 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2749 continue to scan other files.
2750
2751 (o) The -s option was incorrectly implemented. For compatibility with other
2752 greps, it now suppresses the error message for a non-existent or non-
2753 accessible file (but not the return code). There is a new option called
2754 -q that suppresses the output of matching lines, which was what -s was
2755 previously doing.
2756
2757 (p) Added --include and --exclude options to specify files for inclusion
2758 and exclusion when recursing.
2759
2760 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2761 Hopefully, it now does.
2762
2763 12. Missing cast in pcre_study().
2764
2765 13. Added an "uninstall" target to the makefile.
2766
2767 14. Replaced "extern" in the function prototypes in Makefile.in with
2768 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2769 world, but is set differently for Windows.
2770
2771 15. Added a second compiling function called pcre_compile2(). The only
2772 difference is that it has an extra argument, which is a pointer to an
2773 integer error code. When there is a compile-time failure, this is set
2774 non-zero, in addition to the error test pointer being set to point to an
2775 error message. The new argument may be NULL if no error number is required
2776 (but then you may as well call pcre_compile(), which is now just a
2777 wrapper). This facility is provided because some applications need a
2778 numeric error indication, but it has also enabled me to tidy up the way
2779 compile-time errors are handled in the POSIX wrapper.
2780
2781 16. Added VPATH=.libs to the makefile; this should help when building with one
2782 prefix path and installing with another. (Or so I'm told by someone who
2783 knows more about this stuff than I do.)
2784
2785 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2786 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2787 match everything, including newlines. This is not POSIX-compatible, but
2788 somebody wanted the feature. From pcretest it can be activated by using
2789 both the P and the s flags.
2790
2791 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2792
2793 19. libpcre.pc was being incorrectly installed as executable.
2794
2795 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2796 it now also looks for '\r' so that it will work unmodified on Windows.
2797
2798 21. Added Google's contributed C++ wrapper to the distribution.
2799
2800 22. Added some untidy missing memory free() calls in pcretest, to keep
2801 Electric Fence happy when testing.
2802
2803
2804
2805 Version 5.0 13-Sep-04
2806 ---------------------
2807
2808 1. Internal change: literal characters are no longer packed up into items
2809 containing multiple characters in a single byte-string. Each character
2810 is now matched using a separate opcode. However, there may be more than one
2811 byte in the character in UTF-8 mode.
2812
2813 2. The pcre_callout_block structure has two new fields: pattern_position and
2814 next_item_length. These contain the offset in the pattern to the next match
2815 item, and its length, respectively.
2816
2817 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2818 insertion of callouts before each pattern item. Added the /C option to
2819 pcretest to make use of this.
2820
2821 4. On the advice of a Windows user, the lines
2822
2823 #if defined(_WIN32) || defined(WIN32)
2824 _setmode( _fileno( stdout ), 0x8000 );
2825 #endif /* defined(_WIN32) || defined(WIN32) */
2826
2827 have been added to the source of pcretest. This apparently does useful
2828 magic in relation to line terminators.
2829
2830 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2831 for the benefit of those environments where the "b" makes a difference.
2832
2833 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2834 to know about it. I have put a hack into configure.in that adds in code
2835 to set GCC=yes if CC=icc. This seems to end up at a point in the
2836 generated configure script that is early enough to affect the setting of
2837 compiler options, which is what is needed, but I have no means of testing
2838 whether it really works. (The user who reported this had patched the
2839 generated configure script, which of course I cannot do.)
2840
2841 LATER: After change 22 below (new libtool files), the configure script
2842 seems to know about icc (and also ecc). Therefore, I have commented out
2843 this hack in configure.in.
2844
2845 7. Added support for pkg-config (2 patches were sent in).
2846
2847 8. Negated POSIX character classes that used a combination of internal tables
2848 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2849 [[:^ascii]]. Typically, they would match almost any characters. The other
2850 POSIX classes were not broken in this way.
2851
2852 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2853 to find the match, as PCRE was deluded into thinking that the match had to
2854 start at the start point or following a newline. The same bug applied to
2855 patterns with negative forward assertions or any backward assertions
2856 preceding ".*" at the start, unless the pattern required a fixed first
2857 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2858
2859 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2860 starting at the last subject character, bytes beyond the end of the subject
2861 string were read.
2862
2863 11. Renamed the variable "class" as "classbits" to make life easier for C++
2864 users. (Previously there was a macro definition, but it apparently wasn't
2865 enough.)
2866
2867 12. Added the new field "tables" to the extra data so that tables can be passed
2868 in at exec time, or the internal tables can be re-selected. This allows
2869 a compiled regex to be saved and re-used at a later time by a different
2870 program that might have everything at different addresses.
2871
2872 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2873 -R library as well as a -L library.
2874
2875 14. The debugging options of pcretest (-d on the command line or D on a
2876 pattern) showed incorrect output for anything following an extended class
2877 that contained multibyte characters and which was followed by a quantifier.
2878
2879 15. Added optional support for general category Unicode character properties
2880 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2881 support. It adds about 90K to the size of the library. The meanings of the
2882 inbuilt class escapes such as \d and \s have NOT been changed.
2883
2884 16. Updated pcredemo.c to include calls to free() to release the memory for the
2885 compiled pattern.
2886
2887 17. The generated file chartables.c was being created in the source directory
2888 instead of in the building directory. This caused the build to fail if the
2889 source directory was different from the building directory, and was
2890 read-only.
2891
2892 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2893 file. No doubt somebody will tell me if they don't make sense... Also added
2894 Dan Mooney's comments about building on OpenVMS.
2895
2896 19. Added support for partial matching via the PCRE_PARTIAL option for
2897 pcre_exec() and the \P data escape in pcretest.
2898
2899 20. Extended pcretest with 3 new pattern features:
2900
2901 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2902 write the compiled pattern to the file whose name is "rest-of-line".
2903 This is a straight binary dump of the data, with the saved pointer to
2904 the character tables forced to be NULL. The study data, if any, is
2905 written too. After writing, pcretest reads a new pattern.
2906
2907 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2908 compiled pattern from the given file. There must not be any
2909 occurrences of "<" in the file name (pretty unlikely); if there are,
2910 pcretest will instead treat the initial "<" as a pattern delimiter.
2911 After reading in the pattern, pcretest goes on to read data lines as
2912 usual.
2913
2914 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2915 and 16-bit fields in a compiled pattern, to simulate a pattern that
2916 was compiled on a host of opposite endianness.
2917
2918 21. The pcre-exec() function can now cope with patterns that were compiled on
2919 hosts of opposite endianness, with this restriction:
2920
2921 As for any compiled expression that is saved and used later, the tables
2922 pointer field cannot be preserved; the extra_data field in the arguments
2923 to pcre_exec() should be used to pass in a tables address if a value
2924 other than the default internal tables were used at compile time.
2925
2926 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2927 now diagnosed as an error. Previously, most of the time, a negative number
2928 would have been treated as zero, but if in addition "ovector" was passed as
2929 NULL, a crash could occur.
2930
2931 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2932 new versions from the libtool 1.5 distribution (the last one is a copy of
2933 a file called libtool.m4). This seems to have fixed the need to patch
2934 "configure" to support Darwin 1.3 (which I used to do). However, I still
2935 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2936 workstation).
2937
2938 24. Changed the PCRE licence to be the more standard "BSD" licence.
2939
2940
2941 Version 4.5 01-Dec-03
2942 ---------------------
2943
2944 1. There has been some re-arrangement of the code for the match() function so
2945 that it can be compiled in a version that does not call itself recursively.
2946 Instead, it keeps those local variables that need separate instances for
2947 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2948 needs to "recurse". Keeping track of where control must go is done by means
2949 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2950 hide most of the details from the main code, and operates only if
2951 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2952 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2953 operating.
2954
2955 To make it easier for callers to provide specially tailored get/free
2956 functions for this usage, two new functions, pcre_stack_malloc, and
2957 pcre_stack_free, are used. They are always called in strict stacking order,
2958 and the size of block requested is always the same.
2959
2960 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2961 PCRE has been compiled to use the stack or the heap for recursion. The
2962 -C option of pcretest uses this to show which version is compiled.
2963
2964 A new data escape \S, is added to pcretest; it causes the amounts of store
2965 obtained and freed by both kinds of malloc/free at match time to be added
2966 to the output.
2967
2968 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2969 what's available on my current Linux desktop machine.
2970
2971 3. When matching a UTF-8 string, the test for a valid string at the start has
2972 been extended. If start_offset is not zero, PCRE now checks that it points
2973 to a byte that is the start of a UTF-8 character. If not, it returns
2974 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2975 this is necessary because there may be backward assertions in the pattern.
2976 When matching the same subject several times, it may save resources to use
2977 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2978
2979 4. The code for checking the validity of UTF-8 strings has been tightened so
2980 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2981 containing "overlong sequences".
2982
2983 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2984 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2985 should have been "&", but it just so happened that all the cases this let
2986 through by mistake were picked up later in the function.
2987
2988 6. I had used a variable called "isblank" - this is a C99 function, causing
2989 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2990
2991 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2992 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2993 so the version is shown; (c) stop "make test" echoing "./RunTest".
2994
2995 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2996
2997 9. The prototype for memmove() for systems that don't have it was using
2998 size_t, but the inclusion of the header that defines size_t was later. I've
2999 moved the #includes for the C headers earlier to avoid this.
3000
3001 10. Added some adjustments to the code to make it easier to compiler on certain
3002 special systems:
3003
3004 (a) Some "const" qualifiers were missing.
3005 (b) Added the macro EXPORT before all exported functions; by default this
3006 is defined to be empty.
3007 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3008 that it reads its output file name as an argument instead of writing
3009 to the standard output and assuming this can be redirected.
3010
3011 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3012 class containing characters with values greater than 255, PCRE compilation
3013 went into a loop.
3014
3015 12. A recursive reference to a subpattern that was within another subpattern
3016 that had a minimum quantifier of zero caused PCRE to crash. For example,
3017 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3018 recursion. If the recursively-called subpattern itself had a zero repeat,
3019 that was OK.
3020
3021 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3022 buffer into which it was copied (for escape processing) was still set at
3023 1024, so long lines caused crashes.
3024
3025 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3026 "internal error: code overflow...". This applied to any character class
3027 that was followed by a possessive quantifier.
3028
3029 15. Modified the Makefile to add libpcre.la as a prerequisite for
3030 libpcreposix.la because I was told this is needed for a parallel build to
3031 work.
3032
3033 16. If a pattern that contained .* following optional items at the start was
3034 studied, the wrong optimizing data was generated, leading to matching
3035 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3036 matching string must start with a or b or c. The correct conclusion for
3037 this pattern is that a match can start with any character.
3038
3039
3040 Version 4.4 13-Aug-03
3041 ---------------------
3042
3043 1. In UTF-8 mode, a character class containing characters with values between
3044 127 and 255 was not handled correctly if the compiled pattern was studied.
3045 In fixing this, I have also improved the studying algorithm for such
3046 classes (slightly).
3047
3048 2. Three internal functions had redundant arguments passed to them. Removal
3049 might give a very teeny performance improvement.
3050
3051 3. Documentation bug: the value of the capture_top field in a callout is *one
3052 more than* the number of the hightest numbered captured substring.
3053
3054 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3055 in incorrectly linking with a previously installed version. They now link
3056 explicitly with libpcre.la.
3057
3058 5. configure.in no longer needs to recognize Cygwin specially.
3059
3060 6. A problem in pcre.in for Windows platforms is fixed.
3061
3062 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3063 pcretest, it used to include the size of the study block as part of its
3064 output. Unfortunately, the structure contains a field that has a different
3065 size on different hardware architectures. This meant that the tests that
3066 showed this size failed. As the block is currently always of a fixed size,
3067 this information isn't actually particularly useful in pcretest output, so
3068 I have just removed it.
3069
3070 8. Three pre-processor statements accidentally did not start in column 1.
3071 Sadly, there are *still* compilers around that complain, even though
3072 standard C has not required this for well over a decade. Sigh.
3073
3074 9. In pcretest, the code for checking callouts passed small integers in the
3075 callout_data field, which is a void * field. However, some picky compilers
3076 complained about the casts involved for this on 64-bit systems. Now
3077 pcretest passes the address of the small integer instead, which should get
3078 rid of the warnings.
3079
3080 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3081 both compile and run time, and gives an error if an invalid UTF-8 sequence
3082 is found. There is a option for disabling this check in cases where the
3083 string is known to be correct and/or the maximum performance is wanted.
3084
3085 11. In response to a bug report, I changed one line in Makefile.in from
3086
3087 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3088 to
3089 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3090
3091 to look similar to other lines, but I have no way of telling whether this
3092 is the right thing to do, as I do not use Windows. No doubt I'll get told
3093 if it's wrong...
3094
3095
3096 Version 4.3 21-May-03
3097 ---------------------
3098
3099 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3100 Makefile.
3101
3102 2. Some refactoring to improve the quality of the code:
3103
3104 (i) The utf8_table... variables are now declared "const".
3105
3106 (ii) The code for \cx, which used the "case flipping" table to upper case
3107 lower case letters, now just substracts 32. This is ASCII-specific,
3108 but the whole concept of \cx is ASCII-specific, so it seems
3109 reasonable.
3110
3111 (iii) PCRE was using its character types table to recognize decimal and
3112 hexadecimal digits in the pattern. This is silly, because it handles
3113 only 0-9, a-f, and A-F, but the character types table is locale-
3114 specific, which means strange things might happen. A private
3115 table is now used for this - though it costs 256 bytes, a table is
3116 much faster than multiple explicit tests. Of course, the standard
3117 character types table is still used for matching digits in subject
3118 strings against \d.
3119
3120 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3121 ending in _t are). So I've renamed it as ESC_tee.
3122
3123 3. The first argument for regexec() in the POSIX wrapper should have been
3124 defined as "const".
3125
3126 4. Changed pcretest to use malloc() for its buffers so that they can be
3127 Electric Fenced for debugging.
3128
3129 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3130 to read one or more bytes before the start of the subject string. Often this
3131 had no effect on PCRE's behaviour, but in some circumstances it could
3132 provoke a segmentation fault.
3133
3134 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3135 to try to read one or more bytes before the start of the subject string.
3136
3137 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3138 UTF-8 support could misbehave in various ways if the subject string
3139 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3140 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3141 back over UTF-8 characters.)
3142
3143
3144 Version 4.2 14-Apr-03
3145 ---------------------
3146
3147 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3148
3149 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3150 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3151 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3152 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3153 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3154 and BUILD_EXEEXT
3155 Note: automatic setting of the BUILD variables is not yet working
3156 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3157 compile-time but not at link-time
3158 [LINK]: use for linking executables only
3159 make different versions for Windows and non-Windows
3160 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3161 libraries
3162 [LINK_FOR_BUILD]: new variable
3163 [OBJEXT]: use throughout
3164 [EXEEXT]: use throughout
3165 <winshared>: new target
3166 <wininstall>: new target
3167 <dftables.o>: use native compiler
3168 <dftables>: use native linker
3169 <install>: handle Windows platform correctly
3170 <clean>: ditto
3171 <check>: ditto
3172 copy DLL to top builddir before testing
3173
3174 As part of these changes, -no-undefined was removed again. This was reported
3175 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3176 in any case.
3177
3178 3. Some tidies to get rid of compiler warnings:
3179
3180 . In the match_data structure, match_limit was an unsigned long int, whereas
3181 match_call_count was an int. I've made them both unsigned long ints.
3182
3183 . In pcretest the fact that a const uschar * doesn't automatically cast to
3184 a void * provoked a warning.
3185
3186 . Turning on some more compiler warnings threw up some "shadow" variables
3187 and a few more missing casts.
3188
3189 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3190 option, a class that contained a single character with a value between 128
3191 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3192
3193 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3194 option, a class that contained several characters, but with at least one
3195 whose value was between 128 and 255 caused PCRE to crash.
3196
3197
3198 Version 4.1 12-Mar-03
3199 ---------------------
3200
3201 1. Compiling with gcc -pedantic found a couple of places where casts were
3202 needed, and a string in dftables.c that was longer than standard compilers are
3203 required to support.
3204
3205 2. Compiling with Sun's compiler found a few more places where the code could
3206 be tidied up in order to avoid warnings.
3207
3208 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3209 first of these names is deprecated in the latest Autoconf in favour of the name
3210 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3211 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3212 analogy I have changed it to CFLAGS_FOR_BUILD.
3213
3214 4. Added -no-undefined to the linking command in the Makefile, because this is
3215 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3216 linking step for the pcreposix library.
3217
3218 5. PCRE was failing to diagnose the case of two named groups with the same
3219 name.
3220
3221 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3222 literal character that is needed in the subject for a match, and scans along to
3223 ensure that it is present before embarking on the full matching process. This
3224 saves time in cases of nested unlimited repeats that are never going to match.
3225 Problem: the scan can take a lot of time if the subject is very long (e.g.
3226 megabytes), thus penalizing straightforward matches. It is now done only if the
3227 amount of subject to be scanned is less than 1000 bytes.
3228
3229 7. A lesser problem with the same optimization is that it was recording the
3230 first character of an anchored pattern as "needed", thus provoking a search
3231 right along the subject, even when the first match of the pattern was going to
3232 fail. The "needed" character is now not set for anchored patterns, unless it
3233 follows something in the pattern that is of non-fixed length. Thus, it still
3234 fulfils its original purpose of finding quick non-matches in cases of nested
3235 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3236
3237
3238 Version 4.0 17-Feb-03
3239 ---------------------
3240
3241 1. If a comment in an extended regex that started immediately after a meta-item
3242 extended to the end of string, PCRE compiled incorrect data. This could lead to
3243 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3244
3245 2. Moved to autoconf 2.53 and libtool 1.4.2.
3246
3247 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3248 the special perltest8 script is no longer needed - all the tests can be run
3249 from a single perltest script.
3250
3251 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3252 by \s. It has now been removed in PCRE. This means it isn't recognized as
3253 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3254 class [:space:] *does* include VT, thereby creating a mess.
3255
3256 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3257 space and tab.
3258
3259 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3260 its new features into the main test script, reducing the number of scripts.
3261
3262 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3263 were backward compatible, and made the (?i) apply to the whole pattern, as if
3264 /i were given. Now it behaves more logically, and applies the option setting
3265 only to what follows. PCRE has been changed to follow suit. However, if it
3266 finds options settings right at the start of the pattern, it extracts them into
3267 the global options, as before. Thus, they show up in the info data.
3268
3269 8. Added support for the \Q...\E escape sequence. Characters in between are
3270 treated as literals. This is slightly different from Perl in that $ and @ are
3271 also handled as literals inside the quotes. In Perl, they will cause variable
3272 interpolation. Note the following examples:
3273
3274 Pattern PCRE matches Perl matches
3275
3276 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3277 \Qabc\$xyz\E abc\$xyz abc\$xyz
3278 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3279
3280 For compatibility with Perl, \Q...\E sequences are recognized inside character
3281 classes as well as outside them.
3282
3283 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3284 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3285 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3286 signed/unsigned warnings.
3287
3288 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3289 option for pcretest, so I've replaced it by a simple function that does just
3290 that job.
3291
3292 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3293 "pcregrep -".
3294
3295 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3296 Java package. This provides some syntactic sugar for simple cases of what my
3297 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3298 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3299 item, you can use this simplified notation. Note that only makes sense with
3300 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3301 greediness, whatever the setting of the PCRE_UNGREEDY option.
3302
3303 13. A change of greediness default within a pattern was not taking effect at
3304 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3305 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3306 was abstracted outside.
3307
3308 14. PCRE now supports the \G assertion. It is true when the current matching
3309 position is at the start point of the match. This differs from \A when the
3310 starting offset is non-zero. Used with the /g option of pcretest (or similar
3311 code), it works in the same way as it does for Perl's /g option. If all
3312 alternatives of a regex begin with \G, the expression is anchored to the start
3313 match position, and the "anchored" flag is set in the compiled expression.
3314
3315 15. Some bugs concerning the handling of certain option changes within patterns
3316 have been fixed. These applied to options other than (?ims). For example,
3317 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3318 been the other way round. Some of this was related to change 7 above.
3319
3320 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3321 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3322 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3323 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3324
3325 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3326 mode. Unlike ".", it always matches newline, whatever the setting of
3327 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3328 assertions. Perl allows it, but it doesn't (in general) work because it can't
3329 calculate the length of the lookbehind. At least, that's the case for Perl
3330 5.8.0 - I've been told they are going to document that it doesn't work in
3331 future.
3332
3333 18. Added an error diagnosis for escapes that PCRE does not support: these are
3334 \L, \l, \N, \P, \p, \U, \u, and \X.
3335
3336 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3337 reading past the end of the pattern in cases such as /[abcd/.
3338
3339 20. PCRE was getting more memory than necessary for patterns with classes that
3340 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3341
3342 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3343 compiling PCRE for use with Virtual Pascal.
3344
3345 22. Small fix to the Makefile to make it work properly if the build is done
3346 outside the source tree.
3347
3348 23. Added a new extension: a condition to go with recursion. If a conditional
3349 subpattern starts with (?(R) the "true" branch is used if recursion has
3350 happened, whereas the "false" branch is used only at the top level.
3351
3352 24. When there was a very long string of literal characters (over 255 bytes
3353 without UTF support, over 250 bytes with UTF support), the computation of how
3354 much memory was required could be incorrect, leading to segfaults or other
3355 strange effects.
3356
3357 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3358 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3359 there was a subsequent back reference to those brackets. This meant that, for
3360 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3361 possible to check for precisely this case. All we can do is abandon the
3362 optimization if .* occurs inside capturing brackets when there are any back
3363 references whatsoever. (See below for a better fix that came later.)
3364
3365 26. The handling of the optimization for finding the first character of a
3366 non-anchored pattern, and for finding a character that is required later in the
3367 match were failing in some cases. This didn't break the matching; it just
3368 failed to optimize when it could. The way this is done has been re-implemented.
3369
3370 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3371
3372 28. Added a new feature that provides some of the functionality that Perl
3373 provides with (?{...}). The facility is termed a "callout". The way it is done
3374 in PCRE is for the caller to provide an optional function, by setting
3375 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3376 global variable. By default it is unset, which disables all calling out. To get
3377 the function called, the regex must include (?C) at appropriate points. This
3378 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3379 This provides a means of identifying different callout points. When PCRE
3380 reaches such a point in the regex, if pcre_callout has been set, the external
3381 function is called. It is provided with data in a structure called
3382 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3383 matching continues; if it returns a non-zero value, the match at the current
3384 point fails. However, backtracking will occur if possible. [This was changed
3385 later and other features added - see item 49 below.]
3386
3387 29. pcretest is upgraded to test the callout functionality. It provides a
3388 callout function that displays information. By default, it shows the start of
3389 the match and the current position in the text. There are some new data escapes
3390 to vary what happens:
3391
3392 \C+ in addition, show current contents of captured substrings
3393 \C- do not supply a callout function
3394 \C!n return 1 when callout number n is reached
3395 \C!n!m return 1 when callout number n is reached for the mth time
3396
3397 30. If pcregrep was called with the -l option and just a single file name, it
3398 output "<stdin>" if a match was found, instead of the file name.
3399
3400 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3401 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3402 pcre_exec(). This saves a malloc/free per call. The default value of
3403 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3404 when configuring.
3405
3406 32. The default maximum size of a compiled pattern is 64K. There have been a
3407 few cases of people hitting this limit. The code now uses macros to handle the
3408 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3409 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3410 configuring. Tests 2 and 5 work only with 2-byte links because they output
3411 debugging information about compiled patterns.
3412
3413 33. Internal code re-arrangements:
3414
3415 (a) Moved the debugging function for printing out a compiled regex into
3416 its own source file (printint.c) and used #include to pull it into
3417 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3418 separate copies.
3419
3420 (b) Defined the list of op-code names for debugging as a macro in
3421 internal.h so that it is next to the definition of the opcodes.
3422
3423 (c) Defined a table of op-code lengths for simpler skipping along compiled
3424 code. This is again a macro in internal.h so that it is next to the
3425 definition of the opcodes.
3426
3427 34. Added support for recursive calls to individual subpatterns, along the
3428 lines of Robin Houston's patch (but implemented somewhat differently).
3429
3430 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3431 allow it to read and process whole directories in Win32. This code was
3432 contributed by Lionel Fourquaux; it has not been tested by me.
3433
3434 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3435 used to name a group. Names consist of alphanumerics and underscores, and must
3436 be unique. Back references use the syntax (?P=name) and recursive calls use
3437 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3438 numbers. The function pcre_fullinfo() can be used after compilation to extract
3439 a name/number map. There are three relevant calls:
3440
3441 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3442 PCRE_INFO_NAMECOUNT yields the number of entries
3443 PCRE_INFO_NAMETABLE yields a pointer to the map.
3444
3445 The map is a vector of fixed-size entries. The size of each entry depends on
3446 the length of the longest name used. The first two bytes of each entry are the
3447 group number, most significant byte first. There follows the corresponding
3448 name, zero terminated. The names are in alphabetical order.
3449
3450 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3451 case instead of 255. Making it the same both with and without UTF-8 support
3452 means that the same test output works with both.
3453
3454 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3455 calling malloc() with a zero argument.
3456
3457 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3458 optimization. I've improved things by keeping a bitmap of backreferences with
3459 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3460 fact referenced, the optimization can be applied. It is unlikely that a
3461 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3462 the match to follow \n) will appear inside brackets with a number greater than
3463 31, but if it does, any back reference > 31 suppresses the optimization.
3464
3465 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3466 of disabling numbered capturing parentheses. Any opening parenthesis that is
3467 not followed by ? behaves as if it were followed by ?: but named parentheses
3468 can still be used for capturing (and they will acquire numbers in the usual
3469 way).
3470
3471 41. Redesigned the return codes from the match() function into yes/no/error so
3472 that errors can be passed back from deep inside the nested calls. A malloc
3473 failure while inside a recursive subpattern call now causes the
3474 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3475
3476 42. It is now possible to set a limit on the number of times the match()
3477 function is called in a call to pcre_exec(). This facility makes it possible to
3478 limit the amount of recursion and backtracking, though not in a directly
3479 obvious way, because the match() function is used in a number of different
3480 circumstances. The count starts from zero for each position in the subject
3481 string (for non-anchored patterns). The default limit is, for compatibility, a
3482 large number, namely 10 000 000. You can change this in two ways:
3483
3484 (a) When configuring PCRE before making, you can use --with-match-limit=n
3485 to set a default value for the compiled library.
3486
3487 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3488 a different value is set. See 45 below.
3489
3490 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3491
3492 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3493 of things that can be changed at compile time. The first argument specifies
3494 what is wanted and the second points to where the information is to be placed.
3495 The current list of available information is:
3496
3497 PCRE_CONFIG_UTF8
3498
3499 The output is an integer that is set to one if UTF-8 support is available;
3500 otherwise it is set to zero.
3501
3502 PCRE_CONFIG_NEWLINE
3503
3504 The output is an integer that it set to the value of the code that is used for
3505 newline. It is either LF (10) or CR (13).
3506
3507 PCRE_CONFIG_LINK_SIZE
3508
3509 The output is an integer that contains the number of bytes used for internal
3510 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3511
3512 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3513
3514 The output is an integer that contains the threshold above which the POSIX
3515 interface uses malloc() for output vectors. See item 31 above.
3516
3517 PCRE_CONFIG_MATCH_LIMIT
3518
3519 The output is an unsigned integer that contains the default limit of the number
3520 of match() calls in a pcre_exec() execution. See 42 above.
3521
3522 44. pcretest has been upgraded by the addition of the -C option. This causes it
3523 to extract all the available output from the new pcre_config() function, and to
3524 output it. The program then exits immediately.
3525
3526 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3527 order to support additional features. One way would have been to define
3528 pcre_exec2() (for example) with extra arguments, but this would not have been
3529 extensible, and would also have required all calls to the original function to
3530 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3531 is used for passing in "extra" data from pcre_study().
3532
3533 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3534 contains the following fields:
3535
3536 flags a bitmap indicating which of the following fields are set
3537 study_data opaque data from pcre_study()
3538 match_limit a way of specifying a limit on match() calls for a specific
3539 call to pcre_exec()
3540 callout_data data for callouts (see 49 below)
3541
3542 The flag bits are also defined in pcre.h, and are
3543
3544 PCRE_EXTRA_STUDY_DATA
3545 PCRE_EXTRA_MATCH_LIMIT
3546 PCRE_EXTRA_CALLOUT_DATA
3547
3548 The pcre_study() function now returns one of these new pcre_extra blocks, with
3549 the actual study data pointed to by the study_data field, and the
3550 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3551 before. That is, this change is entirely upwards-compatible and requires no
3552 change to existing code.
3553
3554 If you want to pass in additional data to pcre_exec(), you can either place it
3555 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3556 block.
3557
3558 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3559 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3560 times with different match limits, until it finds the minimum value needed for
3561 pcre_exec() to complete. The value is then output. This can be instructive; for
3562 most simple matches the number is quite small, but for pathological cases it
3563 gets very large very quickly.
3564
3565 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3566 returns the size of the data block pointed to by the study_data field in a
3567 pcre_extra block, that is, the value that was passed as the argument to
3568 pcre_malloc() when PCRE was getting memory in which to place the information
3569 created by pcre_study(). The fourth argument should point to a size_t variable.
3570 pcretest has been extended so that this information is shown after a successful
3571 pcre_study() call when information about the compiled regex is being displayed.
3572
3573 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3574 because what follows is always an absolute path. (Later: it turns out that this
3575 is more than cosmetic for MinGW, because it doesn't like empty path
3576 components.)
3577
3578 49. Some changes have been made to the callout feature (see 28 above):
3579
3580 (i) A callout function now has three choices for what it returns:
3581
3582 0 => success, carry on matching
3583 > 0 => failure at this point, but backtrack if possible
3584 < 0 => serious error, return this value from pcre_exec()
3585
3586 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3587 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3588 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3589 use by callout functions. It will never be used by PCRE itself.
3590
3591 (ii) The pcre_extra structure (see 45 above) has a void * field called
3592 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3593 pcre_callout_block structure has a field of the same name. The contents of
3594 the field passed in the pcre_extra structure are passed to the callout
3595 function in the corresponding field in the callout block. This makes it
3596 easier to use the same callout-containing regex from multiple threads. For
3597 testing, the pcretest program has a new data escape
3598
3599 \C*n pass the number n (may be negative) as callout_data
3600
3601 If the callout function in pcretest receives a non-zero value as
3602 callout_data, it returns that value.
3603
3604 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3605 there were some redundant $(CFLAGS) in commands that are now specified as
3606 $(LINK), which already includes $(CFLAGS).
3607
3608 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3609 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3610 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3611 one-byte characters throughout. Note that case-insensitive matching applies
3612 only to characters whose values are less than 256. PCRE doesn't support the
3613 notion of cases for higher-valued characters.
3614
3615 (i) A character class whose characters are all within 0-255 is handled as
3616 a bit map, and the map is inverted for negative classes. Previously, a
3617 character > 255 always failed to match such a class; however it should
3618 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3619
3620 (ii) A negated character class with a single character < 255 is coded as
3621 "not this character" (OP_NOT). This wasn't working properly when the test
3622 character was multibyte, either singly or repeated.
3623
3624 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3625 mode, for example: \x{100}{2,3}.
3626
3627 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3628 singly or repeated) now correctly test multibyte characters. However,
3629 PCRE doesn't recognize any characters with values greater than 255 as
3630 digits, spaces, or word characters. Such characters always match \D, \S,
3631 and \W, and never match \d, \s, or \w.
3632
3633 (v) Classes may now contain characters and character ranges with values
3634 greater than 255. For example: [ab\x{100}-\x{400}].
3635
3636 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3637 PCRE in UTF-8 mode.
3638
3639 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3640 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3641 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3642 value.)
3643
3644 53. The single man page has become too large. I have therefore split it up into
3645 a number of separate man pages. These also give rise to individual HTML pages;
3646 these are now put in a separate directory, and there is an index.html page that
3647 lists them all. Some hyperlinking between the pages has been installed.
3648
3649 54. Added convenience functions for handling named capturing parentheses.
3650
3651 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3652 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3653 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3654 are faulted.
3655
3656 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3657 calling configure. These values are used when compiling the dftables.c program
3658 which is run to generate the source of the default character tables. They
3659 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3660 you will need to set these values.
3661
3662 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3663
3664
3665 Version 3.9 02-Jan-02
3666 ---------------------
3667
3668 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3669
3670 2. If --disable-static was given, the building process failed when trying to
3671 build pcretest and pcregrep. (For some reason it was using libtool to compile
3672 them, which is not right, as they aren't part of the library.)
3673
3674
3675 Version 3.8 18-Dec-01
3676 ---------------------
3677
3678 1. The experimental UTF-8 code was completely screwed up. It was packing the
3679 bytes in the wrong order. How dumb can you get?
3680
3681
3682 Version 3.7 29-Oct-01
3683 ---------------------
3684
3685 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3686 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3687 this didn't happen under Solaris 8, where I normally test things.
3688
3689 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3690 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3691 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3692 this fix an hour or so after the initial 3.7 release.)
3693
3694
3695 Version 3.6 23-Oct-01
3696 ---------------------
3697
3698 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3699 offsets passed as NULL with zero offset count.
3700
3701 2. The config.guess and config.sub files had not been updated when I moved to
3702 the latest autoconf.
3703
3704
3705 Version 3.5 15-Aug-01
3706 ---------------------
3707
3708 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3709 had been forgotten.
3710
3711 2. By using declared but undefined structures, we can avoid using "void"
3712 definitions in pcre.h while keeping the internal definitions of the structures
3713 private.
3714
3715 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3716 user point of view, this means that both static and shared libraries are built
3717 by default, but this can be individually controlled. More of the work of
3718 handling this static/shared cases is now inside libtool instead of PCRE's make
3719 file.
3720
3721 4. The pcretest utility is now installed along with pcregrep because it is
3722 useful for users (to test regexs) and by doing this, it automatically gets
3723 relinked by libtool. The documentation has been turned into a man page, so
3724 there are now .1, .txt, and .html versions in /doc.
3725
3726 5. Upgrades to pcregrep:
3727 (i) Added long-form option names like gnu grep.
3728 (ii) Added --help to list all options with an explanatory phrase.
3729 (iii) Added -r, --recursive to recurse into sub-directories.
3730 (iv) Added -f, --file to read patterns from a file.
3731
3732 6. pcre_exec() was referring to its "code" argument before testing that
3733 argument for NULL (and giving an error if it was NULL).
3734
3735 7. Upgraded Makefile.in to allow for compiling in a different directory from
3736 the source directory.
3737
3738 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3739 options bits, the pointer it was passed was to an int instead of to an unsigned
3740 long int. This mattered only on 64-bit systems.
3741
3742 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3743 generated) instead of pcre.in, which it its source. Also made the same change
3744 in several of the .c files.
3745
3746 10. A new release of gcc defines printf() as a macro, which broke pcretest
3747 because it had an ifdef in the middle of a string argument for printf(). Fixed
3748 by using separate calls to printf().
3749
3750 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3751 script, to force use of CR or LF instead of \n in the source. On non-Unix
3752 systems, the value can be set in config.h.
3753
3754 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3755 absolute limit. Changed the text of the error message to make this clear, and
3756 likewise updated the man page.
3757
3758 13. The limit of 99 on the number of capturing subpatterns has been removed.
3759 The new limit is 65535, which I hope will not be a "real" limit.
3760
3761
3762 Version 3.4 22-Aug-00
3763 ---------------------
3764
3765 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3766
3767 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3768
3769
3770 Version 3.3 01-Aug-00
3771 ---------------------
3772
3773 1. If an octal character was given, but the value was greater than \377, it
3774 was not getting masked to the least significant bits, as documented. This could
3775 lead to crashes in some systems.
3776
3777 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3778 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3779
3780 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3781 These just pass their arguments on to (pcre_free)(), but they are provided
3782 because some uses of PCRE bind it to non-C systems that can call its functions,
3783 but cannot call free() or pcre_free() directly.
3784
3785 4. Add "make test" as a synonym for "make check". Corrected some comments in
3786 the Makefile.
3787
3788 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3789 Makefile.
3790
3791 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3792 command called pgrep for grepping around the active processes.
3793
3794 7. Added the beginnings of support for UTF-8 character strings.
3795
3796 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3797 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3798 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3799 out for the ar command.)
3800
3801
3802 Version 3.2 12-May-00
3803 ---------------------
3804
3805 This is purely a bug fixing release.
3806
3807 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3808 of ZA. This was just one example of several cases that could provoke this bug,
3809 which was introduced by change 9 of version 2.00. The code for breaking
3810 infinite loops after an iteration that matches an empty string was't working
3811 correctly.
3812
3813 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3814 when matched against abbab (for example). After matching an empty string, it
3815 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3816 caused it to match further down the string than it should.
3817
3818 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3819 was there because it doesn't seem to be needed, and it causes trouble on some
3820 systems, as it is not a Standard C header. It has been removed.
3821
3822 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3823 were reported on the Macintosh. The changes were from
3824
3825 while ((c = *(++ptr)) != 0 && c != '\n');
3826 to
3827 while ((c = *(++ptr)) != 0 && c != '\n') ;
3828
3829 Totally extraordinary, but if that's what it takes...
3830
3831 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3832 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3833 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3834 assumes the way PCRE uses memmove() (always moving upwards).
3835
3836 6. PCRE is being used in one environment where strchr() is not available. There
3837 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3838 faster code anyway.
3839
3840
3841 Version 3.1 09-Feb-00
3842 ---------------------
3843
3844 The only change in this release is the fixing of some bugs in Makefile.in for
3845 the "install" target:
3846
3847 (1) It was failing to install pcreposix.h.
3848
3849 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3850
3851
3852 Version 3.0 01-Feb-00
3853 ---------------------
3854
3855 1. Add support for the /+ modifier to perltest (to output $` like it does in
3856 pcretest).
3857
3858 2. Add support for the /g modifier to perltest.
3859
3860 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3861 matches null strings.
3862
3863 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3864 pattern. Perl treats empty patterns specially - it reuses the most recent
3865 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3866 effect.
3867
3868 5. The POSIX interface was broken in that it was just handing over the POSIX
3869 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3870 required a bigger vector, with some working space on the end. This means that
3871 the POSIX wrapper now has to get and free some memory, and copy the results.
3872
3873 6. Added some simple autoconf support, placing the test data and the
3874 documentation in separate directories, re-organizing some of the
3875 information files, and making it build pcre-config (a GNU standard). Also added
3876 libtool support for building PCRE as a shared library, which is now the
3877 default.
3878
3879 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3880 09 are not valid octal constants. Single digits will be used for minor values
3881 less than 10.
3882
3883 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3884 existing programs that set these in the POSIX interface can use PCRE without
3885 modification.
3886
3887 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3888 return all that pcre_info() returns, plus additional data. The pcre_info()
3889 function is retained for compatibility, but is considered to be obsolete.
3890
3891 10. Added experimental recursion feature (?R) to handle one common case that
3892 Perl 5.6 will be able to do with (?p{...}).
3893
3894 11. Added support for POSIX character classes like [:alpha:], which Perl is
3895 adopting.
3896
3897
3898 Version 2.08 31-Aug-99
3899 ----------------------
3900
3901 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3902 trying to match at the startoffset position, but instead was moving forward to
3903 the next newline as if a previous match had failed.
3904
3905 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3906 and could get into a loop if a null string was matched other than at the start
3907 of the subject.
3908
3909 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3910 be distinguished at compile time, and for completeness also added PCRE_DATE.
3911
3912 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3913 in GnuWin32 environments.
3914
3915
3916 Version 2.07 29-Jul-99
3917 ----------------------
3918
3919 1. The documentation is now supplied in plain text form and HTML as well as in
3920 the form of man page sources.
3921
3922 2. C++ compilers don't like assigning (void *) values to other pointer types.
3923 In particular this affects malloc(). Although there is no problem in Standard
3924 C, I've put in casts to keep C++ compilers happy.
3925
3926 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3927 should be (const char *).
3928
3929 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3930 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3931 However, I haven't made this a standard facility. The documentation doesn't
3932 mention it, and the Makefile doesn't support it.
3933
3934 5. The Makefile now contains an "install" target, with editable destinations at
3935 the top of the file. The pcretest program is not installed.
3936
3937 6. pgrep -V now gives the PCRE version number and date.
3938
3939 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3940 causing the entire string to be ignored, instead of just the last character.
3941
3942 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3943 non-matching string, it can take a very, very long time, even for strings of
3944 quite modest length, because of the nested recursion. PCRE now does better in
3945 some of these cases. It does this by remembering the last required literal
3946 character in the pattern, and pre-searching the subject to ensure it is present
3947 before running the real match. In other words, it applies a heuristic to detect
3948 some types of certain failure quickly, and in the above example, if presented
3949 with a string that has no trailing " it gives "no match" very quickly.
3950
3951 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3952 other alternatives are tried instead.
3953
3954
3955 Version 2.06 09-Jun-99
3956 ----------------------
3957
3958 1. Change pcretest's output for amount of store used to show just the code
3959 space, because the remainder (the data block) varies in size between 32-bit and
3960 64-bit systems.
3961
3962 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3963 start matching at. This allows lookbehinds to work when searching for multiple
3964 occurrences in a string.
3965
3966 3. Added additional options to pcretest for testing multiple occurrences:
3967
3968 /+ outputs the rest of the string that follows a match
3969 /g loops for multiple occurrences, using the new startoffset argument
3970 /G loops for multiple occurrences by passing an incremented pointer
3971
3972 4. PCRE wasn't doing the "first character" optimization for patterns starting
3973 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3974 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3975 the letter 'x'. On long subject strings, this gives a significant speed-up.
3976
3977
3978 Version 2.05 21-Apr-99
3979 ----------------------
3980
3981 1. Changed the type of magic_number from int to long int so that it works
3982 properly on 16-bit systems.
3983
3984 2. Fixed a bug which caused patterns starting with .* not to work correctly
3985 when the subject string contained newline characters. PCRE was assuming
3986 anchoring for such patterns in all cases, which is not correct because .* will
3987 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3988 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3989 must be retried after every newline in the subject.
3990
3991
3992 Version 2.04 18-Feb-99
3993 ----------------------
3994
3995 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3996 computation of the store needed to hold the pattern was incorrect (too large).
3997 If such patterns were nested a few deep, this could multiply and become a real
3998 problem.
3999
4000 2. Added /M option to pcretest to show the memory requirement of a specific
4001 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4002
4003 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4004 compiled in such a way that the backtracking after subsequent failure was
4005 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4006 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4007
4008
4009 Version 2.03 02-Feb-99
4010 ----------------------
4011
4012 1. Fixed typo and small mistake in man page.
4013
4014 2. Added 4th condition (GPL supersedes if conflict) and created separate
4015 LICENCE file containing the conditions.
4016
4017 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4018 Perl, that is the internal \ allows the delimiter to be included in the
4019 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4020 the final delimiter, add \ to the end of the pattern (to test the error).
4021
4022 4. Added the convenience functions for extracting substrings after a successful
4023 match. Updated pcretest to make it able to test these functions.
4024
4025
4026 Version 2.02 14-Jan-99
4027 ----------------------
4028
4029 1. Initialized the working variables associated with each extraction so that
4030 their saving and restoring doesn't refer to uninitialized store.
4031
4032 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4033 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4034 fix the problem.
4035
4036 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4037 calls, and wasn't printing the correct value for compiling calls. Increased the
4038 default value of LOOPREPEAT, and the number of significant figures in the
4039 times.
4040
4041 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4042
4043 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4044 a building problem on Windows NT with a FAT file system.
4045
4046
4047 Version 2.01 21-Oct-98
4048 ----------------------
4049
4050 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4051 to character tables built by pcre_maketables() in the current locale. If NULL
4052 is passed, the default tables are used.
4053
4054
4055 Version 2.00 24-Sep-98
4056 ----------------------
4057
4058 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4059 it any more.
4060
4061 2. Allow quantification of (?>) groups, and make it work correctly.
4062
4063 3. The first character computation wasn't working for (?>) groups.
4064
4065 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4066 end of the subject) and add 5.005's \z, which really does match only at the
4067 very end of the subject.
4068
4069 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4070
4071 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4072 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4073 localized options. All options to pcre_study() were also removed.
4074
4075 7. Add other new features from 5.005:
4076
4077 $(?<= positive lookbehind
4078 $(?<! negative lookbehind
4079 (?imsx-imsx) added the unsetting capability
4080 such a setting is global if at outer level; local otherwise
4081 (?imsx-imsx:) non-capturing groups with option setting
4082 (?(cond)re|re) conditional pattern matching
4083
4084 A backreference to itself in a repeated group matches the previous
4085 captured string.
4086
4087 8. General tidying up of studying (both automatic and via "study")
4088 consequential on the addition of new assertions.
4089
4090 9. As in 5.005, unlimited repeated groups that could match an empty substring
4091 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4092 runtime if any iteration does actually match an empty substring.
4093
4094 10. Include the RunTest script in the distribution.
4095
4096 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4097 discrepancies, some of which were old and were also with respect to 5.004. They
4098 have now been fixed.
4099
4100
4101 Version 1.09 28-Apr-98
4102 ----------------------
4103
4104 1. A negated single character class followed by a quantifier with a minimum
4105 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4106 program crashes, or just wrong answers. This did not apply to negated classes
4107 containing more than one character, or to minima other than one.
4108
4109
4110 Version 1.08 27-Mar-98
4111 ----------------------
4112
4113 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4114
4115 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4116 latter must appear before anything that relies on it in the pattern.
4117
4118
4119 Version 1.07 16-Feb-98
4120 ----------------------
4121
4122 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4123 repeat of a potentially empty string).
4124
4125
4126 Version 1.06 23-Jan-98
4127 ----------------------
4128
4129 1. Added Markus Oberhumer's little patches for C++.
4130
4131 2. Literal strings longer than 255 characters were broken.
4132
4133
4134 Version 1.05 23-Dec-97
4135 ----------------------
4136
4137 1. Negated character classes containing more than one character were failing if
4138 PCRE_CASELESS was set at run time.
4139
4140
4141 Version 1.04 19-Dec-97
4142 ----------------------
4143
4144 1. Corrected the man page, where some "const" qualifiers had been omitted.
4145
4146 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4147 input syntax.
4148
4149 3. Fixed memory leak which occurred when a regex with back references was
4150 matched with an offsets vector that wasn't big enough. The temporary memory
4151 that is used in this case wasn't being freed if the match failed.
4152
4153 4. Tidied pcretest to ensure it frees memory that it gets.
4154
4155 5. Temporary memory was being obtained in the case where the passed offsets
4156 vector was exactly big enough.
4157
4158 6. Corrected definition of offsetof() from change 5 below.
4159
4160 7. I had screwed up change 6 below and broken the rules for the use of
4161 setjmp(). Now fixed.
4162
4163
4164 Version 1.03 18-Dec-97
4165 ----------------------
4166
4167 1. A erroneous regex with a missing opening parenthesis was correctly
4168 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4169 on some systems.
4170
4171 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4172 it was reported that one broken compiler failed on the former because "code" is
4173 also an independent variable.
4174
4175 3. The erroneous regex a[]b caused an array overrun reference.
4176
4177 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4178 fail on data ending with that character. (It was going on too far, and checking
4179 the next character, typically a binary zero.) This was specific to the
4180 optimized code for single-character negative classes.
4181
4182 5. Added a contributed patch from the TIN world which does the following:
4183
4184 + Add an undef for memmove, in case the the system defines a macro for it.
4185
4186 + Add a definition of offsetof(), in case there isn't one. (I don't know
4187 the reason behind this - offsetof() is part of the ANSI standard - but
4188 it does no harm).
4189
4190 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4191 most of the places where whitespace preceded '#'. I have given up and
4192 allowed the remaining 2 cases to be at the margin.
4193
4194 + Rename some variables in pcre to eliminate shadowing. This seems very
4195 pedantic, but does no harm, of course.
4196
4197 6. Moved the call to setjmp() into its own function, to get rid of warnings
4198 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4199
4200 7. Constructs such as \d{8,} were compiling into the equivalent of
4201 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4202 outcome, but in this particular case used more store than had been allocated,
4203 which caused the bug to be discovered because it threw up an internal error.
4204
4205 8. The debugging code in both pcre and pcretest for outputting the compiled
4206 form of a regex was going wrong in the case of back references followed by
4207 curly-bracketed repeats.
4208
4209
4210 Version 1.02 12-Dec-97
4211 ----------------------
4212
4213 1. Typos in pcre.3 and comments in the source fixed.
4214
4215 2. Applied a contributed patch to get rid of places where it used to remove
4216 'const' from variables, and fixed some signed/unsigned and uninitialized
4217 variable warnings.
4218
4219 3. Added the "runtest" target to Makefile.
4220
4221 4. Set default compiler flag to -O2 rather than just -O.
4222
4223
4224 Version 1.01 19-Nov-97
4225 ----------------------
4226
4227 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4228 like /([ab]*)*/, that is, for classes with more than one character in them.
4229
4230 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4231 as /((?>a*))*/ (a PCRE_EXTRA facility).
4232
4233
4234 Version 1.00 18-Nov-97
4235 ----------------------
4236
4237 1. Added compile-time macros to support systems such as SunOS4 which don't have
4238 memmove() or strerror() but have other things that can be used instead.
4239
4240 2. Arranged that "make clean" removes the executables.
4241
4242
4243 Version 0.99 27-Oct-97
4244 ----------------------
4245
4246 1. Fixed bug in code for optimizing classes with only one character. It was
4247 initializing a 32-byte map regardless, which could cause it to run off the end
4248 of the memory it had got.
4249
4250 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4251
4252
4253 Version 0.98 22-Oct-97
4254 ----------------------
4255
4256 1. Fixed bug in code for handling temporary memory usage when there are more
4257 back references than supplied space in the ovector. This could cause segfaults.
4258
4259
4260 Version 0.97 21-Oct-97
4261 ----------------------
4262
4263 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4264
4265 2. Optimized negated single characters not to use a bit map.
4266
4267 3. Brought error texts together as macro definitions; clarified some of them;
4268 fixed one that was wrong - it said "range out of order" when it meant "invalid
4269 escape sequence".
4270
4271 4. Changed some char * arguments to const char *.
4272
4273 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4274
4275 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4276 pcretest.
4277
4278
4279 Version 0.96 16-Oct-97
4280 ----------------------
4281
4282 1. Added a simple "pgrep" utility to the distribution.
4283
4284 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4285 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4286 where "ddd" means "one or more decimal digits".
4287
4288 3. Fixed serious bug. If a pattern had a back reference, but the call to
4289 pcre_exec() didn't supply a large enough ovector to record the related
4290 identifying subpattern, the match always failed. PCRE now remembers the number
4291 of the largest back reference, and gets some temporary memory in which to save
4292 the offsets during matching if necessary, in order to ensure that
4293 backreferences always work.
4294
4295 4. Increased the compatibility with Perl in a number of ways:
4296
4297 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4298 to request this handling. The option can be set at compile or exec time.
4299
4300 (b) $ matches before a terminating newline by default; an option
4301 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4302 mode). The option can be set at compile or exec time.
4303
4304 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4305 the same as Perl's. If the decimal number it represents is less than 10
4306 or there aren't that many previous left capturing parentheses, an octal
4307 escape is read. Inside a character class, it's always an octal escape,
4308 even if it is a single digit.
4309
4310 (d) An escaped but undefined alphabetic character is taken as a literal,
4311 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4312 escapes.
4313
4314 (e) {0} is now permitted. (The previous item is removed from the compiled
4315 pattern).
4316
4317 5. Changed all the names of code files so that the basic parts are no longer
4318 than 10 characters, and abolished the teeny "globals.c" file.
4319
4320 6. Changed the handling of character classes; they are now done with a 32-byte
4321 bit map always.
4322
4323 7. Added the -d and /D options to pcretest to make it possible to look at the
4324 internals of compilation without having to recompile pcre.
4325
4326
4327 Version 0.95 23-Sep-97
4328 ----------------------
4329
4330 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4331 \x20 at the start of a run of normal characters. These were being treated as
4332 real characters, instead of the source characters being re-checked.
4333
4334
4335 Version 0.94 18-Sep-97
4336 ----------------------
4337
4338 1. The functions are now thread-safe, with the caveat that the global variables
4339 containing pointers to malloc() and free() or alternative functions are the
4340 same for all threads.
4341
4342 2. Get pcre_study() to generate a bitmap of initial characters for non-
4343 anchored patterns when this is possible, and use it if passed to pcre_exec().
4344
4345
4346 Version 0.93 15-Sep-97
4347 ----------------------
4348
4349 1. /(b)|(:+)/ was computing an incorrect first character.
4350
4351 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4352 but not actually doing anything yet.
4353
4354 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4355 as Perl does (e.g. [-az] or [az-]).
4356
4357 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4358 all possible positions.
4359
4360 5. Split up into different modules to avoid including unneeded functions in a
4361 compiled binary. However, compile and exec are still in one module. The "study"
4362 function is split off.
4363
4364 6. The character tables are now in a separate module whose source is generated
4365 by an auxiliary program - but can then be edited by hand if required. There are
4366 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4367 toupper() in the code.
4368
4369 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4370 make them global. Abolish the function for setting them, as the caller can now
4371 set them directly.
4372
4373
4374 Version 0.92 11-Sep-97
4375 ----------------------
4376
4377 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4378 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4379
4380 2. Caseless matching was not working in character classes if the characters in
4381 the pattern were in upper case.
4382
4383 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4384
4385 4. Make PCRE_ANCHORED public and accept as a compile option.
4386
4387 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4388 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4389 pass them.
4390
4391 6. Give an error if bad option bits passed at compile or run time.
4392
4393 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4394 pcretest to cause it to pass that flag.
4395
4396 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4397 options, and the first character, if set.
4398
4399 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4400
4401
4402 Version 0.91 10-Sep-97
4403 ----------------------
4404
4405 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4406 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4407
4408 2. PCRE was looping on encountering an indefinitely repeated back reference to
4409 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4410 Perl does - treats the match as successful.
4411
4412 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5