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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 687 - (show annotations)
Tue Sep 6 16:09:43 2011 UTC (3 years, 9 months ago) by ph10
File size: 213590 byte(s)
Error occurred while calculating annotation data.
Forgot to update ChangeLog
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. When this is
31 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
32 runtime --no-jit option is given.
33
34 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
35 ovector, the return from the function was zero, implying that there were
36 other matches that did not fit. The correct "exactly full" value is now
37 returned.
38
39 6. If a subpattern that was called recursively or as a subroutine contained
40 (*PRUNE) or any other control that caused it to give a non-standard return,
41 invalid errors such as "Error -26 (nested recursion at the same subject
42 position)" or even infinite loops could occur.
43
44
45 Version 8.13 16-Aug-2011
46 ------------------------
47
48 1. The Unicode data tables have been updated to Unicode 6.0.0.
49
50 2. Two minor typos in pcre_internal.h have been fixed.
51
52 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
53 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
54 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
55
56 4. There were a number of related bugs in the code for matching backrefences
57 caselessly in UTF-8 mode when codes for the characters concerned were
58 different numbers of bytes. For example, U+023A and U+2C65 are an upper
59 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
60 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
61 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
62 2-byte code at the end of the subject (it thought there wasn't enough data
63 left).
64
65 5. Comprehensive information about what went wrong is now returned by
66 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
67 as the output vector has at least 2 elements. The offset of the start of
68 the failing character and a reason code are placed in the vector.
69
70 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
71 now returned is for the first byte of the failing character, instead of the
72 last byte inspected. This is an incompatible change, but I hope it is small
73 enough not to be a problem. It makes the returned offset consistent with
74 pcre_exec() and pcre_dfa_exec().
75
76 7. pcretest now gives a text phrase as well as the error number when
77 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
78 failure, the offset and reason code are output.
79
80 8. When \R was used with a maximizing quantifier it failed to skip backwards
81 over a \r\n pair if the subsequent match failed. Instead, it just skipped
82 back over a single character (\n). This seems wrong (because it treated the
83 two characters as a single entity when going forwards), conflicts with the
84 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
85 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
86 has been changed.
87
88 9. Some internal refactoring has changed the processing so that the handling
89 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
90 time (the PCRE_DOTALL option was changed this way some time ago: version
91 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
92 which was always a bit of a fudge. It also means that there is one less
93 argument for the match() function, which reduces its stack requirements
94 slightly. This change also fixes an incompatibility with Perl: the pattern
95 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
96
97 10. More internal refactoring has drastically reduced the number of recursive
98 calls to match() for possessively repeated groups such as (abc)++ when
99 using pcre_exec().
100
101 11. While implementing 10, a number of bugs in the handling of groups were
102 discovered and fixed:
103
104 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
105 (a|)*(?1) gave a compile-time internal error.
106 ((a|)+)+ did not notice that the outer group could match an empty string.
107 (^a|^)+ was not marked as anchored.
108 (.*a|.*)+ was not marked as matching at start or after a newline.
109
110 12. Yet more internal refactoring has removed another argument from the match()
111 function. Special calls to this function are now indicated by setting a
112 value in a variable in the "match data" data block.
113
114 13. Be more explicit in pcre_study() instead of relying on "default" for
115 opcodes that mean there is no starting character; this means that when new
116 ones are added and accidentally left out of pcre_study(), testing should
117 pick them up.
118
119 14. The -s option of pcretest has been documented for ages as being an old
120 synonym of -m (show memory usage). I have changed it to mean "force study
121 for every regex", that is, assume /S for every regex. This is similar to -i
122 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
123 using it. It makes it easier to run collections of tests with and without
124 study enabled, and thereby test pcre_study() more easily. All the standard
125 tests are now run with and without -s (but some patterns can be marked as
126 "never study" - see 20 below).
127
128 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
129 restoration of the capturing data to the outer values was not happening
130 correctly.
131
132 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
133 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
134 pattern had matched an empty string, and so incorrectly returned a no
135 match.
136
137 17. There was optimizing code for the last branch of non-capturing parentheses,
138 and also for the obeyed branch of a conditional subexpression, which used
139 tail recursion to cut down on stack usage. Unfortunately, now that there is
140 the possibility of (*THEN) occurring in these branches, tail recursion is
141 no longer possible because the return has to be checked for (*THEN). These
142 two optimizations have therefore been removed.
143
144 18. If a pattern containing \R was studied, it was assumed that \R always
145 matched two bytes, thus causing the minimum subject length to be
146 incorrectly computed because \R can also match just one byte.
147
148 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
149 was incorrectly computed.
150
151 20. If /S is present twice on a test pattern in pcretest input, it now
152 *disables* studying, thereby overriding the use of -s on the command line
153 (see 14 above). This is necessary for one or two tests to keep the output
154 identical in both cases.
155
156 21. When (*ACCEPT) was used in an assertion that matched an empty string and
157 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
158
159 22. When an atomic group that contained a capturing parenthesis was
160 successfully matched, but the branch in which it appeared failed, the
161 capturing was not being forgotten if a higher numbered group was later
162 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
163 group 1 to "a", when in fact it should be unset. This applied to multi-
164 branched capturing and non-capturing groups, repeated or not, and also to
165 positive assertions (capturing in negative assertions does not happen
166 in PCRE) and also to nested atomic groups.
167
168 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
169 subject after a captured substring, to make it easier to tell which of a
170 number of identical substrings has been captured.
171
172 24. The way atomic groups are processed by pcre_exec() has been changed so that
173 if they are repeated, backtracking one repetition now resets captured
174 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
175 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
176 "aaa". Previously, it would have been "a". As part of this code
177 refactoring, the way recursive calls are handled has also been changed.
178
179 25. If an assertion condition captured any substrings, they were not passed
180 back unless some other capturing happened later. For example, if
181 (?(?=(a))a) was matched against "a", no capturing was returned.
182
183 26. When studying a pattern that contained subroutine calls or assertions,
184 the code for finding the minimum length of a possible match was handling
185 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
186 group 1 called group 2 while simultaneously a separate group 2 called group
187 1). A stack overflow occurred in this case. I have fixed this by limiting
188 the recursion depth to 10.
189
190 27. Updated RunTest.bat in the distribution to the version supplied by Tom
191 Fortmann. This supports explicit test numbers on the command line, and has
192 argument validation and error reporting.
193
194 28. An instance of \X with an unlimited repeat could fail if at any point the
195 first character it looked at was a mark character.
196
197 29. Some minor code refactoring concerning Unicode properties and scripts
198 should reduce the stack requirement of match() slightly.
199
200 30. Added the '=' option to pcretest to check the setting of unused capturing
201 slots at the end of the pattern, which are documented as being -1, but are
202 not included in the return count.
203
204 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
205 compiled something random. Now it gives a compile-time error (as does
206 Perl).
207
208 32. A *MARK encountered during the processing of a positive assertion is now
209 recorded and passed back (compatible with Perl).
210
211 33. If --only-matching or --colour was set on a pcregrep call whose pattern
212 had alternative anchored branches, the search for a second match in a line
213 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
214 matched the line "0102" twice. The same bug affected patterns that started
215 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
216 twice.
217
218 34. Previously, PCRE did not allow quantification of assertions. However, Perl
219 does, and because of capturing effects, quantifying parenthesized
220 assertions may at times be useful. Quantifiers are now allowed for
221 parenthesized assertions.
222
223 35. A minor code tidy in pcre_compile() when checking options for \R usage.
224
225 36. \g was being checked for fancy things in a character class, when it should
226 just be a literal "g".
227
228 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
229 appearance of a nested POSIX class supersedes an apparent external class.
230 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
231 unescaped square brackets may also appear as part of class names. For
232 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
233 more like Perl. (But see 8.20/1 above.)
234
235 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
236 was because it thought it was \N{name}, which is not supported).
237
238 39. Add minix to OS list not supporting the -S option in pcretest.
239
240 40. PCRE tries to detect cases of infinite recursion at compile time, but it
241 cannot analyze patterns in sufficient detail to catch mutual recursions
242 such as ((?1))((?2)). There is now a runtime test that gives an error if a
243 subgroup is called recursively as a subpattern for a second time at the
244 same position in the subject string. In previous releases this might have
245 been caught by the recursion limit, or it might have run out of stack.
246
247 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
248 happen only once. PCRE was, however incorrectly giving a compile time error
249 "recursive call could loop indefinitely" because it cannot analyze the
250 pattern in sufficient detail. The compile time test no longer happens when
251 PCRE is compiling a conditional subpattern, but actual runaway loops are
252 now caught at runtime (see 40 above).
253
254 42. It seems that Perl allows any characters other than a closing parenthesis
255 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
256 has been changed to be the same.
257
258 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
259 as not to get warnings when autogen.sh is called. Also changed
260 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
261
262 44. To help people who use pcregrep to scan files containing exceedingly long
263 lines, the following changes have been made:
264
265 (a) The default value of the buffer size parameter has been increased from
266 8K to 20K. (The actual buffer used is three times this size.)
267
268 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
269 PCRE is built.
270
271 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
272 to be set at run time.
273
274 (d) Numerical values in pcregrep options can be followed by K or M, for
275 example --buffer-size=50K.
276
277 (e) If a line being scanned overflows pcregrep's buffer, an error is now
278 given and the return code is set to 2.
279
280 45. Add a pointer to the latest mark to the callout data block.
281
282 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
283 partial match of an empty string instead of no match. This was specific to
284 the use of ".".
285
286 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
287 complete match instead of a partial match. This bug was dependent on both
288 the PCRE_UTF8 and PCRE_DOTALL options being set.
289
290 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
291 starting byte set, because \b was not being ignored.
292
293
294 Version 8.12 15-Jan-2011
295 ------------------------
296
297 1. Fixed some typos in the markup of the man pages, and wrote a script that
298 checks for such things as part of the documentation building process.
299
300 2. On a big-endian 64-bit system, pcregrep did not correctly process the
301 --match-limit and --recursion-limit options (added for 8.11). In
302 particular, this made one of the standard tests fail. (The integer value
303 went into the wrong half of a long int.)
304
305 3. If the --colour option was given to pcregrep with -v (invert match), it
306 did strange things, either producing crazy output, or crashing. It should,
307 of course, ignore a request for colour when reporting lines that do not
308 match.
309
310 4. Another pcregrep bug caused similar problems if --colour was specified with
311 -M (multiline) and the pattern match finished with a line ending.
312
313 5. In pcregrep, when a pattern that ended with a literal newline sequence was
314 matched in multiline mode, the following line was shown as part of the
315 match. This seems wrong, so I have changed it.
316
317 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
318 the check for further matches in the same line (so they could be coloured)
319 to overrun the end of the current line. If another match was found, it was
320 incorrectly shown (and then shown again when found in the next line).
321
322 7. If pcregrep was compiled under Windows, there was a reference to the
323 function pcregrep_exit() before it was defined. I am assuming this was
324 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
325 reported by a user. I've moved the definition above the reference.
326
327
328 Version 8.11 10-Dec-2010
329 ------------------------
330
331 1. (*THEN) was not working properly if there were untried alternatives prior
332 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
333 backtracked to try for "b" instead of moving to the next alternative branch
334 at the same level (in this case, to look for "c"). The Perl documentation
335 is clear that when (*THEN) is backtracked onto, it goes to the "next
336 alternative in the innermost enclosing group".
337
338 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
339 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
340 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
341 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
342 (*THEN).
343
344 3. If \s appeared in a character class, it removed the VT character from
345 the class, even if it had been included by some previous item, for example
346 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
347 of \s, but is part of the POSIX "space" class.)
348
349 4. A partial match never returns an empty string (because you can always
350 match an empty string at the end of the subject); however the checking for
351 an empty string was starting at the "start of match" point. This has been
352 changed to the "earliest inspected character" point, because the returned
353 data for a partial match starts at this character. This means that, for
354 example, /(?<=abc)def/ gives a partial match for the subject "abc"
355 (previously it gave "no match").
356
357 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
358 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
359 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
360 has an implication that the given string is incomplete (because a partial
361 match is preferred over a full match). For this reason, these items now
362 give a partial match in this situation. [Aside: previously, the one case
363 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
364 match rather than a full match, which was wrong by the old rules, but is
365 now correct.]
366
367 6. There was a bug in the handling of #-introduced comments, recognized when
368 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
369 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
370 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
371 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
372 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
373 places in pcre_compile().
374
375 7. Related to (6) above, when pcre_compile() was skipping #-introduced
376 comments when looking ahead for named forward references to subpatterns,
377 the only newline sequence it recognized was NL. It now handles newlines
378 according to the set newline convention.
379
380 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
381 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
382 cater for a lack of strerror(). These oversights have been fixed.
383
384 9. Added --match-limit and --recursion-limit to pcregrep.
385
386 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
387
388 11. When the -o option was used, pcregrep was setting a return code of 1, even
389 when matches were found, and --line-buffered was not being honoured.
390
391 12. Added an optional parentheses number to the -o and --only-matching options
392 of pcregrep.
393
394 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
395 can match an empty string. The code to do it in pcretest and pcredemo
396 needed fixing:
397
398 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
399 only one byte after an empty string match just before CRLF (this case
400 just got forgotten; "any" and "anycrlf" were OK).
401
402 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
403 mode when an empty string match preceded an ASCII character followed by
404 a non-ASCII character. (The code for advancing by one character rather
405 than one byte was nonsense.)
406
407 (c) The pcredemo.c sample program did not have any code at all to handle
408 the cases when CRLF is a valid newline sequence.
409
410 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
411 as a starting offset was within the subject string. There is now a new
412 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
413 negative or greater than the length of the string. In order to test this,
414 pcretest is extended to allow the setting of negative starting offsets.
415
416 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
417 starting offset points to the beginning of a UTF-8 character was
418 unnecessarily clumsy. I tidied it up.
419
420 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
421 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
422
423 17. Nobody had reported that the --include_dir option, which was added in
424 release 7.7 should have been called --include-dir (hyphen, not underscore)
425 for compatibility with GNU grep. I have changed it to --include-dir, but
426 left --include_dir as an undocumented synonym, and the same for
427 --exclude-dir, though that is not available in GNU grep, at least as of
428 release 2.5.4.
429
430 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
431 characters from a string of bytes) have been redefined so as not to use
432 loops, in order to improve performance in some environments. At the same
433 time, I abstracted some of the common code into auxiliary macros to save
434 repetition (this should not affect the compiled code).
435
436 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
437 compile-time error is now given if \c is not followed by an ASCII
438 character, that is, a byte less than 128. (In EBCDIC mode, the code is
439 different, and any byte value is allowed.)
440
441 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
442 START_OPTIMIZE option, which is now allowed at compile time - but just
443 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
444 to pcregrep and other applications that have no direct access to PCRE
445 options. The new /Y option in pcretest sets this option when calling
446 pcre_compile().
447
448 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
449 back references. Groups containing recursive back references were forced to
450 be atomic by that change, but in the case of named groups, the amount of
451 memory required was incorrectly computed, leading to "Failed: internal
452 error: code overflow". This has been fixed.
453
454 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
455 pcretest.c, to avoid build problems in some Borland environments.
456
457
458 Version 8.10 25-Jun-2010
459 ------------------------
460
461 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
462 THEN.
463
464 2. (*ACCEPT) was not working when inside an atomic group.
465
466 3. Inside a character class, \B is treated as a literal by default, but
467 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
468 causes the error). The code is unchanged, but I tidied the documentation.
469
470 4. Inside a character class, PCRE always treated \R and \X as literals,
471 whereas Perl faults them if its -w option is set. I have changed PCRE so
472 that it faults them when PCRE_EXTRA is set.
473
474 5. Added support for \N, which always matches any character other than
475 newline. (It is the same as "." when PCRE_DOTALL is not set.)
476
477 6. When compiling pcregrep with newer versions of gcc which may have
478 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
479 declared with attribute warn_unused_result" were given. Just casting the
480 result to (void) does not stop the warnings; a more elaborate fudge is
481 needed. I've used a macro to implement this.
482
483 7. Minor change to pcretest.c to avoid a compiler warning.
484
485 8. Added four artifical Unicode properties to help with an option to make
486 \s etc use properties (see next item). The new properties are: Xan
487 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
488
489 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
490 use Unicode properties. (*UCP) at the start of a pattern can be used to set
491 this option. Modified pcretest to add /W to test this facility. Added
492 REG_UCP to make it available via the POSIX interface.
493
494 10. Added --line-buffered to pcregrep.
495
496 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
497 studied, and the match started with a letter with a code point greater than
498 127 whose first byte was different to the first byte of the other case of
499 the letter, the other case of this starting letter was not recognized
500 (#976).
501
502 12. If a pattern that was studied started with a repeated Unicode property
503 test, for example, \p{Nd}+, there was the theoretical possibility of
504 setting up an incorrect bitmap of starting bytes, but fortunately it could
505 not have actually happened in practice until change 8 above was made (it
506 added property types that matched character-matching opcodes).
507
508 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
509 possible starting bytes for non-anchored patterns.
510
511 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
512 \R, and also a number of cases that involve Unicode properties, both
513 explicit and implicit when PCRE_UCP is set.
514
515 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
516 input, it could crash or give wrong results if characters with values
517 greater than 0xc0 were present in the subject string. (Detail: it assumed
518 UTF-8 input when processing these items.)
519
520 16. Added a lot of (int) casts to avoid compiler warnings in systems where
521 size_t is 64-bit (#991).
522
523 17. Added a check for running out of memory when PCRE is compiled with
524 --disable-stack-for-recursion (#990).
525
526 18. If the last data line in a file for pcretest does not have a newline on
527 the end, a newline was missing in the output.
528
529 19. The default pcre_chartables.c file recognizes only ASCII characters (values
530 less than 128) in its various bitmaps. However, there is a facility for
531 generating tables according to the current locale when PCRE is compiled. It
532 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
533 characters, are recognized by isspace() and therefore were getting set in
534 these tables, and indeed these tables seem to approximate to ISO 8859. This
535 caused a problem in UTF-8 mode when pcre_study() was used to create a list
536 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
537 which of course cannot start UTF-8 characters. I have changed the code so
538 that only real ASCII characters (less than 128) and the correct starting
539 bytes for UTF-8 encodings are set for characters greater than 127 when in
540 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
541 altogether.)
542
543 20. Added the /T option to pcretest so as to be able to run tests with non-
544 standard character tables, thus making it possible to include the tests
545 used for 19 above in the standard set of tests.
546
547 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
548 reference to a subpattern the other side of a comment that contains an
549 opening parenthesis caused either an internal compiling error, or a
550 reference to the wrong subpattern.
551
552
553 Version 8.02 19-Mar-2010
554 ------------------------
555
556 1. The Unicode data tables have been updated to Unicode 5.2.0.
557
558 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
559 configured.
560
561 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
562 original author of that file, following a query about its status.
563
564 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
565 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
566
567 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
568 quantifier applied to a forward-referencing subroutine call, could compile
569 incorrect code or give the error "internal error: previously-checked
570 referenced subpattern not found".
571
572 6. Both MS Visual Studio and Symbian OS have problems with initializing
573 variables to point to external functions. For these systems, therefore,
574 pcre_malloc etc. are now initialized to local functions that call the
575 relevant global functions.
576
577 7. There were two entries missing in the vectors called coptable and poptable
578 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
579 I've fixed the data, and added a kludgy way of testing at compile time that
580 the lengths are correct (equal to the number of opcodes).
581
582 8. Following on from 7, I added a similar kludge to check the length of the
583 eint vector in pcreposix.c.
584
585 9. Error texts for pcre_compile() are held as one long string to avoid too
586 much relocation at load time. To find a text, the string is searched,
587 counting zeros. There was no check for running off the end of the string,
588 which could happen if a new error number was added without updating the
589 string.
590
591 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
592
593 11. \K was not working if it appeared in an atomic group or in a group that
594 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
595 \K is "not well defined" if used in an assertion. PCRE now accepts it if
596 the assertion is positive, but not if it is negative.
597
598 12. Change 11 fortuitously reduced the size of the stack frame used in the
599 "match()" function of pcre_exec.c by one pointer. Forthcoming
600 implementation of support for (*MARK) will need an extra pointer on the
601 stack; I have reserved it now, so that the stack frame size does not
602 decrease.
603
604 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
605 item in branch that calls a recursion is a subroutine call - as in the
606 second branch in the above example - was incorrectly given the compile-
607 time error "recursive call could loop indefinitely" because pcre_compile()
608 was not correctly checking the subroutine for matching a non-empty string.
609
610 14. The checks for overrunning compiling workspace could trigger after an
611 overrun had occurred. This is a "should never occur" error, but it can be
612 triggered by pathological patterns such as hundreds of nested parentheses.
613 The checks now trigger 100 bytes before the end of the workspace.
614
615 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
616
617
618 Version 8.01 19-Jan-2010
619 ------------------------
620
621 1. If a pattern contained a conditional subpattern with only one branch (in
622 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
623 computed the wrong minimum data length (which is of course zero for such
624 subpatterns). This could cause incorrect "no match" results.
625
626 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
627 the pattern is reset in the first branch, pcre_compile() failed with
628 "internal error: code overflow at offset...". This happened only when
629 the reset was to the original external option setting. (An optimization
630 abstracts leading options settings into an external setting, which was the
631 cause of this.)
632
633 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
634 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
635 assertion pattern did not match (meaning that the assertion was true), it
636 was incorrectly treated as false if the SKIP had been reached during the
637 matching. This also applied to assertions used as conditions.
638
639 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
640 assertion subpattern, including such a pattern used as a condition,
641 unpredictable results occurred, instead of the error return
642 PCRE_ERROR_DFA_UITEM.
643
644 5. The C++ GlobalReplace function was not working like Perl for the special
645 situation when an empty string is matched. It now does the fancy magic
646 stuff that is necessary.
647
648 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
649 removed. (These were left over from very, very early versions of PCRE.)
650
651 7. Some cosmetic changes to the code to make life easier when compiling it
652 as part of something else:
653
654 (a) Change DEBUG to PCRE_DEBUG.
655
656 (b) In pcre_compile(), rename the member of the "branch_chain" structure
657 called "current" as "current_branch", to prevent a collision with the
658 Linux macro when compiled as a kernel module.
659
660 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
661 prevent a collision with the Linux macro when compiled as a kernel
662 module.
663
664 8. In pcre_compile() there are some checks for integer overflows that used to
665 cast potentially large values to (double). This has been changed to that
666 when building, a check for int64_t is made, and if it is found, it is used
667 instead, thus avoiding the use of floating point arithmetic. (There is no
668 other use of FP in PCRE.) If int64_t is not found, the fallback is to
669 double.
670
671 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
672 2005 (difference between two addresses compared to an unsigned value).
673
674 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
675 custom one, because of the following reported problem in Windows:
676
677 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
678 under Win32.
679 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
680 therefore missing the function definition.
681 - The compiler thus generates a "C" signature for the test function.
682 - The linker fails to find the "C" function.
683 - PCRE fails to configure if asked to do so against libbz2.
684
685 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
686 messages were output:
687
688 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
689 rerunning libtoolize, to keep the correct libtool macros in-tree.
690 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
691
692 I have done both of these things.
693
694 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
695 most of the time, it *can* run out if it is given a pattern that contains a
696 runaway infinite recursion. I updated the discussion in the pcrestack man
697 page.
698
699 13. Now that we have gone to the x.xx style of version numbers, the minor
700 version may start with zero. Using 08 or 09 is a bad idea because users
701 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
702 interpreted as invalid octal numbers. I've updated the previous comment in
703 configure.ac, and also added a check that gives an error if 08 or 09 are
704 used.
705
706 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
707 causing partial matching to fail when the end of the subject matched \W
708 in a UTF-8 pattern where \W was quantified with a minimum of 3.
709
710 15. There were some discrepancies between the declarations in pcre_internal.h
711 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
712 their definitions. The declarations used "const uschar *" and the
713 definitions used USPTR. Even though USPTR is normally defined as "const
714 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
715 reported that: "This difference in casting confuses some C++ compilers, for
716 example, SunCC recognizes above declarations as different functions and
717 generates broken code for hbpcre." I have changed the declarations to use
718 USPTR.
719
720 16. GNU libtool is named differently on some systems. The autogen.sh script now
721 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
722 (FreeBSD).
723
724 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
725 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
726 comment: "Figure out how to create a longlong from a string: strtoll and
727 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
728 instance, but it only takes 2 args instead of 3!"
729
730 18. A subtle bug concerned with back references has been fixed by a change of
731 specification, with a corresponding code fix. A pattern such as
732 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
733 refers, was giving matches when it shouldn't. For example, xa=xaaa would
734 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
735 same bug. Such groups have to be quantified to be useful, or contained
736 inside another quantified group. (If there's no repetition, the reference
737 can never match.) The problem arises because, having left the group and
738 moved on to the rest of the pattern, a later failure that backtracks into
739 the group uses the captured value from the final iteration of the group
740 rather than the correct earlier one. I have fixed this in PCRE by forcing
741 any group that contains a reference to itself to be an atomic group; that
742 is, there cannot be any backtracking into it once it has completed. This is
743 similar to recursive and subroutine calls.
744
745
746 Version 8.00 19-Oct-09
747 ----------------------
748
749 1. The table for translating pcre_compile() error codes into POSIX error codes
750 was out-of-date, and there was no check on the pcre_compile() error code
751 being within the table. This could lead to an OK return being given in
752 error.
753
754 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
755 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
756 in a Windows environment.
757
758 3. The pcregrep --count option prints the count for each file even when it is
759 zero, as does GNU grep. However, pcregrep was also printing all files when
760 --files-with-matches was added. Now, when both options are given, it prints
761 counts only for those files that have at least one match. (GNU grep just
762 prints the file name in this circumstance, but including the count seems
763 more useful - otherwise, why use --count?) Also ensured that the
764 combination -clh just lists non-zero counts, with no names.
765
766 4. The long form of the pcregrep -F option was incorrectly implemented as
767 --fixed_strings instead of --fixed-strings. This is an incompatible change,
768 but it seems right to fix it, and I didn't think it was worth preserving
769 the old behaviour.
770
771 5. The command line items --regex=pattern and --regexp=pattern were not
772 recognized by pcregrep, which required --regex pattern or --regexp pattern
773 (with a space rather than an '='). The man page documented the '=' forms,
774 which are compatible with GNU grep; these now work.
775
776 6. No libpcreposix.pc file was created for pkg-config; there was just
777 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
778
779 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
780 when UCP support is not needed, by modifying the Python script that
781 generates it from Unicode data files. This should not matter if the module
782 is correctly used as a library, but I received one complaint about 50K of
783 unwanted data. My guess is that the person linked everything into his
784 program rather than using a library. Anyway, it does no harm.
785
786 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
787 was a minimum greater than 1 for a wide character in a possessive
788 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
789 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
790 character. Chaos in the form of incorrect output or a compiling loop could
791 result.
792
793 9. The restrictions on what a pattern can contain when partial matching is
794 requested for pcre_exec() have been removed. All patterns can now be
795 partially matched by this function. In addition, if there are at least two
796 slots in the offset vector, the offset of the earliest inspected character
797 for the match and the offset of the end of the subject are set in them when
798 PCRE_ERROR_PARTIAL is returned.
799
800 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
801 synonymous with PCRE_PARTIAL, for backwards compatibility, and
802 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
803 and may be more useful for multi-segment matching.
804
805 11. Partial matching with pcre_exec() is now more intuitive. A partial match
806 used to be given if ever the end of the subject was reached; now it is
807 given only if matching could not proceed because another character was
808 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
809 string "Z", which now yields "no match" instead of "partial match". In the
810 case of pcre_dfa_exec(), "no match" is given if every matching path for the
811 final character ended with (*FAIL).
812
813 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
814 if the pattern had a "must contain" character that was already found in the
815 earlier partial match, unless partial matching was again requested. For
816 example, with the pattern /dog.(body)?/, the "must contain" character is
817 "g". If the first part-match was for the string "dog", restarting with
818 "sbody" failed. This bug has been fixed.
819
820 13. The string returned by pcre_dfa_exec() after a partial match has been
821 changed so that it starts at the first inspected character rather than the
822 first character of the match. This makes a difference only if the pattern
823 starts with a lookbehind assertion or \b or \B (\K is not supported by
824 pcre_dfa_exec()). It's an incompatible change, but it makes the two
825 matching functions compatible, and I think it's the right thing to do.
826
827 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
828 so that the demonstration program is easily available in environments where
829 PCRE has not been installed from source.
830
831 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
832 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
833 library.
834
835 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
836 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
837 is not the first non-POSIX option to be added. Clearly some people find
838 these options useful.
839
840 17. If a caller to the POSIX matching function regexec() passes a non-zero
841 value for nmatch with a NULL value for pmatch, the value of
842 nmatch is forced to zero.
843
844 18. RunGrepTest did not have a test for the availability of the -u option of
845 the diff command, as RunTest does. It now checks in the same way as
846 RunTest, and also checks for the -b option.
847
848 19. If an odd number of negated classes containing just a single character
849 interposed, within parentheses, between a forward reference to a named
850 subpattern and the definition of the subpattern, compilation crashed with
851 an internal error, complaining that it could not find the referenced
852 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
853 [The bug was that it was starting one character too far in when skipping
854 over the character class, thus treating the ] as data rather than
855 terminating the class. This meant it could skip too much.]
856
857 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
858 /g option in pcretest when the pattern contains \K, which makes it possible
859 to have an empty string match not at the start, even when the pattern is
860 anchored. Updated pcretest and pcredemo to use this option.
861
862 21. If the maximum number of capturing subpatterns in a recursion was greater
863 than the maximum at the outer level, the higher number was returned, but
864 with unset values at the outer level. The correct (outer level) value is
865 now given.
866
867 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
868 PCRE did not set those parentheses (unlike Perl). I have now found a way to
869 make it do so. The string so far is captured, making this feature
870 compatible with Perl.
871
872 23. The tests have been re-organized, adding tests 11 and 12, to make it
873 possible to check the Perl 5.10 features against Perl 5.10.
874
875 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
876 pattern matches a fixed length string. PCRE did not allow this; now it
877 does. Neither allows recursion.
878
879 25. I finally figured out how to implement a request to provide the minimum
880 length of subject string that was needed in order to match a given pattern.
881 (It was back references and recursion that I had previously got hung up
882 on.) This code has now been added to pcre_study(); it finds a lower bound
883 to the length of subject needed. It is not necessarily the greatest lower
884 bound, but using it to avoid searching strings that are too short does give
885 some useful speed-ups. The value is available to calling programs via
886 pcre_fullinfo().
887
888 26. While implementing 25, I discovered to my embarrassment that pcretest had
889 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
890 study optimizations had never been tested with that matching function.
891 Oops. What is worse, even when it was passed study data, there was a bug in
892 pcre_dfa_exec() that meant it never actually used it. Double oops. There
893 were also very few tests of studied patterns with pcre_dfa_exec().
894
895 27. If (?| is used to create subpatterns with duplicate numbers, they are now
896 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
897 on the other side of the coin, they are no longer allowed to have different
898 names, because these cannot be distinguished in PCRE, and this has caused
899 confusion. (This is a difference from Perl.)
900
901 28. When duplicate subpattern names are present (necessarily with different
902 numbers, as required by 27 above), and a test is made by name in a
903 conditional pattern, either for a subpattern having been matched, or for
904 recursion in such a pattern, all the associated numbered subpatterns are
905 tested, and the overall condition is true if the condition is true for any
906 one of them. This is the way Perl works, and is also more like the way
907 testing by number works.
908
909
910 Version 7.9 11-Apr-09
911 ---------------------
912
913 1. When building with support for bzlib/zlib (pcregrep) and/or readline
914 (pcretest), all targets were linked against these libraries. This included
915 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
916 libraries. This caused unwanted dependencies to be created. This problem
917 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
918 pcretest is linked with readline.
919
920 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
921 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
922 moved outside it again, because FALSE and TRUE are already defined in AIX,
923 but BOOL is not.
924
925 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
926 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
927
928 4. The pcregrep documentation said spaces were inserted as well as colons (or
929 hyphens) following file names and line numbers when outputting matching
930 lines. This is not true; no spaces are inserted. I have also clarified the
931 wording for the --colour (or --color) option.
932
933 5. In pcregrep, when --colour was used with -o, the list of matching strings
934 was not coloured; this is different to GNU grep, so I have changed it to be
935 the same.
936
937 6. When --colo(u)r was used in pcregrep, only the first matching substring in
938 each matching line was coloured. Now it goes on to look for further matches
939 of any of the test patterns, which is the same behaviour as GNU grep.
940
941 7. A pattern that could match an empty string could cause pcregrep to loop; it
942 doesn't make sense to accept an empty string match in pcregrep, so I have
943 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
944 seems to be how GNU grep behaves.
945
946 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
947 start or after a newline", because the conditional assertion was not being
948 correctly handled. The rule now is that both the assertion and what follows
949 in the first alternative must satisfy the test.
950
951 9. If auto-callout was enabled in a pattern with a conditional group whose
952 condition was an assertion, PCRE could crash during matching, both with
953 pcre_exec() and pcre_dfa_exec().
954
955 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
956 used for matching.
957
958 11. Unicode property support in character classes was not working for
959 characters (bytes) greater than 127 when not in UTF-8 mode.
960
961 12. Added the -M command line option to pcretest.
962
963 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
964
965 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
966
967 16. Added comments and documentation about mis-use of no_arg in the C++
968 wrapper.
969
970 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
971 from Martin Jerabek that uses macro names for all relevant character and
972 string constants.
973
974 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
975 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
976 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
977 these, but not everybody uses configure.
978
979 19. A conditional group that had only one branch was not being correctly
980 recognized as an item that could match an empty string. This meant that an
981 enclosing group might also not be so recognized, causing infinite looping
982 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
983 with the subject "ab", where knowledge that the repeated group can match
984 nothing is needed in order to break the loop.
985
986 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
987 exec(), but without supplying a callout function, matching went wrong.
988
989 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
990 leak if the size of the offset vector was greater than 30. When the vector
991 is smaller, the saved offsets during recursion go onto a local stack
992 vector, but for larger vectors malloc() is used. It was failing to free
993 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
994 error, in fact).
995
996 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
997 heapframe that is used only when UTF-8 support is enabled. This caused no
998 problem, but was untidy.
999
1000 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1001 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1002 included within another project.
1003
1004 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1005 slightly modified by me:
1006
1007 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1008 not building pcregrep.
1009
1010 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1011 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1012
1013 25. Forward references, both numeric and by name, in patterns that made use of
1014 duplicate group numbers, could behave incorrectly or give incorrect errors,
1015 because when scanning forward to find the reference group, PCRE was not
1016 taking into account the duplicate group numbers. A pattern such as
1017 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1018
1019 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1020 the feature of a custom pointer more persuasive (as requested by a user).
1021
1022 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1023 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1024 pre-defined.
1025
1026 28. Added support for (*UTF8) at the start of a pattern.
1027
1028 29. Arrange for flags added by the "release type" setting in CMake to be shown
1029 in the configuration summary.
1030
1031
1032 Version 7.8 05-Sep-08
1033 ---------------------
1034
1035 1. Replaced UCP searching code with optimized version as implemented for Ad
1036 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1037 stage table and inline lookup instead of a function, giving speed ups of 2
1038 to 5 times on some simple patterns that I tested. Permission was given to
1039 distribute the MultiStage2.py script that generates the tables (it's not in
1040 the tarball, but is in the Subversion repository).
1041
1042 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1043 scripts.
1044
1045 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1046 a group with a zero qualifier. The result of the study could be incorrect,
1047 or the function might crash, depending on the pattern.
1048
1049 4. Caseless matching was not working for non-ASCII characters in back
1050 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1051 It now works when Unicode Property Support is available.
1052
1053 5. In pcretest, an escape such as \x{de} in the data was always generating
1054 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1055 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1056 truncation.
1057
1058 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1059
1060 7. Added two (int) casts to pcregrep when printing the difference of two
1061 pointers, in case they are 64-bit values.
1062
1063 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1064 test 2 if it fails.
1065
1066 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1067 and a #define of that name to empty if it is not externally set. This is to
1068 allow users of MSVC to set it if necessary.
1069
1070 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1071 the convenience functions in the pcre_get.c source file.
1072
1073 11. An option change at the start of a pattern that had top-level alternatives
1074 could cause overwriting and/or a crash. This command provoked a crash in
1075 some environments:
1076
1077 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1078
1079 This potential security problem was recorded as CVE-2008-2371.
1080
1081 12. For a pattern where the match had to start at the beginning or immediately
1082 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1083 pcre_dfa_exec() could read past the end of the passed subject if there was
1084 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1085 pcretest so that it places the subject at the end of its malloc-ed buffer.
1086
1087 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1088 exec() might read past the end of the data buffer in UTF-8 mode.
1089
1090 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1091 the data contained the byte 0x85 as part of a UTF-8 character within its
1092 first line. This applied both to normal and DFA matching.
1093
1094 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1095 /^[^d]*?$/8 failed to match "abc".
1096
1097 16. Added a missing copyright notice to pcrecpp_internal.h.
1098
1099 17. Make it more clear in the documentation that values returned from
1100 pcre_exec() in ovector are byte offsets, not character counts.
1101
1102 18. Tidied a few places to stop certain compilers from issuing warnings.
1103
1104 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1105 supplied by Stefan Weber. I made a further small update for 7.8 because
1106 there is a change of source arrangements: the pcre_searchfuncs.c module is
1107 replaced by pcre_ucd.c.
1108
1109
1110 Version 7.7 07-May-08
1111 ---------------------
1112
1113 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1114 a string to a long long, pretend we don't even have a long long." This is
1115 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1116
1117 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1118 pre-7.6 versions, which defined a global no_arg variable instead of putting
1119 it in the RE class. (See also #8 below.)
1120
1121 3. Remove a line of dead code, identified by coverity and reported by Nuno
1122 Lopes.
1123
1124 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1125
1126 (1) The include/exclude patterns were being applied to the whole pathnames
1127 of files, instead of just to the final components.
1128
1129 (2) If there was more than one level of directory, the subdirectories were
1130 skipped unless they satisfied the include/exclude conditions. This is
1131 inconsistent with GNU grep (and could even be seen as contrary to the
1132 pcregrep specification - which I improved to make it absolutely clear).
1133 The action now is always to scan all levels of directory, and just
1134 apply the include/exclude patterns to regular files.
1135
1136 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1137 --exclude_dir in the tests to avoid scanning .svn directories.
1138
1139 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1140 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1141 doesn't support NULs in patterns.
1142
1143 7. Added some missing "const"s to declarations of static tables in
1144 pcre_compile.c and pcre_dfa_exec.c.
1145
1146 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1147 caused by fix #2 above. (Subsequently also a second patch to fix the
1148 first patch. And a third patch - this was a messy problem.)
1149
1150 9. Applied Craig's patch to remove the use of push_back().
1151
1152 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1153 matching function regexec().
1154
1155 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1156 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1157 references. PCRE supports relative numbers with this syntax (I don't think
1158 Oniguruma does).
1159
1160 12. Previously, a group with a zero repeat such as (...){0} was completely
1161 omitted from the compiled regex. However, this means that if the group
1162 was called as a subroutine from elsewhere in the pattern, things went wrong
1163 (an internal error was given). Such groups are now left in the compiled
1164 pattern, with a new opcode that causes them to be skipped at execution
1165 time.
1166
1167 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1168 to the way PCRE behaves:
1169
1170 (a) A lone ] character is dis-allowed (Perl treats it as data).
1171
1172 (b) A back reference to an unmatched subpattern matches an empty string
1173 (Perl fails the current match path).
1174
1175 (c) A data ] in a character class must be notated as \] because if the
1176 first data character in a class is ], it defines an empty class. (In
1177 Perl it is not possible to have an empty class.) The empty class []
1178 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1179 The negative empty class [^] matches any one character, independently
1180 of the DOTALL setting.
1181
1182 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1183 non-existent subpattern following a character class starting with ']' and
1184 containing () gave an internal compiling error instead of "reference to
1185 non-existent subpattern". Fortunately, when the pattern did exist, the
1186 compiled code was correct. (When scanning forwards to check for the
1187 existencd of the subpattern, it was treating the data ']' as terminating
1188 the class, so got the count wrong. When actually compiling, the reference
1189 was subsequently set up correctly.)
1190
1191 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1192 it was being rejected as not supported by pcre_dfa_exec(), even though
1193 other assertions are supported. I have made pcre_dfa_exec() support
1194 (*FAIL).
1195
1196 16. The implementation of 13c above involved the invention of a new opcode,
1197 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1198 cannot be changed at match time, I realized I could make a small
1199 improvement to matching performance by compiling OP_ALLANY instead of
1200 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1201 on the OP_ANY path.
1202
1203 17. Compiling pcretest on Windows with readline support failed without the
1204 following two fixes: (1) Make the unistd.h include conditional on
1205 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1206
1207 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1208 ncurses library to be included for pcretest when ReadLine support is
1209 requested, but also to allow for it to be overridden. This patch came from
1210 Daniel Bergström.
1211
1212 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1213 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1214 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1215 spotting this.
1216
1217
1218 Version 7.6 28-Jan-08
1219 ---------------------
1220
1221 1. A character class containing a very large number of characters with
1222 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1223 overflow.
1224
1225 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1226 HAVE_LONG_LONG is not defined.
1227
1228 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1229 bring them up to date and include new features. This patch includes:
1230
1231 - Fixed PH's badly added libz and libbz2 support.
1232 - Fixed a problem with static linking.
1233 - Added pcredemo. [But later removed - see 7 below.]
1234 - Fixed dftables problem and added an option.
1235 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1236 HAVE_LONG_LONG.
1237 - Added readline support for pcretest.
1238 - Added an listing of the option settings after cmake has run.
1239
1240 4. A user submitted a patch to Makefile that makes it easy to create
1241 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1242 Makefile.am that cause it to include this special target, without
1243 affecting anything else. Note that the same mingw target plus all
1244 the other distribution libraries and programs are now supported
1245 when configuring with CMake (see 6 below) instead of with
1246 Configure/Make.
1247
1248 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1249 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1250 exported in the Windows port". It has not yet been confirmed that the patch
1251 solves the problem, but it does no harm.
1252
1253 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1254 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1255 with CMake, and also correct the comment about stack recursion.
1256
1257 7. Remove the automatic building of pcredemo from the ./configure system and
1258 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1259 of a program that users should build themselves after PCRE is installed, so
1260 building it automatically is not really right. What is more, it gave
1261 trouble in some build environments.
1262
1263 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1264
1265
1266 Version 7.5 10-Jan-08
1267 ---------------------
1268
1269 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1270 values in parens when parsing an RE using the C++ wrapper."
1271
1272 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1273 Characters greater than 255 were excluded from the class instead of being
1274 included.
1275
1276 3. The same bug as (2) above applied to negated POSIX classes such as
1277 [:^space:].
1278
1279 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1280 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1281 I have changed it.
1282
1283 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1284 first named subpattern) and a construct such as (?&a) would reference the
1285 first named subpattern whose name started with "a" (in other words, the
1286 length check was missing). Both these problems are fixed. "Subpattern name
1287 expected" is now given for (?&) (a zero-length name), and this patch also
1288 makes it give the same error for \k'' (previously it complained that that
1289 was a reference to a non-existent subpattern).
1290
1291 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1292 this is right because (?- can be followed by option settings as well as by
1293 digits. I have, however, made the messages clearer.
1294
1295 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1296 than the number used in the conditional) now cause a compile-time error.
1297 This is actually not compatible with Perl, which accepts such patterns, but
1298 treats the conditional as always being FALSE (as PCRE used to), but it
1299 seems to me that giving a diagnostic is better.
1300
1301 8. Change "alphameric" to the more common word "alphanumeric" in comments
1302 and messages.
1303
1304 9. Fix two occurrences of "backslash" in comments that should have been
1305 "backspace".
1306
1307 10. Remove two redundant lines of code that can never be obeyed (their function
1308 was moved elsewhere).
1309
1310 11. The program that makes PCRE's Unicode character property table had a bug
1311 which caused it to generate incorrect table entries for sequences of
1312 characters that have the same character type, but are in different scripts.
1313 It amalgamated them into a single range, with the script of the first of
1314 them. In other words, some characters were in the wrong script. There were
1315 thirteen such cases, affecting characters in the following ranges:
1316
1317 U+002b0 - U+002c1
1318 U+0060c - U+0060d
1319 U+0061e - U+00612
1320 U+0064b - U+0065e
1321 U+0074d - U+0076d
1322 U+01800 - U+01805
1323 U+01d00 - U+01d77
1324 U+01d9b - U+01dbf
1325 U+0200b - U+0200f
1326 U+030fc - U+030fe
1327 U+03260 - U+0327f
1328 U+0fb46 - U+0fbb1
1329 U+10450 - U+1049d
1330
1331 12. The -o option (show only the matching part of a line) for pcregrep was not
1332 compatible with GNU grep in that, if there was more than one match in a
1333 line, it showed only the first of them. It now behaves in the same way as
1334 GNU grep.
1335
1336 13. If the -o and -v options were combined for pcregrep, it printed a blank
1337 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1338 does the same. The return code can be used to tell if there were any
1339 non-matching lines.
1340
1341 14. Added --file-offsets and --line-offsets to pcregrep.
1342
1343 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1344 infinitely looping recursion. The bug was that positive lookaheads were not
1345 being skipped when checking for a possible empty match (negative lookaheads
1346 and both kinds of lookbehind were skipped).
1347
1348 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1349 inclusion of <windows.h> to before rather than after the definition of
1350 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1351
1352 17. Specifying a possessive quantifier with a specific limit for a Unicode
1353 character property caused pcre_compile() to compile bad code, which led at
1354 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1355 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1356 caused the error; without that there was no problem.
1357
1358 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1359
1360 19. Added --enable-pcretest-libreadline.
1361
1362 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1363 RE::GlobalReplace(). As a result, the number of replacements returned was
1364 double what it should be. I removed one of the increments, but Craig sent a
1365 later patch that removed the other one (the right fix) and added unit tests
1366 that check the return values (which was not done before).
1367
1368 21. Several CMake things:
1369
1370 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1371 the names libpcre and libpcreposix, not just pcre and pcreposix.
1372
1373 (2) The above change means that pcretest and pcregrep are now correctly
1374 linked with the newly-built libraries, not previously installed ones.
1375
1376 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1377
1378 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1379 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1380 UTF-8 newline character). The key issue is that the pattern starts .*;
1381 this means that the match must be either at the beginning, or after a
1382 newline. The bug was in the code for advancing after a failed match and
1383 checking that the new position followed a newline. It was not taking
1384 account of UTF-8 characters correctly.
1385
1386 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1387 character classes. PCRE was not treating the sequence [:...:] as a
1388 character class unless the ... were all letters. Perl, however, seems to
1389 allow any characters between [: and :], though of course it rejects as
1390 unknown any "names" that contain non-letters, because all the known class
1391 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1392 for example, whereas PCRE did not - it did not recognize a POSIX character
1393 class. This seemed a bit dangerous, so the code has been changed to be
1394 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1395 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1396 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1397 Perl does, and where it didn't before.
1398
1399 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1400 Windows environments %n is disabled by default.
1401
1402
1403 Version 7.4 21-Sep-07
1404 ---------------------
1405
1406 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1407 means that a class such as [\s] counted as "explicit reference to CR or
1408 LF". That isn't really right - the whole point of the change was to try to
1409 help when there was an actual mention of one of the two characters. So now
1410 the change happens only if \r or \n (or a literal CR or LF) character is
1411 encountered.
1412
1413 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1414 of both had grown to the point where there were only 3 bits left.
1415 Fortunately, there was spare space in the data structure, and so I have
1416 moved the internal flags into a new 16-bit field to free up more option
1417 bits.
1418
1419 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1420 but did not set the internal JCHANGED flag - either of these is enough to
1421 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1422 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1423 start sets both bits.
1424
1425 4. Added options (at build time, compile time, exec time) to change \R from
1426 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1427
1428 5. doc/pcresyntax.html was missing from the distribution.
1429
1430 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1431 compatibility, even though it is no longer used.
1432
1433 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1434 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1435 windows.h file is present (where different names are used). [This was
1436 reversed later after testing - see 16 below.]
1437
1438 8. Changed all #include <config.h> to #include "config.h". There were also
1439 some further <pcre.h> cases that I changed to "pcre.h".
1440
1441 9. When pcregrep was used with the --colour option, it missed the line ending
1442 sequence off the lines that it output.
1443
1444 10. It was pointed out to me that arrays of string pointers cause lots of
1445 relocations when a shared library is dynamically loaded. A technique of
1446 using a single long string with a table of offsets can drastically reduce
1447 these. I have refactored PCRE in four places to do this. The result is
1448 dramatic:
1449
1450 Originally: 290
1451 After changing UCP table: 187
1452 After changing error message table: 43
1453 After changing table of "verbs" 36
1454 After changing table of Posix names 22
1455
1456 Thanks to the folks working on Gregex for glib for this insight.
1457
1458 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1459 unicode-properties was also set.
1460
1461 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1462
1463 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1464 checked only for CRLF.
1465
1466 14. Added casts to pcretest.c to avoid compiler warnings.
1467
1468 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1469
1470 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1471 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1472 entirely. This removes changes made in 7 above.
1473
1474 17. The CMake files have been updated, and there is now more information about
1475 building with CMake in the NON-UNIX-USE document.
1476
1477
1478 Version 7.3 28-Aug-07
1479 ---------------------
1480
1481 1. In the rejigging of the build system that eventually resulted in 7.1, the
1482 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1483 brackets there is not right, since it causes compilers to look for an
1484 installed pcre.h, not the version that is in the source that is being
1485 compiled (which of course may be different). I have changed it back to:
1486
1487 #include "pcre.h"
1488
1489 I have a vague recollection that the change was concerned with compiling in
1490 different directories, but in the new build system, that is taken care of
1491 by the VPATH setting the Makefile.
1492
1493 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1494 when the subject happened to end in the byte 0x85 (e.g. if the last
1495 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1496 characters but of course it shouldn't be taken as a newline when it is part
1497 of another character. The bug was that, for an unlimited repeat of . in
1498 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1499 characters when looking for a newline.
1500
1501 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1502
1503 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1504 in debug output.
1505
1506 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1507 long printing in the pcrecpp unittest when running under MinGW.
1508
1509 6. ESC_K was left out of the EBCDIC table.
1510
1511 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1512 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1513 limit also applies to "virtual nesting" when a pattern is recursive, and in
1514 this case 1000 isn't so big. I have been able to remove this limit at the
1515 expense of backing off one optimization in certain circumstances. Normally,
1516 when pcre_exec() would call its internal match() function recursively and
1517 immediately return the result unconditionally, it uses a "tail recursion"
1518 feature to save stack. However, when a subpattern that can match an empty
1519 string has an unlimited repetition quantifier, it no longer makes this
1520 optimization. That gives it a stack frame in which to save the data for
1521 checking that an empty string has been matched. Previously this was taken
1522 from the 1000-entry workspace that had been reserved. So now there is no
1523 explicit limit, but more stack is used.
1524
1525 8. Applied Daniel's patches to solve problems with the import/export magic
1526 syntax that is required for Windows, and which was going wrong for the
1527 pcreposix and pcrecpp parts of the library. These were overlooked when this
1528 problem was solved for the main library.
1529
1530 9. There were some crude static tests to avoid integer overflow when computing
1531 the size of patterns that contain repeated groups with explicit upper
1532 limits. As the maximum quantifier is 65535, the maximum group length was
1533 set at 30,000 so that the product of these two numbers did not overflow a
1534 32-bit integer. However, it turns out that people want to use groups that
1535 are longer than 30,000 bytes (though not repeat them that many times).
1536 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1537 made it possible to implement the integer overflow checks in a much more
1538 dynamic way, which I have now done. The artificial limitation on group
1539 length has been removed - we now have only the limit on the total length of
1540 the compiled pattern, which depends on the LINK_SIZE setting.
1541
1542 10. Fixed a bug in the documentation for get/copy named substring when
1543 duplicate names are permitted. If none of the named substrings are set, the
1544 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1545 empty string.
1546
1547 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1548 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1549 because the ] is interpreted as the first data character and the
1550 terminating ] is not found. PCRE has been made compatible with Perl in this
1551 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1552 cause memory overwriting.
1553
1554 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1555 string has been matched (to stop an infinite loop). It was not recognizing
1556 a conditional subpattern that could match an empty string if that
1557 subpattern was within another subpattern. For example, it looped when
1558 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1559 condition was not nested. This bug has been fixed.
1560
1561 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1562 past the start of the subject in the presence of bytes with the top bit
1563 set, for example "\x8aBCD".
1564
1565 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1566 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1567
1568 14. Optimized (?!) to (*FAIL).
1569
1570 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1571 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1572 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1573 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1574 does: it's just the validity check that is more restrictive.
1575
1576 16. Inserted checks for integer overflows during escape sequence (backslash)
1577 processing, and also fixed erroneous offset values for syntax errors during
1578 backslash processing.
1579
1580 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1581 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1582
1583 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1584 caused an overrun.
1585
1586 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1587 something other than just ASCII characters) inside a group that had an
1588 unlimited repeat caused a loop at compile time (while checking to see
1589 whether the group could match an empty string).
1590
1591 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1592 [\P{Any}] did so. (Error in the code for printing property names.)
1593
1594 21. An orphan \E inside a character class could cause a crash.
1595
1596 22. A repeated capturing bracket such as (A)? could cause a wild memory
1597 reference during compilation.
1598
1599 23. There are several functions in pcre_compile() that scan along a compiled
1600 expression for various reasons (e.g. to see if it's fixed length for look
1601 behind). There were bugs in these functions when a repeated \p or \P was
1602 present in the pattern. These operators have additional parameters compared
1603 with \d, etc, and these were not being taken into account when moving along
1604 the compiled data. Specifically:
1605
1606 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1607 length.
1608
1609 (b) An item such as \pL+ within a repeated group could cause crashes or
1610 loops.
1611
1612 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1613 "reference to non-existent subpattern" error.
1614
1615 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1616
1617 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1618 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1619
1620 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1621
1622 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1623 character were causing crashes (broken optimization).
1624
1625 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1626 \p or \P) caused a compile-time loop.
1627
1628 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1629 break. For example, the unstudied pattern [\r\n]A does not match the string
1630 "\r\nA" because change 7.0/46 below moves the current point on by two
1631 characters after failing to match at the start. However, the pattern \nA
1632 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1633 the same is true. There doesn't seem any very clean way out of this, but
1634 what I have chosen to do makes the common cases work: PCRE now takes note
1635 of whether there can be an explicit match for \r or \n anywhere in the
1636 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1637 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1638 pattern has explicit CR or LF references.
1639
1640 29. Added (*CR) etc for changing newline setting at start of pattern.
1641
1642
1643 Version 7.2 19-Jun-07
1644 ---------------------
1645
1646 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1647 which is apparently normally available under Windows.
1648
1649 2. Re-jig the pcregrep tests with different newline settings in an attempt
1650 to make them independent of the local environment's newline setting.
1651
1652 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1653
1654 4. Some of the "internals" tests were previously cut out when the link size
1655 was not 2, because the output contained actual offsets. The recent new
1656 "Z" feature of pcretest means that these can be cut out, making the tests
1657 usable with all link sizes.
1658
1659 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1660 stack recursion. This gives a massive performance boost under BSD, but just
1661 a small improvement under Linux. However, it saves one field in the frame
1662 in all cases.
1663
1664 6. Added more features from the forthcoming Perl 5.10:
1665
1666 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1667 recursion call. It refers to the nth most recently opened parentheses.
1668
1669 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1670 to be opened parentheses.
1671
1672 (c) Conditions that refer to capturing parentheses can be specified
1673 relatively, for example, (?(-2)... or (?(+3)...
1674
1675 (d) \K resets the start of the current match so that everything before
1676 is not part of it.
1677
1678 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1679
1680 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1681 reference syntax.
1682
1683 (g) (?| introduces a group in which the numbering of parentheses in each
1684 alternative starts with the same number.
1685
1686 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1687
1688 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1689 PCRE_INFO_JCHANGED.
1690
1691 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1692 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1693 for detecting groups that can match an empty string.
1694
1695 9. A pattern with a very large number of alternatives (more than several
1696 hundred) was running out of internal workspace during the pre-compile
1697 phase, where pcre_compile() figures out how much memory will be needed. A
1698 bit of new cunning has reduced the workspace needed for groups with
1699 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1700 workspace instead of running out of the 4096 that are available.
1701
1702 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1703
1704 11. Applied patch from Google to remove an optimization that didn't quite work.
1705 The report of the bug said:
1706
1707 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1708 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1709 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1710
1711 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1712 it matched the wrong number of bytes.
1713
1714
1715 Version 7.1 24-Apr-07
1716 ---------------------
1717
1718 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1719 that is more "standard", making use of automake and other Autotools. There
1720 is some re-arrangement of the files and adjustment of comments consequent
1721 on this.
1722
1723 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1724 for recursive directory scanning broke on some systems because the files
1725 are not scanned in any specific order and on different systems the order
1726 was different. A call to "sort" has been inserted into RunGrepTest for the
1727 approprate test as a short-term fix. In the longer term there may be an
1728 alternative.
1729
1730 3. I had an email from Eric Raymond about problems translating some of PCRE's
1731 man pages to HTML (despite the fact that I distribute HTML pages, some
1732 people do their own conversions for various reasons). The problems
1733 concerned the use of low-level troff macros .br and .in. I have therefore
1734 removed all such uses from the man pages (some were redundant, some could
1735 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1736 HTML has been updated to handle .nf/.fi and to complain if it encounters
1737 .br or .in.
1738
1739 4. Updated comments in configure.ac that get placed in config.h.in and also
1740 arranged for config.h to be included in the distribution, with the name
1741 config.h.generic, for the benefit of those who have to compile without
1742 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1743
1744 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1745 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1746 makevp.bat for the current PCRE, using the additional files
1747 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1748
1749 6. A Windows user reported a minor discrepancy with test 2, which turned out
1750 to be caused by a trailing space on an input line that had got lost in his
1751 copy. The trailing space was an accident, so I've just removed it.
1752
1753 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1754 that is needed.
1755
1756 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1757 as "const" (a) because they are and (b) because it helps the PHP
1758 maintainers who have recently made a script to detect big data structures
1759 in the php code that should be moved to the .rodata section. I remembered
1760 to update Builducptable as well, so it won't revert if ucptable.h is ever
1761 re-created.
1762
1763 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1764 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1765 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1766 support is not required. This saves 1.5-2K of code, which is important in
1767 some applications.
1768
1769 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1770 so as not to refer to the tables, even though these functions will never be
1771 called when UTF-8 support is disabled. Otherwise there are problems with a
1772 shared library.
1773
1774 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1775
1776 (a) It was defining its arguments as char * instead of void *.
1777
1778 (b) It was assuming that all moves were upwards in memory; this was true
1779 a long time ago when I wrote it, but is no longer the case.
1780
1781 The emulated memove() is provided for those environments that have neither
1782 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1783 is clearly not the case, as these two bugs were recently reported.
1784
1785 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1786 and Detrail to create the HTML documentation, the .txt form of the man
1787 pages, and it removes trailing spaces from listed files. It also creates
1788 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1789 case, it wraps all the #defines with #ifndefs. This script should be run
1790 before "make dist".
1791
1792 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1793 with Unicode property support.
1794
1795 (a) For a maximizing quantifier, if the two different cases of the
1796 character were of different lengths in their UTF-8 codings (there are
1797 some cases like this - I found 11), and the matching function had to
1798 back up over a mixture of the two cases, it incorrectly assumed they
1799 were both the same length.
1800
1801 (b) When PCRE was configured to use the heap rather than the stack for
1802 recursion during matching, it was not correctly preserving the data for
1803 the other case of a UTF-8 character when checking ahead for a match
1804 while processing a minimizing repeat. If the check also involved
1805 matching a wide character, but failed, corruption could cause an
1806 erroneous result when trying to check for a repeat of the original
1807 character.
1808
1809 13. Some tidying changes to the testing mechanism:
1810
1811 (a) The RunTest script now detects the internal link size and whether there
1812 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1813 values substituted by "configure". (The RunGrepTest script already did
1814 this for UTF-8.) The configure.ac script no longer substitutes the
1815 relevant variables.
1816
1817 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1818 with length and offset values. This means that the output is different
1819 for different internal link sizes. Test 2 is skipped for link sizes
1820 other than 2 because of this, bypassing the problem. Unfortunately,
1821 there was also a test in test 3 (the locale tests) that used /B and
1822 failed for link sizes other than 2. Rather than cut the whole test out,
1823 I have added a new /Z option to pcretest that replaces the length and
1824 offset values with spaces. This is now used to make test 3 independent
1825 of link size. (Test 2 will be tidied up later.)
1826
1827 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1828 segmentation fault instead of returning the appropriate error message.
1829
1830 15. In multiline mode when the newline sequence was set to "any", the pattern
1831 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1832 This doesn't seem right; it now treats the CRLF combination as the line
1833 ending, and so does not match in that case. It's only a pattern such as ^$
1834 that would hit this one: something like ^ABC$ would have failed after \r
1835 and then tried again after \r\n.
1836
1837 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1838 in an attempt to make files that differ only in their line terminators
1839 compare equal. This works on Linux.
1840
1841 17. Under certain error circumstances pcregrep might try to free random memory
1842 as it exited. This is now fixed, thanks to valgrind.
1843
1844 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1845 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1846 was because its rules for how to advance for /g after matching an empty
1847 string at the end of a line did not allow for this case. They now check for
1848 it specially.
1849
1850 20. pcretest is supposed to handle patterns and data of any length, by
1851 extending its buffers when necessary. It was getting this wrong when the
1852 buffer for a data line had to be extended.
1853
1854 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1855 CRLF as a newline sequence.
1856
1857 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1858 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1859 I have nevertheless tidied it up.
1860
1861 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1862
1863 24. Added a man page for pcre-config.
1864
1865
1866 Version 7.0 19-Dec-06
1867 ---------------------
1868
1869 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1870 moving to gcc 4.1.1.
1871
1872 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1873 sys/time.h, which is documented as needed for this function. It doesn't
1874 seem to matter on Linux, but it showed up on some releases of OS X.
1875
1876 3. It seems that there are systems where bytes whose values are greater than
1877 127 match isprint() in the "C" locale. The "C" locale should be the
1878 default when a C program starts up. In most systems, only ASCII printing
1879 characters match isprint(). This difference caused the output from pcretest
1880 to vary, making some of the tests fail. I have changed pcretest so that:
1881
1882 (a) When it is outputting text in the compiled version of a pattern, bytes
1883 other than 32-126 are always shown as hex escapes.
1884
1885 (b) When it is outputting text that is a matched part of a subject string,
1886 it does the same, unless a different locale has been set for the match
1887 (using the /L modifier). In this case, it uses isprint() to decide.
1888
1889 4. Fixed a major bug that caused incorrect computation of the amount of memory
1890 required for a compiled pattern when options that changed within the
1891 pattern affected the logic of the preliminary scan that determines the
1892 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1893 that the computed length was too small. The symptoms of this bug were
1894 either the PCRE error "internal error: code overflow" from pcre_compile(),
1895 or a glibc crash with a message such as "pcretest: free(): invalid next
1896 size (fast)". Examples of patterns that provoked this bug (shown in
1897 pcretest format) are:
1898
1899 /(?-x: )/x
1900 /(?x)(?-x: \s*#\s*)/
1901 /((?i)[\x{c0}])/8
1902 /(?i:[\x{c0}])/8
1903
1904 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1905 is now done differently.
1906
1907 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1908 wrapper classes; (b) implement a new function in the C++ scanner that is
1909 more efficient than the old way of doing things because it avoids levels of
1910 recursion in the regex matching; (c) add a paragraph to the documentation
1911 for the FullMatch() function.
1912
1913 6. The escape sequence \n was being treated as whatever was defined as
1914 "newline". Not only was this contrary to the documentation, which states
1915 that \n is character 10 (hex 0A), but it also went horribly wrong when
1916 "newline" was defined as CRLF. This has been fixed.
1917
1918 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1919 was being set to -1 for the "end of line" case (supposedly a value that no
1920 character can have). Though this value is never used (the check for end of
1921 line is "zero bytes in current character"), it caused compiler complaints.
1922 I've changed it to 0xffffffff.
1923
1924 8. In pcre_version.c, the version string was being built by a sequence of
1925 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1926 string (as it is for production releases) called a macro with an empty
1927 argument. The C standard says the result of this is undefined. The gcc
1928 compiler treats it as an empty string (which was what was wanted) but it is
1929 reported that Visual C gives an error. The source has been hacked around to
1930 avoid this problem.
1931
1932 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1933 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1934 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1935 of them did).
1936
1937 10. Originally, pcretest opened its input and output without "b"; then I was
1938 told that "b" was needed in some environments, so it was added for release
1939 5.0 to both the input and output. (It makes no difference on Unix-like
1940 systems.) Later I was told that it is wrong for the input on Windows. I've
1941 now abstracted the modes into two macros, to make it easier to fiddle with
1942 them, and removed "b" from the input mode under Windows.
1943
1944 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1945
1946 12. Added -help and --help to pcretest as an official way of being reminded
1947 of the options.
1948
1949 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1950 and pcrecpp.cc because they annoy compilers at high warning levels.
1951
1952 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1953
1954 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1955 scripts are not C programs :-) and which was not noticed because it works
1956 on Linux.
1957
1958 16. pcretest is supposed to handle any length of pattern and data line (as one
1959 line or as a continued sequence of lines) by extending its input buffer if
1960 necessary. This feature was broken for very long pattern lines, leading to
1961 a string of junk being passed to pcre_compile() if the pattern was longer
1962 than about 50K.
1963
1964 17. I have done a major re-factoring of the way pcre_compile() computes the
1965 amount of memory needed for a compiled pattern. Previously, there was code
1966 that made a preliminary scan of the pattern in order to do this. That was
1967 OK when PCRE was new, but as the facilities have expanded, it has become
1968 harder and harder to keep it in step with the real compile phase, and there
1969 have been a number of bugs (see for example, 4 above). I have now found a
1970 cunning way of running the real compile function in a "fake" mode that
1971 enables it to compute how much memory it would need, while actually only
1972 ever using a few hundred bytes of working memory and without too many
1973 tests of the mode. This should make future maintenance and development
1974 easier. A side effect of this work is that the limit of 200 on the nesting
1975 depth of parentheses has been removed (though this was never a serious
1976 limitation, I suspect). However, there is a downside: pcre_compile() now
1977 runs more slowly than before (30% or more, depending on the pattern). I
1978 hope this isn't a big issue. There is no effect on runtime performance.
1979
1980 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1981 newline (only possible for the last line of a file) and it was a
1982 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1983
1984 19. Added additional timing features to pcretest. (1) The -tm option now times
1985 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1986 separate command line item, by a number that specifies the number of
1987 repeats to use when timing. The default is 50000; this gives better
1988 precision, but takes uncomfortably long for very large patterns.
1989
1990 20. Extended pcre_study() to be more clever in cases where a branch of a
1991 subpattern has no definite first character. For example, (a*|b*)[cd] would
1992 previously give no result from pcre_study(). Now it recognizes that the
1993 first character must be a, b, c, or d.
1994
1995 21. There was an incorrect error "recursive call could loop indefinitely" if
1996 a subpattern (or the entire pattern) that was being tested for matching an
1997 empty string contained only one non-empty item after a nested subpattern.
1998 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1999 incorrectly, because the \d was being skipped in the check.
2000
2001 22. The pcretest program now has a new pattern option /B and a command line
2002 option -b, which is equivalent to adding /B to every pattern. This causes
2003 it to show the compiled bytecode, without the additional information that
2004 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2005 is the same as /B/I).
2006
2007 23. A new optimization is now able automatically to treat some sequences such
2008 as a*b as a*+b. More specifically, if something simple (such as a character
2009 or a simple class like \d) has an unlimited quantifier, and is followed by
2010 something that cannot possibly match the quantified thing, the quantifier
2011 is automatically "possessified".
2012
2013 24. A recursive reference to a subpattern whose number was greater than 39
2014 went wrong under certain circumstances in UTF-8 mode. This bug could also
2015 have affected the operation of pcre_study().
2016
2017 25. Realized that a little bit of performance could be had by replacing
2018 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2019
2020 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2021
2022 27. Possessive quantifiers such as a++ were previously implemented by turning
2023 them into atomic groups such as ($>a+). Now they have their own opcodes,
2024 which improves performance. This includes the automatically created ones
2025 from 23 above.
2026
2027 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2028 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2029 the first matched character to be a colon. This applied both to named and
2030 numbered groups.
2031
2032 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2033
2034 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2035 building PCRE on Windows easier, so I have included it in the distribution.
2036
2037 31. There is now a check in pcretest against a ridiculously large number being
2038 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2039 loop, the loop is abandoned.
2040
2041 32. Forward references to subpatterns in conditions such as (?(2)...) where
2042 subpattern 2 is defined later cause pcre_compile() to search forwards in
2043 the pattern for the relevant set of parentheses. This search went wrong
2044 when there were unescaped parentheses in a character class, parentheses
2045 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2046
2047 33. "Subroutine" calls and backreferences were previously restricted to
2048 referencing subpatterns earlier in the regex. This restriction has now
2049 been removed.
2050
2051 34. Added a number of extra features that are going to be in Perl 5.10. On the
2052 whole, these are just syntactic alternatives for features that PCRE had
2053 previously implemented using the Python syntax or my own invention. The
2054 other formats are all retained for compatibility.
2055
2056 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2057 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2058 also .NET compatible.
2059
2060 (b) A recursion or subroutine call to a named group can now be defined as
2061 (?&name) as well as (?P>name).
2062
2063 (c) A backreference to a named group can now be defined as \k<name> or
2064 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2065 5.10, are also .NET compatible.
2066
2067 (d) A conditional reference to a named group can now use the syntax
2068 (?(<name>) or (?('name') as well as (?(name).
2069
2070 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2071 groups (named and numbered) that are never evaluated inline, but can be
2072 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2073 is always false. There may be only one alternative in such a group.
2074
2075 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2076 as the simple (?(R). The condition is true only if the most recent
2077 recursion is that of the given number or name. It does not search out
2078 through the entire recursion stack.
2079
2080 (g) The escape \gN or \g{N} has been added, where N is a positive or
2081 negative number, specifying an absolute or relative reference.
2082
2083 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2084 some "unreachable code" warnings.
2085
2086 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2087 things, this adds five new scripts.
2088
2089 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2090 There were also incompatibilities regarding the handling of \Q..\E inside
2091 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2092 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2093
2094 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2095 matches an empty string, and forcibly breaks the loop. There were bugs in
2096 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2097 against aaaa the result was just "a" rather than "aaaa", for example. Two
2098 separate and independent bugs (that affected different cases) have been
2099 fixed.
2100
2101 39. Refactored the code to abolish the use of different opcodes for small
2102 capturing bracket numbers. This is a tidy that I avoided doing when I
2103 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2104 The new approach is not only tidier, it makes it possible to reduce the
2105 memory needed to fix the previous bug (38).
2106
2107 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2108 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2109 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2110 mode.
2111
2112 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2113 report.
2114
2115 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2116 copy construction and assignment in the C++ wrapper.
2117
2118 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2119 couple of bugs that could have given wrong results in the "--newline=crlf"
2120 case.
2121
2122 44. Added a number of casts and did some reorganization of signed/unsigned int
2123 variables following suggestions from Dair Grant. Also renamed the variable
2124 "this" as "item" because it is a C++ keyword.
2125
2126 45. Arranged for dftables to add
2127
2128 #include "pcre_internal.h"
2129
2130 to pcre_chartables.c because without it, gcc 4.x may remove the array
2131 definition from the final binary if PCRE is built into a static library and
2132 dead code stripping is activated.
2133
2134 46. For an unanchored pattern, if a match attempt fails at the start of a
2135 newline sequence, and the newline setting is CRLF or ANY, and the next two
2136 characters are CRLF, advance by two characters instead of one.
2137
2138
2139 Version 6.7 04-Jul-06
2140 ---------------------
2141
2142 1. In order to handle tests when input lines are enormously long, pcretest has
2143 been re-factored so that it automatically extends its buffers when
2144 necessary. The code is crude, but this _is_ just a test program. The
2145 default size has been increased from 32K to 50K.
2146
2147 2. The code in pcre_study() was using the value of the re argument before
2148 testing it for NULL. (Of course, in any sensible call of the function, it
2149 won't be NULL.)
2150
2151 3. The memmove() emulation function in pcre_internal.h, which is used on
2152 systems that lack both memmove() and bcopy() - that is, hardly ever -
2153 was missing a "static" storage class specifier.
2154
2155 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2156 containing an extended class (one that cannot be represented by a bitmap
2157 because it contains high-valued characters or Unicode property items, e.g.
2158 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2159 pattern, but PCRE should not loop if you do not (it no longer does).
2160 [Detail: two cases were found: (a) a repeated subpattern containing an
2161 extended class; (b) a recursive reference to a subpattern that followed a
2162 previous extended class. It wasn't skipping over the extended class
2163 correctly when UTF-8 mode was not set.]
2164
2165 5. A negated single-character class was not being recognized as fixed-length
2166 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2167 compile error "lookbehind assertion is not fixed length".
2168
2169 6. The RunPerlTest auxiliary script was showing an unexpected difference
2170 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2171 write a Perl script that can interpret lines of an input file either as
2172 byte characters or as UTF-8, which is what "perltest" was being required to
2173 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2174 can't do is switch easily at run time between having the "use utf8;" pragma
2175 or not. In the end, I fudged it by using the RunPerlTest script to insert
2176 "use utf8;" explicitly for the UTF-8 tests.
2177
2178 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2179 the end of the subject string, contrary to the documentation and to what
2180 Perl does. This was true of both matching functions. Now it matches only at
2181 the start of the subject and immediately after *internal* newlines.
2182
2183 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2184 a pointer to an int instead of a pointer to an unsigned long int. This
2185 caused problems on 64-bit systems.
2186
2187 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2188 instance of the 'standard' template library not being so standard".
2189
2190 10. There was no check on the number of named subpatterns nor the maximum
2191 length of a subpattern name. The product of these values is used to compute
2192 the size of the memory block for a compiled pattern. By supplying a very
2193 long subpattern name and a large number of named subpatterns, the size
2194 computation could be caused to overflow. This is now prevented by limiting
2195 the length of names to 32 characters, and the number of named subpatterns
2196 to 10,000.
2197
2198 11. Subpatterns that are repeated with specific counts have to be replicated in
2199 the compiled pattern. The size of memory for this was computed from the
2200 length of the subpattern and the repeat count. The latter is limited to
2201 65535, but there was no limit on the former, meaning that integer overflow
2202 could in principle occur. The compiled length of a repeated subpattern is
2203 now limited to 30,000 bytes in order to prevent this.
2204
2205 12. Added the optional facility to have named substrings with the same name.
2206
2207 13. Added the ability to use a named substring as a condition, using the
2208 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2209 are numbers (not recommended). Forward references are permitted.
2210
2211 14. Added forward references in named backreferences (if you see what I mean).
2212
2213 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2214 pattern could run off the end of the subject. For example, the pattern
2215 "(?s)(.{1,5})"8 did this with the subject "ab".
2216
2217 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2218 PCRE_CASELESS was set when matching characters that were quantified with ?
2219 or *.
2220
2221 17. A character class other than a single negated character that had a minimum
2222 but no maximum quantifier - for example [ab]{6,} - was not handled
2223 correctly by pce_dfa_exec(). It would match only one character.
2224
2225 18. A valid (though odd) pattern that looked like a POSIX character
2226 class but used an invalid character after [ (for example [[,abc,]]) caused
2227 pcre_compile() to give the error "Failed: internal error: code overflow" or
2228 in some cases to crash with a glibc free() error. This could even happen if
2229 the pattern terminated after [[ but there just happened to be a sequence of
2230 letters, a binary zero, and a closing ] in the memory that followed.
2231
2232 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2233 over the years. Originally (before any Unicode support), just the bottom 8
2234 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2235 output from "man perlunicode" includes this:
2236
2237 The regular expression compiler produces polymorphic opcodes. That
2238 is, the pattern adapts to the data and automatically switches to
2239 the Unicode character scheme when presented with Unicode data--or
2240 instead uses a traditional byte scheme when presented with byte
2241 data.
2242
2243 Sadly, a wide octal escape does not cause a switch, and in a string with
2244 no other multibyte characters, these octal escapes are treated as before.
2245 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2246 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2247 Unicode string.
2248
2249 I have not perpetrated such confusion in PCRE. Up till now, it took just
2250 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2251 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2252 translate to the appropriate multibyte character.
2253
2254 29. Applied some refactoring to reduce the number of warnings from Microsoft
2255 and Borland compilers. This has included removing the fudge introduced
2256 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2257 a warning about an unused variable.
2258
2259 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2260 characters since release 4.0, because Perl (from release 5.004) does not.
2261 [Or at least, is documented not to: some releases seem to be in conflict
2262 with the documentation.] However, when a pattern was studied with
2263 pcre_study() and all its branches started with \s, PCRE still included VT
2264 as a possible starting character. Of course, this did no harm; it just
2265 caused an unnecessary match attempt.
2266
2267 22. Removed a now-redundant internal flag bit that recorded the fact that case
2268 dependency changed within the pattern. This was once needed for "required
2269 byte" processing, but is no longer used. This recovers a now-scarce options
2270 bit. Also moved the least significant internal flag bit to the most-
2271 significant bit of the word, which was not previously used (hangover from
2272 the days when it was an int rather than a uint) to free up another bit for
2273 the future.
2274
2275 23. Added support for CRLF line endings as well as CR and LF. As well as the
2276 default being selectable at build time, it can now be changed at runtime
2277 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2278 specify that it is scanning data with non-default line endings.
2279
2280 24. Changed the definition of CXXLINK to make it agree with the definition of
2281 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2282
2283 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2284 recursions. This makes a big different to stack usage for some patterns.
2285
2286 26. If a subpattern containing a named recursion or subroutine reference such
2287 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2288 the space required for the compiled pattern went wrong and gave too small a
2289 value. Depending on the environment, this could lead to "Failed: internal
2290 error: code overflow at offset 49" or "glibc detected double free or
2291 corruption" errors.
2292
2293 27. Applied patches from Google (a) to support the new newline modes and (b) to
2294 advance over multibyte UTF-8 characters in GlobalReplace.
2295
2296 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2297 difference for some implementation of PCRE in some Windows version.
2298
2299 29. Added some extra testing facilities to pcretest:
2300
2301 \q<number> in a data line sets the "match limit" value
2302 \Q<number> in a data line sets the "match recursion limt" value
2303 -S <number> sets the stack size, where <number> is in megabytes
2304
2305 The -S option isn't available for Windows.
2306
2307
2308 Version 6.6 06-Feb-06
2309 ---------------------
2310
2311 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2312 in pcreposix.h. I have copied the definition from pcre.h.
2313
2314 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2315 because pcre.h is no longer a built file.
2316
2317 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2318 not normally included in the compiled code.
2319
2320
2321 Version 6.5 01-Feb-06
2322 ---------------------
2323
2324 1. When using the partial match feature with pcre_dfa_exec(), it was not
2325 anchoring the second and subsequent partial matches at the new starting
2326 point. This could lead to incorrect results. For example, with the pattern
2327 /1234/, partially matching against "123" and then "a4" gave a match.
2328
2329 2. Changes to pcregrep:
2330
2331 (a) All non-match returns from pcre_exec() were being treated as failures
2332 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2333 error message is output. Some extra information is given for the
2334 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2335 probably the only errors that are likely to be caused by users (by
2336 specifying a regex that has nested indefinite repeats, for instance).
2337 If there are more than 20 of these errors, pcregrep is abandoned.
2338
2339 (b) A binary zero was treated as data while matching, but terminated the
2340 output line if it was written out. This has been fixed: binary zeroes
2341 are now no different to any other data bytes.
2342
2343 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2344 used to set a locale for matching. The --locale=xxxx long option has
2345 been added (no short equivalent) to specify a locale explicitly on the
2346 pcregrep command, overriding the environment variables.
2347
2348 (d) When -B was used with -n, some line numbers in the output were one less
2349 than they should have been.
2350
2351 (e) Added the -o (--only-matching) option.
2352
2353 (f) If -A or -C was used with -c (count only), some lines of context were
2354 accidentally printed for the final match.
2355
2356 (g) Added the -H (--with-filename) option.
2357
2358 (h) The combination of options -rh failed to suppress file names for files
2359 that were found from directory arguments.
2360
2361 (i) Added the -D (--devices) and -d (--directories) options.
2362
2363 (j) Added the -F (--fixed-strings) option.
2364
2365 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2366
2367 (l) Added the --colo(u)r option.
2368
2369 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2370 is not present by default.
2371
2372 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2373 items such as (?R) or (?1), when the recursion could match a number of
2374 alternatives. If it matched one of the alternatives, but subsequently,
2375 outside the recursion, there was a failure, the code tried to back up into
2376 the recursion. However, because of the way PCRE is implemented, this is not
2377 possible, and the result was an incorrect result from the match.
2378
2379 In order to prevent this happening, the specification of recursion has
2380 been changed so that all such subpatterns are automatically treated as
2381 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2382
2383 4. I had overlooked the fact that, in some locales, there are characters for
2384 which isalpha() is true but neither isupper() nor islower() are true. In
2385 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2386 and ordfeminine) are like this. This affected the treatment of \w and \W
2387 when they appeared in character classes, but not when they appeared outside
2388 a character class. The bit map for "word" characters is now created
2389 separately from the results of isalnum() instead of just taking it from the
2390 upper, lower, and digit maps. (Plus the underscore character, of course.)
2391
2392 5. The above bug also affected the handling of POSIX character classes such as
2393 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2394 permanent tables. Instead, the bit maps for such a class were previously
2395 created as the appropriate unions of the upper, lower, and digit bitmaps.
2396 Now they are created by subtraction from the [[:word:]] class, which has
2397 its own bitmap.
2398
2399 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2400 It is created by subtracting the vertical space characters (\x09, \x0a,
2401 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2402 subtraction was done in the overall bitmap for a character class, meaning
2403 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2404 be recognized. This bug has been fixed.
2405
2406 7. Patches from the folks at Google:
2407
2408 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2409 real life, but is still worth protecting against".
2410
2411 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2412 regular expressions".
2413
2414 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2415 have it.
2416
2417 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2418 "configure" and the latter not, in order to fix a problem somebody had
2419 with compiling the Arg class on HP-UX.
2420
2421 (e) Improve the error-handling of the C++ wrapper a little bit.
2422
2423 (f) New tests for checking recursion limiting.
2424
2425 8. The pcre_memmove() function, which is used only if the environment does not
2426 have a standard memmove() function (and is therefore rarely compiled),
2427 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2428 returning a result (though PCRE never actually uses the result).
2429
2430 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2431 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2432 returned instead of calling malloc() with an overflowing number that would
2433 most likely cause subsequent chaos.
2434
2435 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2436
2437 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2438 with this option is matched, the nmatch and pmatch options of regexec() are
2439 ignored.
2440
2441 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2442 provided in case anyone wants to the the POSIX interface with UTF-8
2443 strings.
2444
2445 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2446 C++ linking (needed for some HP-UX environments).
2447
2448 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2449 (unused parameter) and in the pcre_printint() function (omitted "default"
2450 switch label when the default is to do nothing).
2451
2452 15. Added some code to make it possible, when PCRE is compiled as a C++
2453 library, to replace subject pointers for pcre_exec() with a smart pointer
2454 class, thus making it possible to process discontinuous strings.
2455
2456 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2457 much the same function. They were added by different people who were trying
2458 to make PCRE easy to compile on non-Unix systems. It has been suggested
2459 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2460 for compiling on Windows systems. I have therefore replaced it with
2461 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2462 defaults to "extern" for C or "extern C" for C++, which works fine on
2463 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2464 SCOPE with something explicit in config.h. In addition:
2465
2466 (a) pcreposix.h still had just "extern" instead of either of these macros;
2467 I have replaced it with PCRE_DATA_SCOPE.
2468
2469 (b) Functions such as _pcre_xclass(), which are internal to the library,
2470 but external in the C sense, all had PCRE_EXPORT in their definitions.
2471 This is apparently wrong for the Windows case, so I have removed it.
2472 (It makes no difference on Unix-like systems.)
2473
2474 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2475 of recursive calls to match(). This is different to MATCH_LIMIT because
2476 that limits the total number of calls to match(), not all of which increase
2477 the depth of recursion. Limiting the recursion depth limits the amount of
2478 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2479 when PCRE is compiled, and changed at run time. A patch from Google adds
2480 this functionality to the C++ interface.
2481
2482 18. Changes to the handling of Unicode character properties:
2483
2484 (a) Updated the table to Unicode 4.1.0.
2485
2486 (b) Recognize characters that are not in the table as "Cn" (undefined).
2487
2488 (c) I revised the way the table is implemented to a much improved format
2489 which includes recognition of ranges. It now supports the ranges that
2490 are defined in UnicodeData.txt, and it also amalgamates other
2491 characters into ranges. This has reduced the number of entries in the
2492 table from around 16,000 to around 3,000, thus reducing its size
2493 considerably. I realized I did not need to use a tree structure after
2494 all - a binary chop search is just as efficient. Having reduced the
2495 number of entries, I extended their size from 6 bytes to 8 bytes to
2496 allow for more data.
2497
2498 (d) Added support for Unicode script names via properties such as \p{Han}.
2499
2500 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2501 matching that character.
2502
2503 20. When matching a repeated Unicode property with a minimum greater than zero,
2504 (for example \pL{2,}), PCRE could look past the end of the subject if it
2505 reached it while seeking the minimum number of characters. This could
2506 happen only if some of the characters were more than one byte long, because
2507 there is a check for at least the minimum number of bytes.
2508
2509 21. Refactored the implementation of \p and \P so as to be more general, to
2510 allow for more different types of property in future. This has changed the
2511 compiled form incompatibly. Anybody with saved compiled patterns that use
2512 \p or \P will have to recompile them.
2513
2514 22. Added "Any" and "L&" to the supported property types.
2515
2516 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2517 but give a compile time error if the value is greater than 0xff.
2518
2519 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2520 accidentally not being installed or uninstalled.
2521
2522 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2523 made were to insert the current release number. This seemed silly, because
2524 it made things harder for people building PCRE on systems that don't run
2525 "configure". I have turned pcre.h into a distributed file, no longer built
2526 by "configure", with the version identification directly included. There is
2527 no longer a pcre.h.in file.
2528
2529 However, this change necessitated a change to the pcre-config script as
2530 well. It is built from pcre-config.in, and one of the substitutions was the
2531 release number. I have updated configure.ac so that ./configure now finds
2532 the release number by grepping pcre.h.
2533
2534 26. Added the ability to run the tests under valgrind.
2535
2536
2537 Version 6.4 05-Sep-05
2538 ---------------------
2539
2540 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2541 "--" to be printed when multiple files were scanned, even when none of the
2542 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2543 consider it to be a bug, and have restored the previous behaviour.
2544
2545 2. A couple of code tidies to get rid of compiler warnings.
2546
2547 3. The pcretest program used to cheat by referring to symbols in the library
2548 whose names begin with _pcre_. These are internal symbols that are not
2549 really supposed to be visible externally, and in some environments it is
2550 possible to suppress them. The cheating is now confined to including
2551 certain files from the library's source, which is a bit cleaner.
2552
2553 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2554 file's purpose clearer.
2555
2556 5. Reorganized pcre_ucp_findchar().
2557
2558
2559 Version 6.3 15-Aug-05
2560 ---------------------
2561
2562 1. The file libpcre.pc.in did not have general read permission in the tarball.
2563
2564 2. There were some problems when building without C++ support:
2565
2566 (a) If C++ support was not built, "make install" and "make test" still
2567 tried to test it.
2568
2569 (b) There were problems when the value of CXX was explicitly set. Some
2570 changes have been made to try to fix these, and ...
2571
2572 (c) --disable-cpp can now be used to explicitly disable C++ support.
2573
2574 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2575 backslash in a target when C++ was disabled. This confuses some
2576 versions of "make", apparently. Using an intermediate variable solves
2577 this. (Same for CPP_LOBJ.)
2578
2579 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2580 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2581 necessary on certain architectures.
2582
2583 4. Added a setting of -export-symbols-regex to the link command to remove
2584 those symbols that are exported in the C sense, but actually are local
2585 within the library, and not documented. Their names all begin with
2586 "_pcre_". This is not a perfect job, because (a) we have to except some
2587 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2588 available (and never for static libraries). I have made a note to try to
2589 find a way round (a) in the future.
2590
2591
2592 Version 6.2 01-Aug-05
2593 ---------------------
2594
2595 1. There was no test for integer overflow of quantifier values. A construction
2596 such as {1111111111111111} would give undefined results. What is worse, if
2597 a minimum quantifier for a parenthesized subpattern overflowed and became
2598 negative, the calculation of the memory size went wrong. This could have
2599 led to memory overwriting.
2600
2601 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2602
2603 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2604 operating environments where this matters.
2605
2606 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2607 PCRE options from within the C++ wrapper.
2608
2609 5. Named capturing subpatterns were not being correctly counted when a pattern
2610 was compiled. This caused two problems: (a) If there were more than 100
2611 such subpatterns, the calculation of the memory needed for the whole
2612 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2613 back references of the form \12, where the number was greater than 9, were
2614 not recognized as back references, even though there were sufficient
2615 previous subpatterns.
2616
2617 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2618 versions of gcc, e.g. 2.95.4.
2619
2620
2621 Version 6.1 21-Jun-05
2622 ---------------------
2623
2624 1. There was one reference to the variable "posix" in pcretest.c that was not
2625 surrounded by "#if !defined NOPOSIX".
2626
2627 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2628 the cross-check on the old pcre_info() function, for the benefit of the
2629 cut-down version of PCRE that is currently imported into Exim.
2630
2631 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2632 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2633 patterns that start (?i) but I don't think that matters. The use of (?i) is
2634 just an example; this all applies to the other options as well.
2635
2636 4. Since libtool seems to echo the compile commands it is issuing, the output
2637 from "make" can be reduced a bit by putting "@" in front of each libtool
2638 compile command.
2639
2640 5. Patch from the folks at Google for configure.in to be a bit more thorough
2641 in checking for a suitable C++ installation before trying to compile the
2642 C++ stuff. This should fix a reported problem when a compiler was present,
2643 but no suitable headers.
2644
2645 6. The man pages all had just "PCRE" as their title. I have changed them to
2646 be the relevant file name. I have also arranged that these names are
2647 retained in the file doc/pcre.txt, which is a concatenation in text format
2648 of all the man pages except the little individual ones for each function.
2649
2650 7. The NON-UNIX-USE file had not been updated for the different set of source
2651 files that come with release 6. I also added a few comments about the C++
2652 wrapper.
2653
2654
2655 Version 6.0 07-Jun-05
2656 ---------------------
2657
2658 1. Some minor internal re-organization to help with my DFA experiments.
2659
2660 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2661 didn't matter for the library itself when fully configured, but did matter
2662 when compiling without UCP support, or within Exim, where the ucp files are
2663 not imported.
2664
2665 3. Refactoring of the library code to split up the various functions into
2666 different source modules. The addition of the new DFA matching code (see
2667 below) to a single monolithic source would have made it really too
2668 unwieldy, quite apart from causing all the code to be include in a
2669 statically linked application, when only some functions are used. This is
2670 relevant even without the DFA addition now that patterns can be compiled in
2671 one application and matched in another.
2672
2673 The downside of splitting up is that there have to be some external
2674 functions and data tables that are used internally in different modules of
2675 the library but which are not part of the API. These have all had their
2676 names changed to start with "_pcre_" so that they are unlikely to clash
2677 with other external names.
2678
2679 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2680 a different (DFA) algorithm. Although it is slower than the original
2681 function, it does have some advantages for certain types of matching
2682 problem.
2683
2684 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2685 including restarting after a partial match.
2686
2687 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2688 defined when compiling for Windows was sent to me. I have put it into the
2689 code, though I have no means of testing or verifying it.
2690
2691 7. Added the pcre_refcount() auxiliary function.
2692
2693 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2694 match before or at the first newline in the subject string. In pcretest,
2695 the /f option on a pattern can be used to set this.
2696
2697 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2698 would behave wrongly. This has been present in PCRE since release 4.0.
2699
2700 10. A number of changes to the pcregrep command:
2701
2702 (a) Refactored how -x works; insert ^(...)$ instead of setting
2703 PCRE_ANCHORED and checking the length, in preparation for adding
2704 something similar for -w.
2705
2706 (b) Added the -w (match as a word) option.
2707
2708 (c) Refactored the way lines are read and buffered so as to have more
2709 than one at a time available.
2710
2711 (d) Implemented a pcregrep test script.
2712
2713 (e) Added the -M (multiline match) option. This allows patterns to match
2714 over several lines of the subject. The buffering ensures that at least
2715 8K, or the rest of the document (whichever is the shorter) is available
2716 for matching (and similarly the previous 8K for lookbehind assertions).
2717
2718 (f) Changed the --help output so that it now says
2719
2720 -w, --word-regex(p)
2721
2722 instead of two lines, one with "regex" and the other with "regexp"
2723 because that confused at least one person since the short forms are the
2724 same. (This required a bit of code, as the output is generated
2725 automatically from a table. It wasn't just a text change.)
2726
2727 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2728 option but starts with a hyphen. Could be a pattern or a path name
2729 starting with a hyphen, for instance.
2730
2731 (h) "-" can be given as a file name to represent stdin.
2732
2733 (i) When file names are being printed, "(standard input)" is used for
2734 the standard input, for compatibility with GNU grep. Previously
2735 "<stdin>" was used.
2736
2737 (j) The option --label=xxx can be used to supply a name to be used for
2738 stdin when file names are being printed. There is no short form.
2739
2740 (k) Re-factored the options decoding logic because we are going to add
2741 two more options that take data. Such options can now be given in four
2742 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2743
2744 (l) Added the -A, -B, and -C options for requesting that lines of context
2745 around matches be printed.
2746
2747 (m) Added the -L option to print the names of files that do not contain
2748 any matching lines, that is, the complement of -l.
2749
2750 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2751 continue to scan other files.
2752
2753 (o) The -s option was incorrectly implemented. For compatibility with other
2754 greps, it now suppresses the error message for a non-existent or non-
2755 accessible file (but not the return code). There is a new option called
2756 -q that suppresses the output of matching lines, which was what -s was
2757 previously doing.
2758
2759 (p) Added --include and --exclude options to specify files for inclusion
2760 and exclusion when recursing.
2761
2762 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2763 Hopefully, it now does.
2764
2765 12. Missing cast in pcre_study().
2766
2767 13. Added an "uninstall" target to the makefile.
2768
2769 14. Replaced "extern" in the function prototypes in Makefile.in with
2770 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2771 world, but is set differently for Windows.
2772
2773 15. Added a second compiling function called pcre_compile2(). The only
2774 difference is that it has an extra argument, which is a pointer to an
2775 integer error code. When there is a compile-time failure, this is set
2776 non-zero, in addition to the error test pointer being set to point to an
2777 error message. The new argument may be NULL if no error number is required
2778 (but then you may as well call pcre_compile(), which is now just a
2779 wrapper). This facility is provided because some applications need a
2780 numeric error indication, but it has also enabled me to tidy up the way
2781 compile-time errors are handled in the POSIX wrapper.
2782
2783 16. Added VPATH=.libs to the makefile; this should help when building with one
2784 prefix path and installing with another. (Or so I'm told by someone who
2785 knows more about this stuff than I do.)
2786
2787 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2788 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2789 match everything, including newlines. This is not POSIX-compatible, but
2790 somebody wanted the feature. From pcretest it can be activated by using
2791 both the P and the s flags.
2792
2793 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2794
2795 19. libpcre.pc was being incorrectly installed as executable.
2796
2797 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2798 it now also looks for '\r' so that it will work unmodified on Windows.
2799
2800 21. Added Google's contributed C++ wrapper to the distribution.
2801
2802 22. Added some untidy missing memory free() calls in pcretest, to keep
2803 Electric Fence happy when testing.
2804
2805
2806
2807 Version 5.0 13-Sep-04
2808 ---------------------
2809
2810 1. Internal change: literal characters are no longer packed up into items
2811 containing multiple characters in a single byte-string. Each character
2812 is now matched using a separate opcode. However, there may be more than one
2813 byte in the character in UTF-8 mode.
2814
2815 2. The pcre_callout_block structure has two new fields: pattern_position and
2816 next_item_length. These contain the offset in the pattern to the next match
2817 item, and its length, respectively.
2818
2819 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2820 insertion of callouts before each pattern item. Added the /C option to
2821 pcretest to make use of this.
2822
2823 4. On the advice of a Windows user, the lines
2824
2825 #if defined(_WIN32) || defined(WIN32)
2826 _setmode( _fileno( stdout ), 0x8000 );
2827 #endif /* defined(_WIN32) || defined(WIN32) */
2828
2829 have been added to the source of pcretest. This apparently does useful
2830 magic in relation to line terminators.
2831
2832 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2833 for the benefit of those environments where the "b" makes a difference.
2834
2835 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2836 to know about it. I have put a hack into configure.in that adds in code
2837 to set GCC=yes if CC=icc. This seems to end up at a point in the
2838 generated configure script that is early enough to affect the setting of
2839 compiler options, which is what is needed, but I have no means of testing
2840 whether it really works. (The user who reported this had patched the
2841 generated configure script, which of course I cannot do.)
2842
2843 LATER: After change 22 below (new libtool files), the configure script
2844 seems to know about icc (and also ecc). Therefore, I have commented out
2845 this hack in configure.in.
2846
2847 7. Added support for pkg-config (2 patches were sent in).
2848
2849 8. Negated POSIX character classes that used a combination of internal tables
2850 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2851 [[:^ascii]]. Typically, they would match almost any characters. The other
2852 POSIX classes were not broken in this way.
2853
2854 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2855 to find the match, as PCRE was deluded into thinking that the match had to
2856 start at the start point or following a newline. The same bug applied to
2857 patterns with negative forward assertions or any backward assertions
2858 preceding ".*" at the start, unless the pattern required a fixed first
2859 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2860
2861 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2862 starting at the last subject character, bytes beyond the end of the subject
2863 string were read.
2864
2865 11. Renamed the variable "class" as "classbits" to make life easier for C++
2866 users. (Previously there was a macro definition, but it apparently wasn't
2867 enough.)
2868
2869 12. Added the new field "tables" to the extra data so that tables can be passed
2870 in at exec time, or the internal tables can be re-selected. This allows
2871 a compiled regex to be saved and re-used at a later time by a different
2872 program that might have everything at different addresses.
2873
2874 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2875 -R library as well as a -L library.
2876
2877 14. The debugging options of pcretest (-d on the command line or D on a
2878 pattern) showed incorrect output for anything following an extended class
2879 that contained multibyte characters and which was followed by a quantifier.
2880
2881 15. Added optional support for general category Unicode character properties
2882 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2883 support. It adds about 90K to the size of the library. The meanings of the
2884 inbuilt class escapes such as \d and \s have NOT been changed.
2885
2886 16. Updated pcredemo.c to include calls to free() to release the memory for the
2887 compiled pattern.
2888
2889 17. The generated file chartables.c was being created in the source directory
2890 instead of in the building directory. This caused the build to fail if the
2891 source directory was different from the building directory, and was
2892 read-only.
2893
2894 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2895 file. No doubt somebody will tell me if they don't make sense... Also added
2896 Dan Mooney's comments about building on OpenVMS.
2897
2898 19. Added support for partial matching via the PCRE_PARTIAL option for
2899 pcre_exec() and the \P data escape in pcretest.
2900
2901 20. Extended pcretest with 3 new pattern features:
2902
2903 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2904 write the compiled pattern to the file whose name is "rest-of-line".
2905 This is a straight binary dump of the data, with the saved pointer to
2906 the character tables forced to be NULL. The study data, if any, is
2907 written too. After writing, pcretest reads a new pattern.
2908
2909 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2910 compiled pattern from the given file. There must not be any
2911 occurrences of "<" in the file name (pretty unlikely); if there are,
2912 pcretest will instead treat the initial "<" as a pattern delimiter.
2913 After reading in the pattern, pcretest goes on to read data lines as
2914 usual.
2915
2916 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2917 and 16-bit fields in a compiled pattern, to simulate a pattern that
2918 was compiled on a host of opposite endianness.
2919
2920 21. The pcre-exec() function can now cope with patterns that were compiled on
2921 hosts of opposite endianness, with this restriction:
2922
2923 As for any compiled expression that is saved and used later, the tables
2924 pointer field cannot be preserved; the extra_data field in the arguments
2925 to pcre_exec() should be used to pass in a tables address if a value
2926 other than the default internal tables were used at compile time.
2927
2928 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2929 now diagnosed as an error. Previously, most of the time, a negative number
2930 would have been treated as zero, but if in addition "ovector" was passed as
2931 NULL, a crash could occur.
2932
2933 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2934 new versions from the libtool 1.5 distribution (the last one is a copy of
2935 a file called libtool.m4). This seems to have fixed the need to patch
2936 "configure" to support Darwin 1.3 (which I used to do). However, I still
2937 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2938 workstation).
2939
2940 24. Changed the PCRE licence to be the more standard "BSD" licence.
2941
2942
2943 Version 4.5 01-Dec-03
2944 ---------------------
2945
2946 1. There has been some re-arrangement of the code for the match() function so
2947 that it can be compiled in a version that does not call itself recursively.
2948 Instead, it keeps those local variables that need separate instances for
2949 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2950 needs to "recurse". Keeping track of where control must go is done by means
2951 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2952 hide most of the details from the main code, and operates only if
2953 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2954 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2955 operating.
2956
2957 To make it easier for callers to provide specially tailored get/free
2958 functions for this usage, two new functions, pcre_stack_malloc, and
2959 pcre_stack_free, are used. They are always called in strict stacking order,
2960 and the size of block requested is always the same.
2961
2962 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2963 PCRE has been compiled to use the stack or the heap for recursion. The
2964 -C option of pcretest uses this to show which version is compiled.
2965
2966 A new data escape \S, is added to pcretest; it causes the amounts of store
2967 obtained and freed by both kinds of malloc/free at match time to be added
2968 to the output.
2969
2970 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2971 what's available on my current Linux desktop machine.
2972
2973 3. When matching a UTF-8 string, the test for a valid string at the start has
2974 been extended. If start_offset is not zero, PCRE now checks that it points
2975 to a byte that is the start of a UTF-8 character. If not, it returns
2976 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2977 this is necessary because there may be backward assertions in the pattern.
2978 When matching the same subject several times, it may save resources to use
2979 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2980
2981 4. The code for checking the validity of UTF-8 strings has been tightened so
2982 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2983 containing "overlong sequences".
2984
2985 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2986 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2987 should have been "&", but it just so happened that all the cases this let
2988 through by mistake were picked up later in the function.
2989
2990 6. I had used a variable called "isblank" - this is a C99 function, causing
2991 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2992
2993 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2994 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2995 so the version is shown; (c) stop "make test" echoing "./RunTest".
2996
2997 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2998
2999 9. The prototype for memmove() for systems that don't have it was using
3000 size_t, but the inclusion of the header that defines size_t was later. I've
3001 moved the #includes for the C headers earlier to avoid this.
3002
3003 10. Added some adjustments to the code to make it easier to compiler on certain
3004 special systems:
3005
3006 (a) Some "const" qualifiers were missing.
3007 (b) Added the macro EXPORT before all exported functions; by default this
3008 is defined to be empty.
3009 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3010 that it reads its output file name as an argument instead of writing
3011 to the standard output and assuming this can be redirected.
3012
3013 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3014 class containing characters with values greater than 255, PCRE compilation
3015 went into a loop.
3016
3017 12. A recursive reference to a subpattern that was within another subpattern
3018 that had a minimum quantifier of zero caused PCRE to crash. For example,
3019 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3020 recursion. If the recursively-called subpattern itself had a zero repeat,
3021 that was OK.
3022
3023 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3024 buffer into which it was copied (for escape processing) was still set at
3025 1024, so long lines caused crashes.
3026
3027 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3028 "internal error: code overflow...". This applied to any character class
3029 that was followed by a possessive quantifier.
3030
3031 15. Modified the Makefile to add libpcre.la as a prerequisite for
3032 libpcreposix.la because I was told this is needed for a parallel build to
3033 work.
3034
3035 16. If a pattern that contained .* following optional items at the start was
3036 studied, the wrong optimizing data was generated, leading to matching
3037 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3038 matching string must start with a or b or c. The correct conclusion for
3039 this pattern is that a match can start with any character.
3040
3041
3042 Version 4.4 13-Aug-03
3043 ---------------------
3044
3045 1. In UTF-8 mode, a character class containing characters with values between
3046 127 and 255 was not handled correctly if the compiled pattern was studied.
3047 In fixing this, I have also improved the studying algorithm for such
3048 classes (slightly).
3049
3050 2. Three internal functions had redundant arguments passed to them. Removal
3051 might give a very teeny performance improvement.
3052
3053 3. Documentation bug: the value of the capture_top field in a callout is *one
3054 more than* the number of the hightest numbered captured substring.
3055
3056 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3057 in incorrectly linking with a previously installed version. They now link
3058 explicitly with libpcre.la.
3059
3060 5. configure.in no longer needs to recognize Cygwin specially.
3061
3062 6. A problem in pcre.in for Windows platforms is fixed.
3063
3064 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3065 pcretest, it used to include the size of the study block as part of its
3066 output. Unfortunately, the structure contains a field that has a different
3067 size on different hardware architectures. This meant that the tests that
3068 showed this size failed. As the block is currently always of a fixed size,
3069 this information isn't actually particularly useful in pcretest output, so
3070 I have just removed it.
3071
3072 8. Three pre-processor statements accidentally did not start in column 1.
3073 Sadly, there are *still* compilers around that complain, even though
3074 standard C has not required this for well over a decade. Sigh.
3075
3076 9. In pcretest, the code for checking callouts passed small integers in the
3077 callout_data field, which is a void * field. However, some picky compilers
3078 complained about the casts involved for this on 64-bit systems. Now
3079 pcretest passes the address of the small integer instead, which should get
3080 rid of the warnings.
3081
3082 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3083 both compile and run time, and gives an error if an invalid UTF-8 sequence
3084 is found. There is a option for disabling this check in cases where the
3085 string is known to be correct and/or the maximum performance is wanted.
3086
3087 11. In response to a bug report, I changed one line in Makefile.in from
3088
3089 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3090 to
3091 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3092
3093 to look similar to other lines, but I have no way of telling whether this
3094 is the right thing to do, as I do not use Windows. No doubt I'll get told
3095 if it's wrong...
3096
3097
3098 Version 4.3 21-May-03
3099 ---------------------
3100
3101 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3102 Makefile.
3103
3104 2. Some refactoring to improve the quality of the code:
3105
3106 (i) The utf8_table... variables are now declared "const".
3107
3108 (ii) The code for \cx, which used the "case flipping" table to upper case
3109 lower case letters, now just substracts 32. This is ASCII-specific,
3110 but the whole concept of \cx is ASCII-specific, so it seems
3111 reasonable.
3112
3113 (iii) PCRE was using its character types table to recognize decimal and
3114 hexadecimal digits in the pattern. This is silly, because it handles
3115 only 0-9, a-f, and A-F, but the character types table is locale-
3116 specific, which means strange things might happen. A private
3117 table is now used for this - though it costs 256 bytes, a table is
3118 much faster than multiple explicit tests. Of course, the standard
3119 character types table is still used for matching digits in subject
3120 strings against \d.
3121
3122 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3123 ending in _t are). So I've renamed it as ESC_tee.
3124
3125 3. The first argument for regexec() in the POSIX wrapper should have been
3126 defined as "const".
3127
3128 4. Changed pcretest to use malloc() for its buffers so that they can be
3129 Electric Fenced for debugging.
3130
3131 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3132 to read one or more bytes before the start of the subject string. Often this
3133 had no effect on PCRE's behaviour, but in some circumstances it could
3134 provoke a segmentation fault.
3135
3136 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3137 to try to read one or more bytes before the start of the subject string.
3138
3139 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3140 UTF-8 support could misbehave in various ways if the subject string
3141 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3142 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3143 back over UTF-8 characters.)
3144
3145
3146 Version 4.2 14-Apr-03
3147 ---------------------
3148
3149 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3150
3151 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3152 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3153 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3154 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3155 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3156 and BUILD_EXEEXT
3157 Note: automatic setting of the BUILD variables is not yet working
3158 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3159 compile-time but not at link-time
3160 [LINK]: use for linking executables only
3161 make different versions for Windows and non-Windows
3162 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3163 libraries
3164 [LINK_FOR_BUILD]: new variable
3165 [OBJEXT]: use throughout
3166 [EXEEXT]: use throughout
3167 <winshared>: new target
3168 <wininstall>: new target
3169 <dftables.o>: use native compiler
3170 <dftables>: use native linker
3171 <install>: handle Windows platform correctly
3172 <clean>: ditto
3173 <check>: ditto
3174 copy DLL to top builddir before testing
3175
3176 As part of these changes, -no-undefined was removed again. This was reported
3177 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3178 in any case.
3179
3180 3. Some tidies to get rid of compiler warnings:
3181
3182 . In the match_data structure, match_limit was an unsigned long int, whereas
3183 match_call_count was an int. I've made them both unsigned long ints.
3184
3185 . In pcretest the fact that a const uschar * doesn't automatically cast to
3186 a void * provoked a warning.
3187
3188 . Turning on some more compiler warnings threw up some "shadow" variables
3189 and a few more missing casts.
3190
3191 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3192 option, a class that contained a single character with a value between 128
3193 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3194
3195 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3196 option, a class that contained several characters, but with at least one
3197 whose value was between 128 and 255 caused PCRE to crash.
3198
3199
3200 Version 4.1 12-Mar-03
3201 ---------------------
3202
3203 1. Compiling with gcc -pedantic found a couple of places where casts were
3204 needed, and a string in dftables.c that was longer than standard compilers are
3205 required to support.
3206
3207 2. Compiling with Sun's compiler found a few more places where the code could
3208 be tidied up in order to avoid warnings.
3209
3210 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3211 first of these names is deprecated in the latest Autoconf in favour of the name
3212 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3213 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3214 analogy I have changed it to CFLAGS_FOR_BUILD.
3215
3216 4. Added -no-undefined to the linking command in the Makefile, because this is
3217 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3218 linking step for the pcreposix library.
3219
3220 5. PCRE was failing to diagnose the case of two named groups with the same
3221 name.
3222
3223 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3224 literal character that is needed in the subject for a match, and scans along to
3225 ensure that it is present before embarking on the full matching process. This
3226 saves time in cases of nested unlimited repeats that are never going to match.
3227 Problem: the scan can take a lot of time if the subject is very long (e.g.
3228 megabytes), thus penalizing straightforward matches. It is now done only if the
3229 amount of subject to be scanned is less than 1000 bytes.
3230
3231 7. A lesser problem with the same optimization is that it was recording the
3232 first character of an anchored pattern as "needed", thus provoking a search
3233 right along the subject, even when the first match of the pattern was going to
3234 fail. The "needed" character is now not set for anchored patterns, unless it
3235 follows something in the pattern that is of non-fixed length. Thus, it still
3236 fulfils its original purpose of finding quick non-matches in cases of nested
3237 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3238
3239
3240 Version 4.0 17-Feb-03
3241 ---------------------
3242
3243 1. If a comment in an extended regex that started immediately after a meta-item
3244 extended to the end of string, PCRE compiled incorrect data. This could lead to
3245 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3246
3247 2. Moved to autoconf 2.53 and libtool 1.4.2.
3248
3249 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3250 the special perltest8 script is no longer needed - all the tests can be run
3251 from a single perltest script.
3252
3253 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3254 by \s. It has now been removed in PCRE. This means it isn't recognized as
3255 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3256 class [:space:] *does* include VT, thereby creating a mess.
3257
3258 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3259 space and tab.
3260
3261 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3262 its new features into the main test script, reducing the number of scripts.
3263
3264 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3265 were backward compatible, and made the (?i) apply to the whole pattern, as if
3266 /i were given. Now it behaves more logically, and applies the option setting
3267 only to what follows. PCRE has been changed to follow suit. However, if it
3268 finds options settings right at the start of the pattern, it extracts them into
3269 the global options, as before. Thus, they show up in the info data.
3270
3271 8. Added support for the \Q...\E escape sequence. Characters in between are
3272 treated as literals. This is slightly different from Perl in that $ and @ are
3273 also handled as literals inside the quotes. In Perl, they will cause variable
3274 interpolation. Note the following examples:
3275
3276 Pattern PCRE matches Perl matches
3277
3278 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3279 \Qabc\$xyz\E abc\$xyz abc\$xyz
3280 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3281
3282 For compatibility with Perl, \Q...\E sequences are recognized inside character
3283 classes as well as outside them.
3284
3285 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3286 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3287 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3288 signed/unsigned warnings.
3289
3290 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3291 option for pcretest, so I've replaced it by a simple function that does just
3292 that job.
3293
3294 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3295 "pcregrep -".
3296
3297 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3298 Java package. This provides some syntactic sugar for simple cases of what my
3299 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3300 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3301 item, you can use this simplified notation. Note that only makes sense with
3302 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3303 greediness, whatever the setting of the PCRE_UNGREEDY option.
3304
3305 13. A change of greediness default within a pattern was not taking effect at
3306 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3307 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3308 was abstracted outside.
3309
3310 14. PCRE now supports the \G assertion. It is true when the current matching
3311 position is at the start point of the match. This differs from \A when the
3312 starting offset is non-zero. Used with the /g option of pcretest (or similar
3313 code), it works in the same way as it does for Perl's /g option. If all
3314 alternatives of a regex begin with \G, the expression is anchored to the start
3315 match position, and the "anchored" flag is set in the compiled expression.
3316
3317 15. Some bugs concerning the handling of certain option changes within patterns
3318 have been fixed. These applied to options other than (?ims). For example,
3319 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3320 been the other way round. Some of this was related to change 7 above.
3321
3322 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3323 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3324 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3325 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3326
3327 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3328 mode. Unlike ".", it always matches newline, whatever the setting of
3329 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3330 assertions. Perl allows it, but it doesn't (in general) work because it can't
3331 calculate the length of the lookbehind. At least, that's the case for Perl
3332 5.8.0 - I've been told they are going to document that it doesn't work in
3333 future.
3334
3335 18. Added an error diagnosis for escapes that PCRE does not support: these are
3336 \L, \l, \N, \P, \p, \U, \u, and \X.
3337
3338 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3339 reading past the end of the pattern in cases such as /[abcd/.
3340
3341 20. PCRE was getting more memory than necessary for patterns with classes that
3342 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3343
3344 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3345 compiling PCRE for use with Virtual Pascal.
3346
3347 22. Small fix to the Makefile to make it work properly if the build is done
3348 outside the source tree.
3349
3350 23. Added a new extension: a condition to go with recursion. If a conditional
3351 subpattern starts with (?(R) the "true" branch is used if recursion has
3352 happened, whereas the "false" branch is used only at the top level.
3353
3354 24. When there was a very long string of literal characters (over 255 bytes
3355 without UTF support, over 250 bytes with UTF support), the computation of how
3356 much memory was required could be incorrect, leading to segfaults or other
3357 strange effects.
3358
3359 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3360 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3361 there was a subsequent back reference to those brackets. This meant that, for
3362 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3363 possible to check for precisely this case. All we can do is abandon the
3364 optimization if .* occurs inside capturing brackets when there are any back
3365 references whatsoever. (See below for a better fix that came later.)
3366
3367 26. The handling of the optimization for finding the first character of a
3368 non-anchored pattern, and for finding a character that is required later in the
3369 match were failing in some cases. This didn't break the matching; it just
3370 failed to optimize when it could. The way this is done has been re-implemented.
3371
3372 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3373
3374 28. Added a new feature that provides some of the functionality that Perl
3375 provides with (?{...}). The facility is termed a "callout". The way it is done
3376 in PCRE is for the caller to provide an optional function, by setting
3377 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3378 global variable. By default it is unset, which disables all calling out. To get
3379 the function called, the regex must include (?C) at appropriate points. This
3380 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3381 This provides a means of identifying different callout points. When PCRE
3382 reaches such a point in the regex, if pcre_callout has been set, the external
3383 function is called. It is provided with data in a structure called
3384 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3385 matching continues; if it returns a non-zero value, the match at the current
3386 point fails. However, backtracking will occur if possible. [This was changed
3387 later and other features added - see item 49 below.]
3388
3389 29. pcretest is upgraded to test the callout functionality. It provides a
3390 callout function that displays information. By default, it shows the start of
3391 the match and the current position in the text. There are some new data escapes
3392 to vary what happens:
3393
3394 \C+ in addition, show current contents of captured substrings
3395 \C- do not supply a callout function
3396 \C!n return 1 when callout number n is reached
3397 \C!n!m return 1 when callout number n is reached for the mth time
3398
3399 30. If pcregrep was called with the -l option and just a single file name, it
3400 output "<stdin>" if a match was found, instead of the file name.
3401
3402 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3403 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3404 pcre_exec(). This saves a malloc/free per call. The default value of
3405 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3406 when configuring.
3407
3408 32. The default maximum size of a compiled pattern is 64K. There have been a
3409 few cases of people hitting this limit. The code now uses macros to handle the
3410 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3411 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3412 configuring. Tests 2 and 5 work only with 2-byte links because they output
3413 debugging information about compiled patterns.
3414
3415 33. Internal code re-arrangements:
3416
3417 (a) Moved the debugging function for printing out a compiled regex into
3418 its own source file (printint.c) and used #include to pull it into
3419 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3420 separate copies.
3421
3422 (b) Defined the list of op-code names for debugging as a macro in
3423 internal.h so that it is next to the definition of the opcodes.
3424
3425 (c) Defined a table of op-code lengths for simpler skipping along compiled
3426 code. This is again a macro in internal.h so that it is next to the
3427 definition of the opcodes.
3428
3429 34. Added support for recursive calls to individual subpatterns, along the
3430 lines of Robin Houston's patch (but implemented somewhat differently).
3431
3432 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3433 allow it to read and process whole directories in Win32. This code was
3434 contributed by Lionel Fourquaux; it has not been tested by me.
3435
3436 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3437 used to name a group. Names consist of alphanumerics and underscores, and must
3438 be unique. Back references use the syntax (?P=name) and recursive calls use
3439 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3440 numbers. The function pcre_fullinfo() can be used after compilation to extract
3441 a name/number map. There are three relevant calls:
3442
3443 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3444 PCRE_INFO_NAMECOUNT yields the number of entries
3445 PCRE_INFO_NAMETABLE yields a pointer to the map.
3446
3447 The map is a vector of fixed-size entries. The size of each entry depends on
3448 the length of the longest name used. The first two bytes of each entry are the
3449 group number, most significant byte first. There follows the corresponding
3450 name, zero terminated. The names are in alphabetical order.
3451
3452 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3453 case instead of 255. Making it the same both with and without UTF-8 support
3454 means that the same test output works with both.
3455
3456 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3457 calling malloc() with a zero argument.
3458
3459 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3460 optimization. I've improved things by keeping a bitmap of backreferences with
3461 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3462 fact referenced, the optimization can be applied. It is unlikely that a
3463 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3464 the match to follow \n) will appear inside brackets with a number greater than
3465 31, but if it does, any back reference > 31 suppresses the optimization.
3466
3467 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3468 of disabling numbered capturing parentheses. Any opening parenthesis that is
3469 not followed by ? behaves as if it were followed by ?: but named parentheses
3470 can still be used for capturing (and they will acquire numbers in the usual
3471 way).
3472
3473 41. Redesigned the return codes from the match() function into yes/no/error so
3474 that errors can be passed back from deep inside the nested calls. A malloc
3475 failure while inside a recursive subpattern call now causes the
3476 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3477
3478 42. It is now possible to set a limit on the number of times the match()
3479 function is called in a call to pcre_exec(). This facility makes it possible to
3480 limit the amount of recursion and backtracking, though not in a directly
3481 obvious way, because the match() function is used in a number of different
3482 circumstances. The count starts from zero for each position in the subject
3483 string (for non-anchored patterns). The default limit is, for compatibility, a
3484 large number, namely 10 000 000. You can change this in two ways:
3485
3486 (a) When configuring PCRE before making, you can use --with-match-limit=n
3487 to set a default value for the compiled library.
3488
3489 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3490 a different value is set. See 45 below.
3491
3492 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3493
3494 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3495 of things that can be changed at compile time. The first argument specifies
3496 what is wanted and the second points to where the information is to be placed.
3497 The current list of available information is:
3498
3499 PCRE_CONFIG_UTF8
3500
3501 The output is an integer that is set to one if UTF-8 support is available;
3502 otherwise it is set to zero.
3503
3504 PCRE_CONFIG_NEWLINE
3505
3506 The output is an integer that it set to the value of the code that is used for
3507 newline. It is either LF (10) or CR (13).
3508
3509 PCRE_CONFIG_LINK_SIZE
3510
3511 The output is an integer that contains the number of bytes used for internal
3512 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3513
3514 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3515
3516 The output is an integer that contains the threshold above which the POSIX
3517 interface uses malloc() for output vectors. See item 31 above.
3518
3519 PCRE_CONFIG_MATCH_LIMIT
3520
3521 The output is an unsigned integer that contains the default limit of the number
3522 of match() calls in a pcre_exec() execution. See 42 above.
3523
3524 44. pcretest has been upgraded by the addition of the -C option. This causes it
3525 to extract all the available output from the new pcre_config() function, and to
3526 output it. The program then exits immediately.
3527
3528 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3529 order to support additional features. One way would have been to define
3530 pcre_exec2() (for example) with extra arguments, but this would not have been
3531 extensible, and would also have required all calls to the original function to
3532 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3533 is used for passing in "extra" data from pcre_study().
3534
3535 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3536 contains the following fields:
3537
3538 flags a bitmap indicating which of the following fields are set
3539 study_data opaque data from pcre_study()
3540 match_limit a way of specifying a limit on match() calls for a specific
3541 call to pcre_exec()
3542 callout_data data for callouts (see 49 below)
3543
3544 The flag bits are also defined in pcre.h, and are
3545
3546 PCRE_EXTRA_STUDY_DATA
3547 PCRE_EXTRA_MATCH_LIMIT
3548 PCRE_EXTRA_CALLOUT_DATA
3549
3550 The pcre_study() function now returns one of these new pcre_extra blocks, with
3551 the actual study data pointed to by the study_data field, and the
3552 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3553 before. That is, this change is entirely upwards-compatible and requires no
3554 change to existing code.
3555
3556 If you want to pass in additional data to pcre_exec(), you can either place it
3557 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3558 block.
3559
3560 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3561 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3562 times with different match limits, until it finds the minimum value needed for
3563 pcre_exec() to complete. The value is then output. This can be instructive; for
3564 most simple matches the number is quite small, but for pathological cases it
3565 gets very large very quickly.
3566
3567 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3568 returns the size of the data block pointed to by the study_data field in a
3569 pcre_extra block, that is, the value that was passed as the argument to
3570 pcre_malloc() when PCRE was getting memory in which to place the information
3571 created by pcre_study(). The fourth argument should point to a size_t variable.
3572 pcretest has been extended so that this information is shown after a successful
3573 pcre_study() call when information about the compiled regex is being displayed.
3574
3575 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3576 because what follows is always an absolute path. (Later: it turns out that this
3577 is more than cosmetic for MinGW, because it doesn't like empty path
3578 components.)
3579
3580 49. Some changes have been made to the callout feature (see 28 above):
3581
3582 (i) A callout function now has three choices for what it returns:
3583
3584 0 => success, carry on matching
3585 > 0 => failure at this point, but backtrack if possible
3586 < 0 => serious error, return this value from pcre_exec()
3587
3588 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3589 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3590 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3591 use by callout functions. It will never be used by PCRE itself.
3592
3593 (ii) The pcre_extra structure (see 45 above) has a void * field called
3594 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3595 pcre_callout_block structure has a field of the same name. The contents of
3596 the field passed in the pcre_extra structure are passed to the callout
3597 function in the corresponding field in the callout block. This makes it
3598 easier to use the same callout-containing regex from multiple threads. For
3599 testing, the pcretest program has a new data escape
3600
3601 \C*n pass the number n (may be negative) as callout_data
3602
3603 If the callout function in pcretest receives a non-zero value as
3604 callout_data, it returns that value.
3605
3606 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3607 there were some redundant $(CFLAGS) in commands that are now specified as
3608 $(LINK), which already includes $(CFLAGS).
3609
3610 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3611 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3612 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3613 one-byte characters throughout. Note that case-insensitive matching applies
3614 only to characters whose values are less than 256. PCRE doesn't support the
3615 notion of cases for higher-valued characters.
3616
3617 (i) A character class whose characters are all within 0-255 is handled as
3618 a bit map, and the map is inverted for negative classes. Previously, a
3619 character > 255 always failed to match such a class; however it should
3620 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3621
3622 (ii) A negated character class with a single character < 255 is coded as
3623 "not this character" (OP_NOT). This wasn't working properly when the test
3624 character was multibyte, either singly or repeated.
3625
3626 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3627 mode, for example: \x{100}{2,3}.
3628
3629 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3630 singly or repeated) now correctly test multibyte characters. However,
3631 PCRE doesn't recognize any characters with values greater than 255 as
3632 digits, spaces, or word characters. Such characters always match \D, \S,
3633 and \W, and never match \d, \s, or \w.
3634
3635 (v) Classes may now contain characters and character ranges with values
3636 greater than 255. For example: [ab\x{100}-\x{400}].
3637
3638 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3639 PCRE in UTF-8 mode.
3640
3641 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3642 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3643 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3644 value.)
3645
3646 53. The single man page has become too large. I have therefore split it up into
3647 a number of separate man pages. These also give rise to individual HTML pages;
3648 these are now put in a separate directory, and there is an index.html page that
3649 lists them all. Some hyperlinking between the pages has been installed.
3650
3651 54. Added convenience functions for handling named capturing parentheses.
3652
3653 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3654 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3655 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3656 are faulted.
3657
3658 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3659 calling configure. These values are used when compiling the dftables.c program
3660 which is run to generate the source of the default character tables. They
3661 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3662 you will need to set these values.
3663
3664 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3665
3666
3667 Version 3.9 02-Jan-02
3668 ---------------------
3669
3670 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3671
3672 2. If --disable-static was given, the building process failed when trying to
3673 build pcretest and pcregrep. (For some reason it was using libtool to compile
3674 them, which is not right, as they aren't part of the library.)
3675
3676
3677 Version 3.8 18-Dec-01
3678 ---------------------
3679
3680 1. The experimental UTF-8 code was completely screwed up. It was packing the
3681 bytes in the wrong order. How dumb can you get?
3682
3683
3684 Version 3.7 29-Oct-01
3685 ---------------------
3686
3687 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3688 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3689 this didn't happen under Solaris 8, where I normally test things.
3690
3691 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3692 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3693 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3694 this fix an hour or so after the initial 3.7 release.)
3695
3696
3697 Version 3.6 23-Oct-01
3698 ---------------------
3699
3700 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3701 offsets passed as NULL with zero offset count.
3702
3703 2. The config.guess and config.sub files had not been updated when I moved to
3704 the latest autoconf.
3705
3706
3707 Version 3.5 15-Aug-01
3708 ---------------------
3709
3710 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3711 had been forgotten.
3712
3713 2. By using declared but undefined structures, we can avoid using "void"
3714 definitions in pcre.h while keeping the internal definitions of the structures
3715 private.
3716
3717 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3718 user point of view, this means that both static and shared libraries are built
3719 by default, but this can be individually controlled. More of the work of
3720 handling this static/shared cases is now inside libtool instead of PCRE's make
3721 file.
3722
3723 4. The pcretest utility is now installed along with pcregrep because it is
3724 useful for users (to test regexs) and by doing this, it automatically gets
3725 relinked by libtool. The documentation has been turned into a man page, so
3726 there are now .1, .txt, and .html versions in /doc.
3727
3728 5. Upgrades to pcregrep:
3729 (i) Added long-form option names like gnu grep.
3730 (ii) Added --help to list all options with an explanatory phrase.
3731 (iii) Added -r, --recursive to recurse into sub-directories.
3732 (iv) Added -f, --file to read patterns from a file.
3733
3734 6. pcre_exec() was referring to its "code" argument before testing that
3735 argument for NULL (and giving an error if it was NULL).
3736
3737 7. Upgraded Makefile.in to allow for compiling in a different directory from
3738 the source directory.
3739
3740 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3741 options bits, the pointer it was passed was to an int instead of to an unsigned
3742 long int. This mattered only on 64-bit systems.
3743
3744 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3745 generated) instead of pcre.in, which it its source. Also made the same change
3746 in several of the .c files.
3747
3748 10. A new release of gcc defines printf() as a macro, which broke pcretest
3749 because it had an ifdef in the middle of a string argument for printf(). Fixed
3750 by using separate calls to printf().
3751
3752 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3753 script, to force use of CR or LF instead of \n in the source. On non-Unix
3754 systems, the value can be set in config.h.
3755
3756 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3757 absolute limit. Changed the text of the error message to make this clear, and
3758 likewise updated the man page.
3759
3760 13. The limit of 99 on the number of capturing subpatterns has been removed.
3761 The new limit is 65535, which I hope will not be a "real" limit.
3762
3763
3764 Version 3.4 22-Aug-00
3765 ---------------------
3766
3767 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3768
3769 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3770
3771
3772 Version 3.3 01-Aug-00
3773 ---------------------
3774
3775 1. If an octal character was given, but the value was greater than \377, it
3776 was not getting masked to the least significant bits, as documented. This could
3777 lead to crashes in some systems.
3778
3779 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3780 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3781
3782 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3783 These just pass their arguments on to (pcre_free)(), but they are provided
3784 because some uses of PCRE bind it to non-C systems that can call its functions,
3785 but cannot call free() or pcre_free() directly.
3786
3787 4. Add "make test" as a synonym for "make check". Corrected some comments in
3788 the Makefile.
3789
3790 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3791 Makefile.
3792
3793 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3794 command called pgrep for grepping around the active processes.
3795
3796 7. Added the beginnings of support for UTF-8 character strings.
3797
3798 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3799 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3800 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3801 out for the ar command.)
3802
3803
3804 Version 3.2 12-May-00
3805 ---------------------
3806
3807 This is purely a bug fixing release.
3808
3809 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3810 of ZA. This was just one example of several cases that could provoke this bug,
3811 which was introduced by change 9 of version 2.00. The code for breaking
3812 infinite loops after an iteration that matches an empty string was't working
3813 correctly.
3814
3815 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3816 when matched against abbab (for example). After matching an empty string, it
3817 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3818 caused it to match further down the string than it should.
3819
3820 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3821 was there because it doesn't seem to be needed, and it causes trouble on some
3822 systems, as it is not a Standard C header. It has been removed.
3823
3824 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3825 were reported on the Macintosh. The changes were from
3826
3827 while ((c = *(++ptr)) != 0 && c != '\n');
3828 to
3829 while ((c = *(++ptr)) != 0 && c != '\n') ;
3830
3831 Totally extraordinary, but if that's what it takes...
3832
3833 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3834 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3835 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3836 assumes the way PCRE uses memmove() (always moving upwards).
3837
3838 6. PCRE is being used in one environment where strchr() is not available. There
3839 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3840 faster code anyway.
3841
3842
3843 Version 3.1 09-Feb-00
3844 ---------------------
3845
3846 The only change in this release is the fixing of some bugs in Makefile.in for
3847 the "install" target:
3848
3849 (1) It was failing to install pcreposix.h.
3850
3851 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3852
3853
3854 Version 3.0 01-Feb-00
3855 ---------------------
3856
3857 1. Add support for the /+ modifier to perltest (to output $` like it does in
3858 pcretest).
3859
3860 2. Add support for the /g modifier to perltest.
3861
3862 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3863 matches null strings.
3864
3865 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3866 pattern. Perl treats empty patterns specially - it reuses the most recent
3867 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3868 effect.
3869
3870 5. The POSIX interface was broken in that it was just handing over the POSIX
3871 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3872 required a bigger vector, with some working space on the end. This means that
3873 the POSIX wrapper now has to get and free some memory, and copy the results.
3874
3875 6. Added some simple autoconf support, placing the test data and the
3876 documentation in separate directories, re-organizing some of the
3877 information files, and making it build pcre-config (a GNU standard). Also added
3878 libtool support for building PCRE as a shared library, which is now the
3879 default.
3880
3881 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3882 09 are not valid octal constants. Single digits will be used for minor values
3883 less than 10.
3884
3885 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3886 existing programs that set these in the POSIX interface can use PCRE without
3887 modification.
3888
3889 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3890 return all that pcre_info() returns, plus additional data. The pcre_info()
3891 function is retained for compatibility, but is considered to be obsolete.
3892
3893 10. Added experimental recursion feature (?R) to handle one common case that
3894 Perl 5.6 will be able to do with (?p{...}).
3895
3896 11. Added support for POSIX character classes like [:alpha:], which Perl is
3897 adopting.
3898
3899
3900 Version 2.08 31-Aug-99
3901 ----------------------
3902
3903 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3904 trying to match at the startoffset position, but instead was moving forward to
3905 the next newline as if a previous match had failed.
3906
3907 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3908 and could get into a loop if a null string was matched other than at the start
3909 of the subject.
3910
3911 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3912 be distinguished at compile time, and for completeness also added PCRE_DATE.
3913
3914 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3915 in GnuWin32 environments.
3916
3917
3918 Version 2.07 29-Jul-99
3919 ----------------------
3920
3921 1. The documentation is now supplied in plain text form and HTML as well as in
3922 the form of man page sources.
3923
3924 2. C++ compilers don't like assigning (void *) values to other pointer types.
3925 In particular this affects malloc(). Although there is no problem in Standard
3926 C, I've put in casts to keep C++ compilers happy.
3927
3928 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3929 should be (const char *).
3930
3931 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3932 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3933 However, I haven't made this a standard facility. The documentation doesn't
3934 mention it, and the Makefile doesn't support it.
3935
3936 5. The Makefile now contains an "install" target, with editable destinations at
3937 the top of the file. The pcretest program is not installed.
3938
3939 6. pgrep -V now gives the PCRE version number and date.
3940
3941 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3942 causing the entire string to be ignored, instead of just the last character.
3943
3944 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3945 non-matching string, it can take a very, very long time, even for strings of
3946 quite modest length, because of the nested recursion. PCRE now does better in
3947 some of these cases. It does this by remembering the last required literal
3948 character in the pattern, and pre-searching the subject to ensure it is present
3949 before running the real match. In other words, it applies a heuristic to detect
3950 some types of certain failure quickly, and in the above example, if presented
3951 with a string that has no trailing " it gives "no match" very quickly.
3952
3953 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3954 other alternatives are tried instead.
3955
3956
3957 Version 2.06 09-Jun-99
3958 ----------------------
3959
3960 1. Change pcretest's output for amount of store used to show just the code
3961 space, because the remainder (the data block) varies in size between 32-bit and
3962 64-bit systems.
3963
3964 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3965 start matching at. This allows lookbehinds to work when searching for multiple
3966 occurrences in a string.
3967
3968 3. Added additional options to pcretest for testing multiple occurrences:
3969
3970 /+ outputs the rest of the string that follows a match
3971 /g loops for multiple occurrences, using the new startoffset argument
3972 /G loops for multiple occurrences by passing an incremented pointer
3973
3974 4. PCRE wasn't doing the "first character" optimization for patterns starting
3975 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3976 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3977 the letter 'x'. On long subject strings, this gives a significant speed-up.
3978
3979
3980 Version 2.05 21-Apr-99
3981 ----------------------
3982
3983 1. Changed the type of magic_number from int to long int so that it works
3984 properly on 16-bit systems.
3985
3986 2. Fixed a bug which caused patterns starting with .* not to work correctly
3987 when the subject string contained newline characters. PCRE was assuming
3988 anchoring for such patterns in all cases, which is not correct because .* will
3989 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3990 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3991 must be retried after every newline in the subject.
3992
3993
3994 Version 2.04 18-Feb-99
3995 ----------------------
3996
3997 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3998 computation of the store needed to hold the pattern was incorrect (too large).
3999 If such patterns were nested a few deep, this could multiply and become a real
4000 problem.
4001
4002 2. Added /M option to pcretest to show the memory requirement of a specific
4003 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4004
4005 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4006 compiled in such a way that the backtracking after subsequent failure was
4007 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4008 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4009
4010
4011 Version 2.03 02-Feb-99
4012 ----------------------
4013
4014 1. Fixed typo and small mistake in man page.
4015
4016 2. Added 4th condition (GPL supersedes if conflict) and created separate
4017 LICENCE file containing the conditions.
4018
4019 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4020 Perl, that is the internal \ allows the delimiter to be included in the
4021 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4022 the final delimiter, add \ to the end of the pattern (to test the error).
4023
4024 4. Added the convenience functions for extracting substrings after a successful
4025 match. Updated pcretest to make it able to test these functions.
4026
4027
4028 Version 2.02 14-Jan-99
4029 ----------------------
4030
4031 1. Initialized the working variables associated with each extraction so that
4032 their saving and restoring doesn't refer to uninitialized store.
4033
4034 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4035 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4036 fix the problem.
4037
4038 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4039 calls, and wasn't printing the correct value for compiling calls. Increased the
4040 default value of LOOPREPEAT, and the number of significant figures in the
4041 times.
4042
4043 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4044
4045 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4046 a building problem on Windows NT with a FAT file system.
4047
4048
4049 Version 2.01 21-Oct-98
4050 ----------------------
4051
4052 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4053 to character tables built by pcre_maketables() in the current locale. If NULL
4054 is passed, the default tables are used.
4055
4056
4057 Version 2.00 24-Sep-98
4058 ----------------------
4059
4060 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4061 it any more.
4062
4063 2. Allow quantification of (?>) groups, and make it work correctly.
4064
4065 3. The first character computation wasn't working for (?>) groups.
4066
4067 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4068 end of the subject) and add 5.005's \z, which really does match only at the
4069 very end of the subject.
4070
4071 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4072
4073 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4074 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4075 localized options. All options to pcre_study() were also removed.
4076
4077 7. Add other new features from 5.005:
4078
4079 $(?<= positive lookbehind
4080 $(?<! negative lookbehind
4081 (?imsx-imsx) added the unsetting capability
4082 such a setting is global if at outer level; local otherwise
4083 (?imsx-imsx:) non-capturing groups with option setting
4084 (?(cond)re|re) conditional pattern matching
4085
4086 A backreference to itself in a repeated group matches the previous
4087 captured string.
4088
4089 8. General tidying up of studying (both automatic and via "study")
4090 consequential on the addition of new assertions.
4091
4092 9. As in 5.005, unlimited repeated groups that could match an empty substring
4093 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4094 runtime if any iteration does actually match an empty substring.
4095
4096 10. Include the RunTest script in the distribution.
4097
4098 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4099 discrepancies, some of which were old and were also with respect to 5.004. They
4100 have now been fixed.
4101
4102
4103 Version 1.09 28-Apr-98
4104 ----------------------
4105
4106 1. A negated single character class followed by a quantifier with a minimum
4107 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4108 program crashes, or just wrong answers. This did not apply to negated classes
4109 containing more than one character, or to minima other than one.
4110
4111
4112 Version 1.08 27-Mar-98
4113 ----------------------
4114
4115 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4116
4117 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4118 latter must appear before anything that relies on it in the pattern.
4119
4120
4121 Version 1.07 16-Feb-98
4122 ----------------------
4123
4124 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4125 repeat of a potentially empty string).
4126
4127
4128 Version 1.06 23-Jan-98
4129 ----------------------
4130
4131 1. Added Markus Oberhumer's little patches for C++.
4132
4133 2. Literal strings longer than 255 characters were broken.
4134
4135
4136 Version 1.05 23-Dec-97
4137 ----------------------
4138
4139 1. Negated character classes containing more than one character were failing if
4140 PCRE_CASELESS was set at run time.
4141
4142
4143 Version 1.04 19-Dec-97
4144 ----------------------
4145
4146 1. Corrected the man page, where some "const" qualifiers had been omitted.
4147
4148 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4149 input syntax.
4150
4151 3. Fixed memory leak which occurred when a regex with back references was
4152 matched with an offsets vector that wasn't big enough. The temporary memory
4153 that is used in this case wasn't being freed if the match failed.
4154
4155 4. Tidied pcretest to ensure it frees memory that it gets.
4156
4157 5. Temporary memory was being obtained in the case where the passed offsets
4158 vector was exactly big enough.
4159
4160 6. Corrected definition of offsetof() from change 5 below.
4161
4162 7. I had screwed up change 6 below and broken the rules for the use of
4163 setjmp(). Now fixed.
4164
4165
4166 Version 1.03 18-Dec-97
4167 ----------------------
4168
4169 1. A erroneous regex with a missing opening parenthesis was correctly
4170 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4171 on some systems.
4172
4173 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4174 it was reported that one broken compiler failed on the former because "code" is
4175 also an independent variable.
4176
4177 3. The erroneous regex a[]b caused an array overrun reference.
4178
4179 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4180 fail on data ending with that character. (It was going on too far, and checking
4181 the next character, typically a binary zero.) This was specific to the
4182 optimized code for single-character negative classes.
4183
4184 5. Added a contributed patch from the TIN world which does the following:
4185
4186 + Add an undef for memmove, in case the the system defines a macro for it.
4187
4188 + Add a definition of offsetof(), in case there isn't one. (I don't know
4189 the reason behind this - offsetof() is part of the ANSI standard - but
4190 it does no harm).
4191
4192 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4193 most of the places where whitespace preceded '#'. I have given up and
4194 allowed the remaining 2 cases to be at the margin.
4195
4196 + Rename some variables in pcre to eliminate shadowing. This seems very
4197 pedantic, but does no harm, of course.
4198
4199 6. Moved the call to setjmp() into its own function, to get rid of warnings
4200 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4201
4202 7. Constructs such as \d{8,} were compiling into the equivalent of
4203 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4204 outcome, but in this particular case used more store than had been allocated,
4205 which caused the bug to be discovered because it threw up an internal error.
4206
4207 8. The debugging code in both pcre and pcretest for outputting the compiled
4208 form of a regex was going wrong in the case of back references followed by
4209 curly-bracketed repeats.
4210
4211
4212 Version 1.02 12-Dec-97
4213 ----------------------
4214
4215 1. Typos in pcre.3 and comments in the source fixed.
4216
4217 2. Applied a contributed patch to get rid of places where it used to remove
4218 'const' from variables, and fixed some signed/unsigned and uninitialized
4219 variable warnings.
4220
4221 3. Added the "runtest" target to Makefile.
4222
4223 4. Set default compiler flag to -O2 rather than just -O.
4224
4225
4226 Version 1.01 19-Nov-97
4227 ----------------------
4228
4229 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4230 like /([ab]*)*/, that is, for classes with more than one character in them.
4231
4232 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4233 as /((?>a*))*/ (a PCRE_EXTRA facility).
4234
4235
4236 Version 1.00 18-Nov-97
4237 ----------------------
4238
4239 1. Added compile-time macros to support systems such as SunOS4 which don't have
4240 memmove() or strerror() but have other things that can be used instead.
4241
4242 2. Arranged that "make clean" removes the executables.
4243
4244
4245 Version 0.99 27-Oct-97
4246 ----------------------
4247
4248 1. Fixed bug in code for optimizing classes with only one character. It was
4249 initializing a 32-byte map regardless, which could cause it to run off the end
4250 of the memory it had got.
4251
4252 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4253
4254
4255 Version 0.98 22-Oct-97
4256 ----------------------
4257
4258 1. Fixed bug in code for handling temporary memory usage when there are more
4259 back references than supplied space in the ovector. This could cause segfaults.
4260
4261
4262 Version 0.97 21-Oct-97
4263 ----------------------
4264
4265 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4266
4267 2. Optimized negated single characters not to use a bit map.
4268
4269 3. Brought error texts together as macro definitions; clarified some of them;
4270 fixed one that was wrong - it said "range out of order" when it meant "invalid
4271 escape sequence".
4272
4273 4. Changed some char * arguments to const char *.
4274
4275 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4276
4277 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4278 pcretest.
4279
4280
4281 Version 0.96 16-Oct-97
4282 ----------------------
4283
4284 1. Added a simple "pgrep" utility to the distribution.
4285
4286 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4287 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4288 where "ddd" means "one or more decimal digits".
4289
4290 3. Fixed serious bug. If a pattern had a back reference, but the call to
4291 pcre_exec() didn't supply a large enough ovector to record the related
4292 identifying subpattern, the match always failed. PCRE now remembers the number
4293 of the largest back reference, and gets some temporary memory in which to save
4294 the offsets during matching if necessary, in order to ensure that
4295 backreferences always work.
4296
4297 4. Increased the compatibility with Perl in a number of ways:
4298
4299 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4300 to request this handling. The option can be set at compile or exec time.
4301
4302 (b) $ matches before a terminating newline by default; an option
4303 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4304 mode). The option can be set at compile or exec time.
4305
4306 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4307 the same as Perl's. If the decimal number it represents is less than 10
4308 or there aren't that many previous left capturing parentheses, an octal
4309 escape is read. Inside a character class, it's always an octal escape,
4310 even if it is a single digit.
4311
4312 (d) An escaped but undefined alphabetic character is taken as a literal,
4313 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4314 escapes.
4315
4316 (e) {0} is now permitted. (The previous item is removed from the compiled
4317 pattern).
4318
4319 5. Changed all the names of code files so that the basic parts are no longer
4320 than 10 characters, and abolished the teeny "globals.c" file.
4321
4322 6. Changed the handling of character classes; they are now done with a 32-byte
4323 bit map always.
4324
4325 7. Added the -d and /D options to pcretest to make it possible to look at the
4326 internals of compilation without having to recompile pcre.
4327
4328
4329 Version 0.95 23-Sep-97
4330 ----------------------
4331
4332 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4333 \x20 at the start of a run of normal characters. These were being treated as
4334 real characters, instead of the source characters being re-checked.
4335
4336
4337 Version 0.94 18-Sep-97
4338 ----------------------
4339
4340 1. The functions are now thread-safe, with the caveat that the global variables
4341 containing pointers to malloc() and free() or alternative functions are the
4342 same for all threads.
4343
4344 2. Get pcre_study() to generate a bitmap of initial characters for non-
4345 anchored patterns when this is possible, and use it if passed to pcre_exec().
4346
4347
4348 Version 0.93 15-Sep-97
4349 ----------------------
4350
4351 1. /(b)|(:+)/ was computing an incorrect first character.
4352
4353 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4354 but not actually doing anything yet.
4355
4356 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4357 as Perl does (e.g. [-az] or [az-]).
4358
4359 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4360 all possible positions.
4361
4362 5. Split up into different modules to avoid including unneeded functions in a
4363 compiled binary. However, compile and exec are still in one module. The "study"
4364 function is split off.
4365
4366 6. The character tables are now in a separate module whose source is generated
4367 by an auxiliary program - but can then be edited by hand if required. There are
4368 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4369 toupper() in the code.
4370
4371 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4372 make them global. Abolish the function for setting them, as the caller can now
4373 set them directly.
4374
4375
4376 Version 0.92 11-Sep-97
4377 ----------------------
4378
4379 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4380 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4381
4382 2. Caseless matching was not working in character classes if the characters in
4383 the pattern were in upper case.
4384
4385 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4386
4387 4. Make PCRE_ANCHORED public and accept as a compile option.
4388
4389 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4390 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4391 pass them.
4392
4393 6. Give an error if bad option bits passed at compile or run time.
4394
4395 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4396 pcretest to cause it to pass that flag.
4397
4398 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4399 options, and the first character, if set.
4400
4401 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4402
4403
4404 Version 0.91 10-Sep-97
4405 ----------------------
4406
4407 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4408 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4409
4410 2. PCRE was looping on encountering an indefinitely repeated back reference to
4411 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4412 Perl does - treats the match as successful.
4413
4414 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5