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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 654 - (show annotations) (download)
Tue Aug 2 11:00:40 2011 UTC (2 years, 11 months ago) by ph10
File size: 211165 byte(s)
Documentation and general text tidies in preparation for test release.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12