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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 645 - (show annotations)
Sun Jul 31 17:02:18 2011 UTC (4 years, 1 month ago) by ph10
File size: 211044 byte(s)
Error occurred while calculating annotation data.
Pass *MARK name to callouts
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.13 30-Apr-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, not 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 *disables*
111 studying, thereby overriding the use of -s on the command line. This is
112 necessary for one or two tests to keep the output identical in both cases.
113
114 21. When (*ACCEPT) was used in an assertion that matched an empty string and
115 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
116
117 22. When an atomic group that contained a capturing parenthesis was
118 successfully matched, but the branch in which it appeared failed, the
119 capturing was not being forgotten if a higher numbered group was later
120 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
121 group 1 to "a", when in fact it should be unset. This applied to multi-
122 branched capturing and non-capturing groups, repeated or not, and also to
123 positive assertions (capturing in negative assertions is not well defined
124 in PCRE) and also to nested atomic groups.
125
126 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
127 subject after a captured substring (to make it easier to tell which of a
128 number of identical substrings has been captured).
129
130 24. The way atomic groups are processed by pcre_exec() has been changed so that
131 if they are repeated, backtracking one repetition now resets captured
132 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
133 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
134 "aaa". Previously, it would have been "a". As part of this code
135 refactoring, the way recursive calls are handled has also been changed.
136
137 24. If an assertion condition captured any substrings, they were not passed
138 back unless some other capturing happened later. For example, if
139 (?(?=(a))a) was matched against "a", no capturing was returned.
140
141 25. When studying a pattern that contained subroutine calls or assertions,
142 the code for finding the minimum length of a possible match was handling
143 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
144 group 1 called group 2 while simultaneously a separate group 2 called group
145 1). A stack overflow occurred in this case. I have fixed this by limiting
146 the recursion depth to 10.
147
148 26. Updated RunTest.bat in the distribution to the version supplied by Tom
149 Fortmann. This supports explicit test numbers on the command line, and has
150 argument validation and error reporting.
151
152 27. An instance of \X with an unlimited repeat could fail if at any point the
153 first character it looked at was a mark character.
154
155 28. Some minor code refactoring concerning Unicode properties and scripts
156 should reduce the stack requirement of match() slightly.
157
158 29. Added the '=' option to pcretest to check the setting of unused capturing
159 slots at the end of the pattern, which are documented as being -1, but are
160 not included in the return count.
161
162 30. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
163 compiled something random. Now it gives a compile-time error (as does
164 Perl).
165
166 31. A *MARK encountered during the processing of a positive assertion is now
167 recorded and passed back (compatible with Perl).
168
169 32. If --only-matching or --colour was set on a pcregrep call whose pattern
170 had alternative anchored branches, the search for a second match in a line
171 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
172 matched the line "0102" twice. The same bug affected patterns that started
173 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
174 twice.
175
176 33. Previously, PCRE did not allow quantification of assertions. However, Perl
177 does, and because of capturing effects, quantifying parenthesized
178 assertions may at times be useful. Quantifiers are now allowed for
179 parenthesized assertions.
180
181 34. A minor code tidy in pcre_compile() when checking options for \R usage.
182
183 35. \g was being checked for fancy things in a character class, when it should
184 just be a literal "g".
185
186 36. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
187 appearance of a nested POSIX class supersedes an apparent external class.
188 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
189 unescaped square brackets may also appear as part of class names. For
190 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
191 more like Perl.
192
193 37. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
194 was because it thought it was \N{name}, which is not supported).
195
196 38. Add minix to OS list not supporting the -S option in pcretest.
197
198 39. PCRE tries to detect cases of infinite recursion at compile time, but it
199 cannot analyze patterns in sufficient detail to catch mutual recursions
200 such as ((?1))((?2)). There is now a runtime test that gives an error if a
201 subgroup is called recursively as a subpattern for a second time at the
202 same position in the subject string. In previous releases this might have
203 been caught by the recursion limit, or it might have run out of stack.
204
205 40. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
206 happen only once. PCRE was, however incorrectly giving a compile time error
207 "recursive call could loop indefinitely" because it cannot analyze the
208 pattern in sufficient detail. The compile time test no longer happens when
209 PCRE is compiling a conditional subpattern, but actual runaway loops are
210 now caught at runtime (see 39 above).
211
212 41. It seems that Perl allows any characters other than a closing parenthesis
213 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
214 has been changed to be the same.
215
216 42. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
217 as not to get warnings when autogen.sh is called. Also changed
218 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
219
220 43. To help people who use pcregrep to scan files containing exceedingly long
221 lines, the following changes have been made:
222
223 (a) The default value of the buffer size parameter has been increased from
224 8K to 20K. (A buffer three times this size is actually used.)
225
226 (b) The default can be changed by ./configure --with-pcregrep-bufsiz when
227 PCRE is built.
228
229 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
230 to be set at run time.
231
232 (d) Numerical values in pcregrep options can be followed by K or M, for
233 example --buffer-size=50K.
234
235 (e) If a line being scanned overflows pcregrep's buffer, an error is now
236 given and the return code is set to 2.
237
238 44. Add a pointer to the latest mark to the callout data block.
239
240
241 Version 8.12 15-Jan-2011
242 ------------------------
243
244 1. Fixed some typos in the markup of the man pages, and wrote a script that
245 checks for such things as part of the documentation building process.
246
247 2. On a big-endian 64-bit system, pcregrep did not correctly process the
248 --match-limit and --recursion-limit options (added for 8.11). In
249 particular, this made one of the standard tests fail. (The integer value
250 went into the wrong half of a long int.)
251
252 3. If the --colour option was given to pcregrep with -v (invert match), it
253 did strange things, either producing crazy output, or crashing. It should,
254 of course, ignore a request for colour when reporting lines that do not
255 match.
256
257 4. Another pcregrep bug caused similar problems if --colour was specified with
258 -M (multiline) and the pattern match finished with a line ending.
259
260 5. In pcregrep, when a pattern that ended with a literal newline sequence was
261 matched in multiline mode, the following line was shown as part of the
262 match. This seems wrong, so I have changed it.
263
264 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
265 the check for further matches in the same line (so they could be coloured)
266 to overrun the end of the current line. If another match was found, it was
267 incorrectly shown (and then shown again when found in the next line).
268
269 7. If pcregrep was compiled under Windows, there was a reference to the
270 function pcregrep_exit() before it was defined. I am assuming this was
271 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
272 reported by a user. I've moved the definition above the reference.
273
274
275 Version 8.11 10-Dec-2010
276 ------------------------
277
278 1. (*THEN) was not working properly if there were untried alternatives prior
279 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
280 backtracked to try for "b" instead of moving to the next alternative branch
281 at the same level (in this case, to look for "c"). The Perl documentation
282 is clear that when (*THEN) is backtracked onto, it goes to the "next
283 alternative in the innermost enclosing group".
284
285 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
286 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
287 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
288 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
289 (*THEN).
290
291 3. If \s appeared in a character class, it removed the VT character from
292 the class, even if it had been included by some previous item, for example
293 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
294 of \s, but is part of the POSIX "space" class.)
295
296 4. A partial match never returns an empty string (because you can always
297 match an empty string at the end of the subject); however the checking for
298 an empty string was starting at the "start of match" point. This has been
299 changed to the "earliest inspected character" point, because the returned
300 data for a partial match starts at this character. This means that, for
301 example, /(?<=abc)def/ gives a partial match for the subject "abc"
302 (previously it gave "no match").
303
304 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
305 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
306 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
307 has an implication that the given string is incomplete (because a partial
308 match is preferred over a full match). For this reason, these items now
309 give a partial match in this situation. [Aside: previously, the one case
310 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
311 match rather than a full match, which was wrong by the old rules, but is
312 now correct.]
313
314 6. There was a bug in the handling of #-introduced comments, recognized when
315 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
316 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
317 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
318 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
319 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
320 places in pcre_compile().
321
322 7. Related to (6) above, when pcre_compile() was skipping #-introduced
323 comments when looking ahead for named forward references to subpatterns,
324 the only newline sequence it recognized was NL. It now handles newlines
325 according to the set newline convention.
326
327 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
328 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
329 cater for a lack of strerror(). These oversights have been fixed.
330
331 9. Added --match-limit and --recursion-limit to pcregrep.
332
333 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
334
335 11. When the -o option was used, pcregrep was setting a return code of 1, even
336 when matches were found, and --line-buffered was not being honoured.
337
338 12. Added an optional parentheses number to the -o and --only-matching options
339 of pcregrep.
340
341 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
342 can match an empty string. The code to do it in pcretest and pcredemo
343 needed fixing:
344
345 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
346 only one byte after an empty string match just before CRLF (this case
347 just got forgotten; "any" and "anycrlf" were OK).
348
349 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
350 mode when an empty string match preceded an ASCII character followed by
351 a non-ASCII character. (The code for advancing by one character rather
352 than one byte was nonsense.)
353
354 (c) The pcredemo.c sample program did not have any code at all to handle
355 the cases when CRLF is a valid newline sequence.
356
357 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
358 as a starting offset was within the subject string. There is now a new
359 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
360 negative or greater than the length of the string. In order to test this,
361 pcretest is extended to allow the setting of negative starting offsets.
362
363 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
364 starting offset points to the beginning of a UTF-8 character was
365 unnecessarily clumsy. I tidied it up.
366
367 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
368 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
369
370 17. Nobody had reported that the --include_dir option, which was added in
371 release 7.7 should have been called --include-dir (hyphen, not underscore)
372 for compatibility with GNU grep. I have changed it to --include-dir, but
373 left --include_dir as an undocumented synonym, and the same for
374 --exclude-dir, though that is not available in GNU grep, at least as of
375 release 2.5.4.
376
377 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
378 characters from a string of bytes) have been redefined so as not to use
379 loops, in order to improve performance in some environments. At the same
380 time, I abstracted some of the common code into auxiliary macros to save
381 repetition (this should not affect the compiled code).
382
383 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
384 compile-time error is now given if \c is not followed by an ASCII
385 character, that is, a byte less than 128. (In EBCDIC mode, the code is
386 different, and any byte value is allowed.)
387
388 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
389 START_OPTIMIZE option, which is now allowed at compile time - but just
390 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
391 to pcregrep and other applications that have no direct access to PCRE
392 options. The new /Y option in pcretest sets this option when calling
393 pcre_compile().
394
395 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
396 back references. Groups containing recursive back references were forced to
397 be atomic by that change, but in the case of named groups, the amount of
398 memory required was incorrectly computed, leading to "Failed: internal
399 error: code overflow". This has been fixed.
400
401 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
402 pcretest.c, to avoid build problems in some Borland environments.
403
404
405 Version 8.10 25-Jun-2010
406 ------------------------
407
408 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
409 THEN.
410
411 2. (*ACCEPT) was not working when inside an atomic group.
412
413 3. Inside a character class, \B is treated as a literal by default, but
414 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
415 causes the error). The code is unchanged, but I tidied the documentation.
416
417 4. Inside a character class, PCRE always treated \R and \X as literals,
418 whereas Perl faults them if its -w option is set. I have changed PCRE so
419 that it faults them when PCRE_EXTRA is set.
420
421 5. Added support for \N, which always matches any character other than
422 newline. (It is the same as "." when PCRE_DOTALL is not set.)
423
424 6. When compiling pcregrep with newer versions of gcc which may have
425 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
426 declared with attribute warn_unused_result" were given. Just casting the
427 result to (void) does not stop the warnings; a more elaborate fudge is
428 needed. I've used a macro to implement this.
429
430 7. Minor change to pcretest.c to avoid a compiler warning.
431
432 8. Added four artifical Unicode properties to help with an option to make
433 \s etc use properties (see next item). The new properties are: Xan
434 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
435
436 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
437 use Unicode properties. (*UCP) at the start of a pattern can be used to set
438 this option. Modified pcretest to add /W to test this facility. Added
439 REG_UCP to make it available via the POSIX interface.
440
441 10. Added --line-buffered to pcregrep.
442
443 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
444 studied, and the match started with a letter with a code point greater than
445 127 whose first byte was different to the first byte of the other case of
446 the letter, the other case of this starting letter was not recognized
447 (#976).
448
449 12. If a pattern that was studied started with a repeated Unicode property
450 test, for example, \p{Nd}+, there was the theoretical possibility of
451 setting up an incorrect bitmap of starting bytes, but fortunately it could
452 not have actually happened in practice until change 8 above was made (it
453 added property types that matched character-matching opcodes).
454
455 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
456 possible starting bytes for non-anchored patterns.
457
458 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
459 \R, and also a number of cases that involve Unicode properties, both
460 explicit and implicit when PCRE_UCP is set.
461
462 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
463 input, it could crash or give wrong results if characters with values
464 greater than 0xc0 were present in the subject string. (Detail: it assumed
465 UTF-8 input when processing these items.)
466
467 16. Added a lot of (int) casts to avoid compiler warnings in systems where
468 size_t is 64-bit (#991).
469
470 17. Added a check for running out of memory when PCRE is compiled with
471 --disable-stack-for-recursion (#990).
472
473 18. If the last data line in a file for pcretest does not have a newline on
474 the end, a newline was missing in the output.
475
476 19. The default pcre_chartables.c file recognizes only ASCII characters (values
477 less than 128) in its various bitmaps. However, there is a facility for
478 generating tables according to the current locale when PCRE is compiled. It
479 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
480 characters, are recognized by isspace() and therefore were getting set in
481 these tables, and indeed these tables seem to approximate to ISO 8859. This
482 caused a problem in UTF-8 mode when pcre_study() was used to create a list
483 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
484 which of course cannot start UTF-8 characters. I have changed the code so
485 that only real ASCII characters (less than 128) and the correct starting
486 bytes for UTF-8 encodings are set for characters greater than 127 when in
487 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
488 altogether.)
489
490 20. Added the /T option to pcretest so as to be able to run tests with non-
491 standard character tables, thus making it possible to include the tests
492 used for 19 above in the standard set of tests.
493
494 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
495 reference to a subpattern the other side of a comment that contains an
496 opening parenthesis caused either an internal compiling error, or a
497 reference to the wrong subpattern.
498
499
500 Version 8.02 19-Mar-2010
501 ------------------------
502
503 1. The Unicode data tables have been updated to Unicode 5.2.0.
504
505 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
506 configured.
507
508 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
509 original author of that file, following a query about its status.
510
511 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
512 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
513
514 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
515 quantifier applied to a forward-referencing subroutine call, could compile
516 incorrect code or give the error "internal error: previously-checked
517 referenced subpattern not found".
518
519 6. Both MS Visual Studio and Symbian OS have problems with initializing
520 variables to point to external functions. For these systems, therefore,
521 pcre_malloc etc. are now initialized to local functions that call the
522 relevant global functions.
523
524 7. There were two entries missing in the vectors called coptable and poptable
525 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
526 I've fixed the data, and added a kludgy way of testing at compile time that
527 the lengths are correct (equal to the number of opcodes).
528
529 8. Following on from 7, I added a similar kludge to check the length of the
530 eint vector in pcreposix.c.
531
532 9. Error texts for pcre_compile() are held as one long string to avoid too
533 much relocation at load time. To find a text, the string is searched,
534 counting zeros. There was no check for running off the end of the string,
535 which could happen if a new error number was added without updating the
536 string.
537
538 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
539
540 11. \K was not working if it appeared in an atomic group or in a group that
541 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
542 \K is "not well defined" if used in an assertion. PCRE now accepts it if
543 the assertion is positive, but not if it is negative.
544
545 12. Change 11 fortuitously reduced the size of the stack frame used in the
546 "match()" function of pcre_exec.c by one pointer. Forthcoming
547 implementation of support for (*MARK) will need an extra pointer on the
548 stack; I have reserved it now, so that the stack frame size does not
549 decrease.
550
551 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
552 item in branch that calls a recursion is a subroutine call - as in the
553 second branch in the above example - was incorrectly given the compile-
554 time error "recursive call could loop indefinitely" because pcre_compile()
555 was not correctly checking the subroutine for matching a non-empty string.
556
557 14. The checks for overrunning compiling workspace could trigger after an
558 overrun had occurred. This is a "should never occur" error, but it can be
559 triggered by pathological patterns such as hundreds of nested parentheses.
560 The checks now trigger 100 bytes before the end of the workspace.
561
562 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
563
564
565 Version 8.01 19-Jan-2010
566 ------------------------
567
568 1. If a pattern contained a conditional subpattern with only one branch (in
569 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
570 computed the wrong minimum data length (which is of course zero for such
571 subpatterns). This could cause incorrect "no match" results.
572
573 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
574 the pattern is reset in the first branch, pcre_compile() failed with
575 "internal error: code overflow at offset...". This happened only when
576 the reset was to the original external option setting. (An optimization
577 abstracts leading options settings into an external setting, which was the
578 cause of this.)
579
580 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
581 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
582 assertion pattern did not match (meaning that the assertion was true), it
583 was incorrectly treated as false if the SKIP had been reached during the
584 matching. This also applied to assertions used as conditions.
585
586 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
587 assertion subpattern, including such a pattern used as a condition,
588 unpredictable results occurred, instead of the error return
589 PCRE_ERROR_DFA_UITEM.
590
591 5. The C++ GlobalReplace function was not working like Perl for the special
592 situation when an empty string is matched. It now does the fancy magic
593 stuff that is necessary.
594
595 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
596 removed. (These were left over from very, very early versions of PCRE.)
597
598 7. Some cosmetic changes to the code to make life easier when compiling it
599 as part of something else:
600
601 (a) Change DEBUG to PCRE_DEBUG.
602
603 (b) In pcre_compile(), rename the member of the "branch_chain" structure
604 called "current" as "current_branch", to prevent a collision with the
605 Linux macro when compiled as a kernel module.
606
607 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
608 prevent a collision with the Linux macro when compiled as a kernel
609 module.
610
611 8. In pcre_compile() there are some checks for integer overflows that used to
612 cast potentially large values to (double). This has been changed to that
613 when building, a check for int64_t is made, and if it is found, it is used
614 instead, thus avoiding the use of floating point arithmetic. (There is no
615 other use of FP in PCRE.) If int64_t is not found, the fallback is to
616 double.
617
618 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
619 2005 (difference between two addresses compared to an unsigned value).
620
621 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
622 custom one, because of the following reported problem in Windows:
623
624 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
625 under Win32.
626 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
627 therefore missing the function definition.
628 - The compiler thus generates a "C" signature for the test function.
629 - The linker fails to find the "C" function.
630 - PCRE fails to configure if asked to do so against libbz2.
631
632 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
633 messages were output:
634
635 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
636 rerunning libtoolize, to keep the correct libtool macros in-tree.
637 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
638
639 I have done both of these things.
640
641 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
642 most of the time, it *can* run out if it is given a pattern that contains a
643 runaway infinite recursion. I updated the discussion in the pcrestack man
644 page.
645
646 13. Now that we have gone to the x.xx style of version numbers, the minor
647 version may start with zero. Using 08 or 09 is a bad idea because users
648 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
649 interpreted as invalid octal numbers. I've updated the previous comment in
650 configure.ac, and also added a check that gives an error if 08 or 09 are
651 used.
652
653 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
654 causing partial matching to fail when the end of the subject matched \W
655 in a UTF-8 pattern where \W was quantified with a minimum of 3.
656
657 15. There were some discrepancies between the declarations in pcre_internal.h
658 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
659 their definitions. The declarations used "const uschar *" and the
660 definitions used USPTR. Even though USPTR is normally defined as "const
661 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
662 reported that: "This difference in casting confuses some C++ compilers, for
663 example, SunCC recognizes above declarations as different functions and
664 generates broken code for hbpcre." I have changed the declarations to use
665 USPTR.
666
667 16. GNU libtool is named differently on some systems. The autogen.sh script now
668 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
669 (FreeBSD).
670
671 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
672 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
673 comment: "Figure out how to create a longlong from a string: strtoll and
674 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
675 instance, but it only takes 2 args instead of 3!"
676
677 18. A subtle bug concerned with back references has been fixed by a change of
678 specification, with a corresponding code fix. A pattern such as
679 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
680 refers, was giving matches when it shouldn't. For example, xa=xaaa would
681 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
682 same bug. Such groups have to be quantified to be useful, or contained
683 inside another quantified group. (If there's no repetition, the reference
684 can never match.) The problem arises because, having left the group and
685 moved on to the rest of the pattern, a later failure that backtracks into
686 the group uses the captured value from the final iteration of the group
687 rather than the correct earlier one. I have fixed this in PCRE by forcing
688 any group that contains a reference to itself to be an atomic group; that
689 is, there cannot be any backtracking into it once it has completed. This is
690 similar to recursive and subroutine calls.
691
692
693 Version 8.00 19-Oct-09
694 ----------------------
695
696 1. The table for translating pcre_compile() error codes into POSIX error codes
697 was out-of-date, and there was no check on the pcre_compile() error code
698 being within the table. This could lead to an OK return being given in
699 error.
700
701 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
702 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
703 in a Windows environment.
704
705 3. The pcregrep --count option prints the count for each file even when it is
706 zero, as does GNU grep. However, pcregrep was also printing all files when
707 --files-with-matches was added. Now, when both options are given, it prints
708 counts only for those files that have at least one match. (GNU grep just
709 prints the file name in this circumstance, but including the count seems
710 more useful - otherwise, why use --count?) Also ensured that the
711 combination -clh just lists non-zero counts, with no names.
712
713 4. The long form of the pcregrep -F option was incorrectly implemented as
714 --fixed_strings instead of --fixed-strings. This is an incompatible change,
715 but it seems right to fix it, and I didn't think it was worth preserving
716 the old behaviour.
717
718 5. The command line items --regex=pattern and --regexp=pattern were not
719 recognized by pcregrep, which required --regex pattern or --regexp pattern
720 (with a space rather than an '='). The man page documented the '=' forms,
721 which are compatible with GNU grep; these now work.
722
723 6. No libpcreposix.pc file was created for pkg-config; there was just
724 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
725
726 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
727 when UCP support is not needed, by modifying the Python script that
728 generates it from Unicode data files. This should not matter if the module
729 is correctly used as a library, but I received one complaint about 50K of
730 unwanted data. My guess is that the person linked everything into his
731 program rather than using a library. Anyway, it does no harm.
732
733 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
734 was a minimum greater than 1 for a wide character in a possessive
735 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
736 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
737 character. Chaos in the form of incorrect output or a compiling loop could
738 result.
739
740 9. The restrictions on what a pattern can contain when partial matching is
741 requested for pcre_exec() have been removed. All patterns can now be
742 partially matched by this function. In addition, if there are at least two
743 slots in the offset vector, the offset of the earliest inspected character
744 for the match and the offset of the end of the subject are set in them when
745 PCRE_ERROR_PARTIAL is returned.
746
747 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
748 synonymous with PCRE_PARTIAL, for backwards compatibility, and
749 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
750 and may be more useful for multi-segment matching.
751
752 11. Partial matching with pcre_exec() is now more intuitive. A partial match
753 used to be given if ever the end of the subject was reached; now it is
754 given only if matching could not proceed because another character was
755 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
756 string "Z", which now yields "no match" instead of "partial match". In the
757 case of pcre_dfa_exec(), "no match" is given if every matching path for the
758 final character ended with (*FAIL).
759
760 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
761 if the pattern had a "must contain" character that was already found in the
762 earlier partial match, unless partial matching was again requested. For
763 example, with the pattern /dog.(body)?/, the "must contain" character is
764 "g". If the first part-match was for the string "dog", restarting with
765 "sbody" failed. This bug has been fixed.
766
767 13. The string returned by pcre_dfa_exec() after a partial match has been
768 changed so that it starts at the first inspected character rather than the
769 first character of the match. This makes a difference only if the pattern
770 starts with a lookbehind assertion or \b or \B (\K is not supported by
771 pcre_dfa_exec()). It's an incompatible change, but it makes the two
772 matching functions compatible, and I think it's the right thing to do.
773
774 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
775 so that the demonstration program is easily available in environments where
776 PCRE has not been installed from source.
777
778 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
779 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
780 library.
781
782 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
783 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
784 is not the first non-POSIX option to be added. Clearly some people find
785 these options useful.
786
787 17. If a caller to the POSIX matching function regexec() passes a non-zero
788 value for nmatch with a NULL value for pmatch, the value of
789 nmatch is forced to zero.
790
791 18. RunGrepTest did not have a test for the availability of the -u option of
792 the diff command, as RunTest does. It now checks in the same way as
793 RunTest, and also checks for the -b option.
794
795 19. If an odd number of negated classes containing just a single character
796 interposed, within parentheses, between a forward reference to a named
797 subpattern and the definition of the subpattern, compilation crashed with
798 an internal error, complaining that it could not find the referenced
799 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
800 [The bug was that it was starting one character too far in when skipping
801 over the character class, thus treating the ] as data rather than
802 terminating the class. This meant it could skip too much.]
803
804 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
805 /g option in pcretest when the pattern contains \K, which makes it possible
806 to have an empty string match not at the start, even when the pattern is
807 anchored. Updated pcretest and pcredemo to use this option.
808
809 21. If the maximum number of capturing subpatterns in a recursion was greater
810 than the maximum at the outer level, the higher number was returned, but
811 with unset values at the outer level. The correct (outer level) value is
812 now given.
813
814 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
815 PCRE did not set those parentheses (unlike Perl). I have now found a way to
816 make it do so. The string so far is captured, making this feature
817 compatible with Perl.
818
819 23. The tests have been re-organized, adding tests 11 and 12, to make it
820 possible to check the Perl 5.10 features against Perl 5.10.
821
822 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
823 pattern matches a fixed length string. PCRE did not allow this; now it
824 does. Neither allows recursion.
825
826 25. I finally figured out how to implement a request to provide the minimum
827 length of subject string that was needed in order to match a given pattern.
828 (It was back references and recursion that I had previously got hung up
829 on.) This code has now been added to pcre_study(); it finds a lower bound
830 to the length of subject needed. It is not necessarily the greatest lower
831 bound, but using it to avoid searching strings that are too short does give
832 some useful speed-ups. The value is available to calling programs via
833 pcre_fullinfo().
834
835 26. While implementing 25, I discovered to my embarrassment that pcretest had
836 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
837 study optimizations had never been tested with that matching function.
838 Oops. What is worse, even when it was passed study data, there was a bug in
839 pcre_dfa_exec() that meant it never actually used it. Double oops. There
840 were also very few tests of studied patterns with pcre_dfa_exec().
841
842 27. If (?| is used to create subpatterns with duplicate numbers, they are now
843 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
844 on the other side of the coin, they are no longer allowed to have different
845 names, because these cannot be distinguished in PCRE, and this has caused
846 confusion. (This is a difference from Perl.)
847
848 28. When duplicate subpattern names are present (necessarily with different
849 numbers, as required by 27 above), and a test is made by name in a
850 conditional pattern, either for a subpattern having been matched, or for
851 recursion in such a pattern, all the associated numbered subpatterns are
852 tested, and the overall condition is true if the condition is true for any
853 one of them. This is the way Perl works, and is also more like the way
854 testing by number works.
855
856
857 Version 7.9 11-Apr-09
858 ---------------------
859
860 1. When building with support for bzlib/zlib (pcregrep) and/or readline
861 (pcretest), all targets were linked against these libraries. This included
862 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
863 libraries. This caused unwanted dependencies to be created. This problem
864 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
865 pcretest is linked with readline.
866
867 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
868 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
869 moved outside it again, because FALSE and TRUE are already defined in AIX,
870 but BOOL is not.
871
872 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
873 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
874
875 4. The pcregrep documentation said spaces were inserted as well as colons (or
876 hyphens) following file names and line numbers when outputting matching
877 lines. This is not true; no spaces are inserted. I have also clarified the
878 wording for the --colour (or --color) option.
879
880 5. In pcregrep, when --colour was used with -o, the list of matching strings
881 was not coloured; this is different to GNU grep, so I have changed it to be
882 the same.
883
884 6. When --colo(u)r was used in pcregrep, only the first matching substring in
885 each matching line was coloured. Now it goes on to look for further matches
886 of any of the test patterns, which is the same behaviour as GNU grep.
887
888 7. A pattern that could match an empty string could cause pcregrep to loop; it
889 doesn't make sense to accept an empty string match in pcregrep, so I have
890 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
891 seems to be how GNU grep behaves.
892
893 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
894 start or after a newline", because the conditional assertion was not being
895 correctly handled. The rule now is that both the assertion and what follows
896 in the first alternative must satisfy the test.
897
898 9. If auto-callout was enabled in a pattern with a conditional group whose
899 condition was an assertion, PCRE could crash during matching, both with
900 pcre_exec() and pcre_dfa_exec().
901
902 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
903 used for matching.
904
905 11. Unicode property support in character classes was not working for
906 characters (bytes) greater than 127 when not in UTF-8 mode.
907
908 12. Added the -M command line option to pcretest.
909
910 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
911
912 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
913
914 16. Added comments and documentation about mis-use of no_arg in the C++
915 wrapper.
916
917 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
918 from Martin Jerabek that uses macro names for all relevant character and
919 string constants.
920
921 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
922 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
923 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
924 these, but not everybody uses configure.
925
926 19. A conditional group that had only one branch was not being correctly
927 recognized as an item that could match an empty string. This meant that an
928 enclosing group might also not be so recognized, causing infinite looping
929 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
930 with the subject "ab", where knowledge that the repeated group can match
931 nothing is needed in order to break the loop.
932
933 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
934 exec(), but without supplying a callout function, matching went wrong.
935
936 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
937 leak if the size of the offset vector was greater than 30. When the vector
938 is smaller, the saved offsets during recursion go onto a local stack
939 vector, but for larger vectors malloc() is used. It was failing to free
940 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
941 error, in fact).
942
943 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
944 heapframe that is used only when UTF-8 support is enabled. This caused no
945 problem, but was untidy.
946
947 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
948 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
949 included within another project.
950
951 24. Steven Van Ingelgem's patches to add more options to the CMake support,
952 slightly modified by me:
953
954 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
955 not building pcregrep.
956
957 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
958 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
959
960 25. Forward references, both numeric and by name, in patterns that made use of
961 duplicate group numbers, could behave incorrectly or give incorrect errors,
962 because when scanning forward to find the reference group, PCRE was not
963 taking into account the duplicate group numbers. A pattern such as
964 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
965
966 26. Changed a few more instances of "const unsigned char *" to USPTR, making
967 the feature of a custom pointer more persuasive (as requested by a user).
968
969 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
970 pcretest.c, inside #ifndefs, because it seems they are sometimes already
971 pre-defined.
972
973 28. Added support for (*UTF8) at the start of a pattern.
974
975 29. Arrange for flags added by the "release type" setting in CMake to be shown
976 in the configuration summary.
977
978
979 Version 7.8 05-Sep-08
980 ---------------------
981
982 1. Replaced UCP searching code with optimized version as implemented for Ad
983 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
984 stage table and inline lookup instead of a function, giving speed ups of 2
985 to 5 times on some simple patterns that I tested. Permission was given to
986 distribute the MultiStage2.py script that generates the tables (it's not in
987 the tarball, but is in the Subversion repository).
988
989 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
990 scripts.
991
992 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
993 a group with a zero qualifier. The result of the study could be incorrect,
994 or the function might crash, depending on the pattern.
995
996 4. Caseless matching was not working for non-ASCII characters in back
997 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
998 It now works when Unicode Property Support is available.
999
1000 5. In pcretest, an escape such as \x{de} in the data was always generating
1001 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1002 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1003 truncation.
1004
1005 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1006
1007 7. Added two (int) casts to pcregrep when printing the difference of two
1008 pointers, in case they are 64-bit values.
1009
1010 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1011 test 2 if it fails.
1012
1013 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1014 and a #define of that name to empty if it is not externally set. This is to
1015 allow users of MSVC to set it if necessary.
1016
1017 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1018 the convenience functions in the pcre_get.c source file.
1019
1020 11. An option change at the start of a pattern that had top-level alternatives
1021 could cause overwriting and/or a crash. This command provoked a crash in
1022 some environments:
1023
1024 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1025
1026 This potential security problem was recorded as CVE-2008-2371.
1027
1028 12. For a pattern where the match had to start at the beginning or immediately
1029 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1030 pcre_dfa_exec() could read past the end of the passed subject if there was
1031 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1032 pcretest so that it places the subject at the end of its malloc-ed buffer.
1033
1034 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1035 exec() might read past the end of the data buffer in UTF-8 mode.
1036
1037 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1038 the data contained the byte 0x85 as part of a UTF-8 character within its
1039 first line. This applied both to normal and DFA matching.
1040
1041 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1042 /^[^d]*?$/8 failed to match "abc".
1043
1044 16. Added a missing copyright notice to pcrecpp_internal.h.
1045
1046 17. Make it more clear in the documentation that values returned from
1047 pcre_exec() in ovector are byte offsets, not character counts.
1048
1049 18. Tidied a few places to stop certain compilers from issuing warnings.
1050
1051 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1052 supplied by Stefan Weber. I made a further small update for 7.8 because
1053 there is a change of source arrangements: the pcre_searchfuncs.c module is
1054 replaced by pcre_ucd.c.
1055
1056
1057 Version 7.7 07-May-08
1058 ---------------------
1059
1060 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1061 a string to a long long, pretend we don't even have a long long." This is
1062 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1063
1064 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1065 pre-7.6 versions, which defined a global no_arg variable instead of putting
1066 it in the RE class. (See also #8 below.)
1067
1068 3. Remove a line of dead code, identified by coverity and reported by Nuno
1069 Lopes.
1070
1071 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1072
1073 (1) The include/exclude patterns were being applied to the whole pathnames
1074 of files, instead of just to the final components.
1075
1076 (2) If there was more than one level of directory, the subdirectories were
1077 skipped unless they satisfied the include/exclude conditions. This is
1078 inconsistent with GNU grep (and could even be seen as contrary to the
1079 pcregrep specification - which I improved to make it absolutely clear).
1080 The action now is always to scan all levels of directory, and just
1081 apply the include/exclude patterns to regular files.
1082
1083 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1084 --exclude_dir in the tests to avoid scanning .svn directories.
1085
1086 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1087 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1088 doesn't support NULs in patterns.
1089
1090 7. Added some missing "const"s to declarations of static tables in
1091 pcre_compile.c and pcre_dfa_exec.c.
1092
1093 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1094 caused by fix #2 above. (Subsequently also a second patch to fix the
1095 first patch. And a third patch - this was a messy problem.)
1096
1097 9. Applied Craig's patch to remove the use of push_back().
1098
1099 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1100 matching function regexec().
1101
1102 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1103 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1104 references. PCRE supports relative numbers with this syntax (I don't think
1105 Oniguruma does).
1106
1107 12. Previously, a group with a zero repeat such as (...){0} was completely
1108 omitted from the compiled regex. However, this means that if the group
1109 was called as a subroutine from elsewhere in the pattern, things went wrong
1110 (an internal error was given). Such groups are now left in the compiled
1111 pattern, with a new opcode that causes them to be skipped at execution
1112 time.
1113
1114 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1115 to the way PCRE behaves:
1116
1117 (a) A lone ] character is dis-allowed (Perl treats it as data).
1118
1119 (b) A back reference to an unmatched subpattern matches an empty string
1120 (Perl fails the current match path).
1121
1122 (c) A data ] in a character class must be notated as \] because if the
1123 first data character in a class is ], it defines an empty class. (In
1124 Perl it is not possible to have an empty class.) The empty class []
1125 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1126 The negative empty class [^] matches any one character, independently
1127 of the DOTALL setting.
1128
1129 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1130 non-existent subpattern following a character class starting with ']' and
1131 containing () gave an internal compiling error instead of "reference to
1132 non-existent subpattern". Fortunately, when the pattern did exist, the
1133 compiled code was correct. (When scanning forwards to check for the
1134 existencd of the subpattern, it was treating the data ']' as terminating
1135 the class, so got the count wrong. When actually compiling, the reference
1136 was subsequently set up correctly.)
1137
1138 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1139 it was being rejected as not supported by pcre_dfa_exec(), even though
1140 other assertions are supported. I have made pcre_dfa_exec() support
1141 (*FAIL).
1142
1143 16. The implementation of 13c above involved the invention of a new opcode,
1144 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1145 cannot be changed at match time, I realized I could make a small
1146 improvement to matching performance by compiling OP_ALLANY instead of
1147 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1148 on the OP_ANY path.
1149
1150 17. Compiling pcretest on Windows with readline support failed without the
1151 following two fixes: (1) Make the unistd.h include conditional on
1152 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1153
1154 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1155 ncurses library to be included for pcretest when ReadLine support is
1156 requested, but also to allow for it to be overridden. This patch came from
1157 Daniel Bergström.
1158
1159 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1160 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1161 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1162 spotting this.
1163
1164
1165 Version 7.6 28-Jan-08
1166 ---------------------
1167
1168 1. A character class containing a very large number of characters with
1169 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1170 overflow.
1171
1172 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1173 HAVE_LONG_LONG is not defined.
1174
1175 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1176 bring them up to date and include new features. This patch includes:
1177
1178 - Fixed PH's badly added libz and libbz2 support.
1179 - Fixed a problem with static linking.
1180 - Added pcredemo. [But later removed - see 7 below.]
1181 - Fixed dftables problem and added an option.
1182 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1183 HAVE_LONG_LONG.
1184 - Added readline support for pcretest.
1185 - Added an listing of the option settings after cmake has run.
1186
1187 4. A user submitted a patch to Makefile that makes it easy to create
1188 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1189 Makefile.am that cause it to include this special target, without
1190 affecting anything else. Note that the same mingw target plus all
1191 the other distribution libraries and programs are now supported
1192 when configuring with CMake (see 6 below) instead of with
1193 Configure/Make.
1194
1195 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1196 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1197 exported in the Windows port". It has not yet been confirmed that the patch
1198 solves the problem, but it does no harm.
1199
1200 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1201 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1202 with CMake, and also correct the comment about stack recursion.
1203
1204 7. Remove the automatic building of pcredemo from the ./configure system and
1205 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1206 of a program that users should build themselves after PCRE is installed, so
1207 building it automatically is not really right. What is more, it gave
1208 trouble in some build environments.
1209
1210 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1211
1212
1213 Version 7.5 10-Jan-08
1214 ---------------------
1215
1216 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1217 values in parens when parsing an RE using the C++ wrapper."
1218
1219 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1220 Characters greater than 255 were excluded from the class instead of being
1221 included.
1222
1223 3. The same bug as (2) above applied to negated POSIX classes such as
1224 [:^space:].
1225
1226 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1227 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1228 I have changed it.
1229
1230 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1231 first named subpattern) and a construct such as (?&a) would reference the
1232 first named subpattern whose name started with "a" (in other words, the
1233 length check was missing). Both these problems are fixed. "Subpattern name
1234 expected" is now given for (?&) (a zero-length name), and this patch also
1235 makes it give the same error for \k'' (previously it complained that that
1236 was a reference to a non-existent subpattern).
1237
1238 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1239 this is right because (?- can be followed by option settings as well as by
1240 digits. I have, however, made the messages clearer.
1241
1242 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1243 than the number used in the conditional) now cause a compile-time error.
1244 This is actually not compatible with Perl, which accepts such patterns, but
1245 treats the conditional as always being FALSE (as PCRE used to), but it
1246 seems to me that giving a diagnostic is better.
1247
1248 8. Change "alphameric" to the more common word "alphanumeric" in comments
1249 and messages.
1250
1251 9. Fix two occurrences of "backslash" in comments that should have been
1252 "backspace".
1253
1254 10. Remove two redundant lines of code that can never be obeyed (their function
1255 was moved elsewhere).
1256
1257 11. The program that makes PCRE's Unicode character property table had a bug
1258 which caused it to generate incorrect table entries for sequences of
1259 characters that have the same character type, but are in different scripts.
1260 It amalgamated them into a single range, with the script of the first of
1261 them. In other words, some characters were in the wrong script. There were
1262 thirteen such cases, affecting characters in the following ranges:
1263
1264 U+002b0 - U+002c1
1265 U+0060c - U+0060d
1266 U+0061e - U+00612
1267 U+0064b - U+0065e
1268 U+0074d - U+0076d
1269 U+01800 - U+01805
1270 U+01d00 - U+01d77
1271 U+01d9b - U+01dbf
1272 U+0200b - U+0200f
1273 U+030fc - U+030fe
1274 U+03260 - U+0327f
1275 U+0fb46 - U+0fbb1
1276 U+10450 - U+1049d
1277
1278 12. The -o option (show only the matching part of a line) for pcregrep was not
1279 compatible with GNU grep in that, if there was more than one match in a
1280 line, it showed only the first of them. It now behaves in the same way as
1281 GNU grep.
1282
1283 13. If the -o and -v options were combined for pcregrep, it printed a blank
1284 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1285 does the same. The return code can be used to tell if there were any
1286 non-matching lines.
1287
1288 14. Added --file-offsets and --line-offsets to pcregrep.
1289
1290 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1291 infinitely looping recursion. The bug was that positive lookaheads were not
1292 being skipped when checking for a possible empty match (negative lookaheads
1293 and both kinds of lookbehind were skipped).
1294
1295 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1296 inclusion of <windows.h> to before rather than after the definition of
1297 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1298
1299 17. Specifying a possessive quantifier with a specific limit for a Unicode
1300 character property caused pcre_compile() to compile bad code, which led at
1301 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1302 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1303 caused the error; without that there was no problem.
1304
1305 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1306
1307 19. Added --enable-pcretest-libreadline.
1308
1309 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1310 RE::GlobalReplace(). As a result, the number of replacements returned was
1311 double what it should be. I removed one of the increments, but Craig sent a
1312 later patch that removed the other one (the right fix) and added unit tests
1313 that check the return values (which was not done before).
1314
1315 21. Several CMake things:
1316
1317 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1318 the names libpcre and libpcreposix, not just pcre and pcreposix.
1319
1320 (2) The above change means that pcretest and pcregrep are now correctly
1321 linked with the newly-built libraries, not previously installed ones.
1322
1323 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1324
1325 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1326 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1327 UTF-8 newline character). The key issue is that the pattern starts .*;
1328 this means that the match must be either at the beginning, or after a
1329 newline. The bug was in the code for advancing after a failed match and
1330 checking that the new position followed a newline. It was not taking
1331 account of UTF-8 characters correctly.
1332
1333 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1334 character classes. PCRE was not treating the sequence [:...:] as a
1335 character class unless the ... were all letters. Perl, however, seems to
1336 allow any characters between [: and :], though of course it rejects as
1337 unknown any "names" that contain non-letters, because all the known class
1338 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1339 for example, whereas PCRE did not - it did not recognize a POSIX character
1340 class. This seemed a bit dangerous, so the code has been changed to be
1341 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1342 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1343 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1344 Perl does, and where it didn't before.
1345
1346 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1347 Windows environments %n is disabled by default.
1348
1349
1350 Version 7.4 21-Sep-07
1351 ---------------------
1352
1353 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1354 means that a class such as [\s] counted as "explicit reference to CR or
1355 LF". That isn't really right - the whole point of the change was to try to
1356 help when there was an actual mention of one of the two characters. So now
1357 the change happens only if \r or \n (or a literal CR or LF) character is
1358 encountered.
1359
1360 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1361 of both had grown to the point where there were only 3 bits left.
1362 Fortunately, there was spare space in the data structure, and so I have
1363 moved the internal flags into a new 16-bit field to free up more option
1364 bits.
1365
1366 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1367 but did not set the internal JCHANGED flag - either of these is enough to
1368 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1369 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1370 start sets both bits.
1371
1372 4. Added options (at build time, compile time, exec time) to change \R from
1373 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1374
1375 5. doc/pcresyntax.html was missing from the distribution.
1376
1377 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1378 compatibility, even though it is no longer used.
1379
1380 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1381 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1382 windows.h file is present (where different names are used). [This was
1383 reversed later after testing - see 16 below.]
1384
1385 8. Changed all #include <config.h> to #include "config.h". There were also
1386 some further <pcre.h> cases that I changed to "pcre.h".
1387
1388 9. When pcregrep was used with the --colour option, it missed the line ending
1389 sequence off the lines that it output.
1390
1391 10. It was pointed out to me that arrays of string pointers cause lots of
1392 relocations when a shared library is dynamically loaded. A technique of
1393 using a single long string with a table of offsets can drastically reduce
1394 these. I have refactored PCRE in four places to do this. The result is
1395 dramatic:
1396
1397 Originally: 290
1398 After changing UCP table: 187
1399 After changing error message table: 43
1400 After changing table of "verbs" 36
1401 After changing table of Posix names 22
1402
1403 Thanks to the folks working on Gregex for glib for this insight.
1404
1405 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1406 unicode-properties was also set.
1407
1408 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1409
1410 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1411 checked only for CRLF.
1412
1413 14. Added casts to pcretest.c to avoid compiler warnings.
1414
1415 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1416
1417 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1418 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1419 entirely. This removes changes made in 7 above.
1420
1421 17. The CMake files have been updated, and there is now more information about
1422 building with CMake in the NON-UNIX-USE document.
1423
1424
1425 Version 7.3 28-Aug-07
1426 ---------------------
1427
1428 1. In the rejigging of the build system that eventually resulted in 7.1, the
1429 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1430 brackets there is not right, since it causes compilers to look for an
1431 installed pcre.h, not the version that is in the source that is being
1432 compiled (which of course may be different). I have changed it back to:
1433
1434 #include "pcre.h"
1435
1436 I have a vague recollection that the change was concerned with compiling in
1437 different directories, but in the new build system, that is taken care of
1438 by the VPATH setting the Makefile.
1439
1440 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1441 when the subject happened to end in the byte 0x85 (e.g. if the last
1442 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1443 characters but of course it shouldn't be taken as a newline when it is part
1444 of another character. The bug was that, for an unlimited repeat of . in
1445 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1446 characters when looking for a newline.
1447
1448 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1449
1450 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1451 in debug output.
1452
1453 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1454 long printing in the pcrecpp unittest when running under MinGW.
1455
1456 6. ESC_K was left out of the EBCDIC table.
1457
1458 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1459 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1460 limit also applies to "virtual nesting" when a pattern is recursive, and in
1461 this case 1000 isn't so big. I have been able to remove this limit at the
1462 expense of backing off one optimization in certain circumstances. Normally,
1463 when pcre_exec() would call its internal match() function recursively and
1464 immediately return the result unconditionally, it uses a "tail recursion"
1465 feature to save stack. However, when a subpattern that can match an empty
1466 string has an unlimited repetition quantifier, it no longer makes this
1467 optimization. That gives it a stack frame in which to save the data for
1468 checking that an empty string has been matched. Previously this was taken
1469 from the 1000-entry workspace that had been reserved. So now there is no
1470 explicit limit, but more stack is used.
1471
1472 8. Applied Daniel's patches to solve problems with the import/export magic
1473 syntax that is required for Windows, and which was going wrong for the
1474 pcreposix and pcrecpp parts of the library. These were overlooked when this
1475 problem was solved for the main library.
1476
1477 9. There were some crude static tests to avoid integer overflow when computing
1478 the size of patterns that contain repeated groups with explicit upper
1479 limits. As the maximum quantifier is 65535, the maximum group length was
1480 set at 30,000 so that the product of these two numbers did not overflow a
1481 32-bit integer. However, it turns out that people want to use groups that
1482 are longer than 30,000 bytes (though not repeat them that many times).
1483 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1484 made it possible to implement the integer overflow checks in a much more
1485 dynamic way, which I have now done. The artificial limitation on group
1486 length has been removed - we now have only the limit on the total length of
1487 the compiled pattern, which depends on the LINK_SIZE setting.
1488
1489 10. Fixed a bug in the documentation for get/copy named substring when
1490 duplicate names are permitted. If none of the named substrings are set, the
1491 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1492 empty string.
1493
1494 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1495 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1496 because the ] is interpreted as the first data character and the
1497 terminating ] is not found. PCRE has been made compatible with Perl in this
1498 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1499 cause memory overwriting.
1500
1501 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1502 string has been matched (to stop an infinite loop). It was not recognizing
1503 a conditional subpattern that could match an empty string if that
1504 subpattern was within another subpattern. For example, it looped when
1505 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1506 condition was not nested. This bug has been fixed.
1507
1508 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1509 past the start of the subject in the presence of bytes with the top bit
1510 set, for example "\x8aBCD".
1511
1512 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1513 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1514
1515 14. Optimized (?!) to (*FAIL).
1516
1517 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1518 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1519 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1520 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1521 does: it's just the validity check that is more restrictive.
1522
1523 16. Inserted checks for integer overflows during escape sequence (backslash)
1524 processing, and also fixed erroneous offset values for syntax errors during
1525 backslash processing.
1526
1527 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1528 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1529
1530 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1531 caused an overrun.
1532
1533 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1534 something other than just ASCII characters) inside a group that had an
1535 unlimited repeat caused a loop at compile time (while checking to see
1536 whether the group could match an empty string).
1537
1538 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1539 [\P{Any}] did so. (Error in the code for printing property names.)
1540
1541 21. An orphan \E inside a character class could cause a crash.
1542
1543 22. A repeated capturing bracket such as (A)? could cause a wild memory
1544 reference during compilation.
1545
1546 23. There are several functions in pcre_compile() that scan along a compiled
1547 expression for various reasons (e.g. to see if it's fixed length for look
1548 behind). There were bugs in these functions when a repeated \p or \P was
1549 present in the pattern. These operators have additional parameters compared
1550 with \d, etc, and these were not being taken into account when moving along
1551 the compiled data. Specifically:
1552
1553 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1554 length.
1555
1556 (b) An item such as \pL+ within a repeated group could cause crashes or
1557 loops.
1558
1559 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1560 "reference to non-existent subpattern" error.
1561
1562 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1563
1564 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1565 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1566
1567 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1568
1569 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1570 character were causing crashes (broken optimization).
1571
1572 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1573 \p or \P) caused a compile-time loop.
1574
1575 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1576 break. For example, the unstudied pattern [\r\n]A does not match the string
1577 "\r\nA" because change 7.0/46 below moves the current point on by two
1578 characters after failing to match at the start. However, the pattern \nA
1579 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1580 the same is true. There doesn't seem any very clean way out of this, but
1581 what I have chosen to do makes the common cases work: PCRE now takes note
1582 of whether there can be an explicit match for \r or \n anywhere in the
1583 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1584 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1585 pattern has explicit CR or LF references.
1586
1587 29. Added (*CR) etc for changing newline setting at start of pattern.
1588
1589
1590 Version 7.2 19-Jun-07
1591 ---------------------
1592
1593 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1594 which is apparently normally available under Windows.
1595
1596 2. Re-jig the pcregrep tests with different newline settings in an attempt
1597 to make them independent of the local environment's newline setting.
1598
1599 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1600
1601 4. Some of the "internals" tests were previously cut out when the link size
1602 was not 2, because the output contained actual offsets. The recent new
1603 "Z" feature of pcretest means that these can be cut out, making the tests
1604 usable with all link sizes.
1605
1606 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1607 stack recursion. This gives a massive performance boost under BSD, but just
1608 a small improvement under Linux. However, it saves one field in the frame
1609 in all cases.
1610
1611 6. Added more features from the forthcoming Perl 5.10:
1612
1613 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1614 recursion call. It refers to the nth most recently opened parentheses.
1615
1616 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1617 to be opened parentheses.
1618
1619 (c) Conditions that refer to capturing parentheses can be specified
1620 relatively, for example, (?(-2)... or (?(+3)...
1621
1622 (d) \K resets the start of the current match so that everything before
1623 is not part of it.
1624
1625 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1626
1627 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1628 reference syntax.
1629
1630 (g) (?| introduces a group in which the numbering of parentheses in each
1631 alternative starts with the same number.
1632
1633 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1634
1635 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1636 PCRE_INFO_JCHANGED.
1637
1638 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1639 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1640 for detecting groups that can match an empty string.
1641
1642 9. A pattern with a very large number of alternatives (more than several
1643 hundred) was running out of internal workspace during the pre-compile
1644 phase, where pcre_compile() figures out how much memory will be needed. A
1645 bit of new cunning has reduced the workspace needed for groups with
1646 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1647 workspace instead of running out of the 4096 that are available.
1648
1649 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1650
1651 11. Applied patch from Google to remove an optimization that didn't quite work.
1652 The report of the bug said:
1653
1654 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1655 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1656 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1657
1658 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1659 it matched the wrong number of bytes.
1660
1661
1662 Version 7.1 24-Apr-07
1663 ---------------------
1664
1665 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1666 that is more "standard", making use of automake and other Autotools. There
1667 is some re-arrangement of the files and adjustment of comments consequent
1668 on this.
1669
1670 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1671 for recursive directory scanning broke on some systems because the files
1672 are not scanned in any specific order and on different systems the order
1673 was different. A call to "sort" has been inserted into RunGrepTest for the
1674 approprate test as a short-term fix. In the longer term there may be an
1675 alternative.
1676
1677 3. I had an email from Eric Raymond about problems translating some of PCRE's
1678 man pages to HTML (despite the fact that I distribute HTML pages, some
1679 people do their own conversions for various reasons). The problems
1680 concerned the use of low-level troff macros .br and .in. I have therefore
1681 removed all such uses from the man pages (some were redundant, some could
1682 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1683 HTML has been updated to handle .nf/.fi and to complain if it encounters
1684 .br or .in.
1685
1686 4. Updated comments in configure.ac that get placed in config.h.in and also
1687 arranged for config.h to be included in the distribution, with the name
1688 config.h.generic, for the benefit of those who have to compile without
1689 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1690
1691 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1692 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1693 makevp.bat for the current PCRE, using the additional files
1694 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1695
1696 6. A Windows user reported a minor discrepancy with test 2, which turned out
1697 to be caused by a trailing space on an input line that had got lost in his
1698 copy. The trailing space was an accident, so I've just removed it.
1699
1700 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1701 that is needed.
1702
1703 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1704 as "const" (a) because they are and (b) because it helps the PHP
1705 maintainers who have recently made a script to detect big data structures
1706 in the php code that should be moved to the .rodata section. I remembered
1707 to update Builducptable as well, so it won't revert if ucptable.h is ever
1708 re-created.
1709
1710 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1711 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1712 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1713 support is not required. This saves 1.5-2K of code, which is important in
1714 some applications.
1715
1716 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1717 so as not to refer to the tables, even though these functions will never be
1718 called when UTF-8 support is disabled. Otherwise there are problems with a
1719 shared library.
1720
1721 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1722
1723 (a) It was defining its arguments as char * instead of void *.
1724
1725 (b) It was assuming that all moves were upwards in memory; this was true
1726 a long time ago when I wrote it, but is no longer the case.
1727
1728 The emulated memove() is provided for those environments that have neither
1729 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1730 is clearly not the case, as these two bugs were recently reported.
1731
1732 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1733 and Detrail to create the HTML documentation, the .txt form of the man
1734 pages, and it removes trailing spaces from listed files. It also creates
1735 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1736 case, it wraps all the #defines with #ifndefs. This script should be run
1737 before "make dist".
1738
1739 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1740 with Unicode property support.
1741
1742 (a) For a maximizing quantifier, if the two different cases of the
1743 character were of different lengths in their UTF-8 codings (there are
1744 some cases like this - I found 11), and the matching function had to
1745 back up over a mixture of the two cases, it incorrectly assumed they
1746 were both the same length.
1747
1748 (b) When PCRE was configured to use the heap rather than the stack for
1749 recursion during matching, it was not correctly preserving the data for
1750 the other case of a UTF-8 character when checking ahead for a match
1751 while processing a minimizing repeat. If the check also involved
1752 matching a wide character, but failed, corruption could cause an
1753 erroneous result when trying to check for a repeat of the original
1754 character.
1755
1756 13. Some tidying changes to the testing mechanism:
1757
1758 (a) The RunTest script now detects the internal link size and whether there
1759 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1760 values substituted by "configure". (The RunGrepTest script already did
1761 this for UTF-8.) The configure.ac script no longer substitutes the
1762 relevant variables.
1763
1764 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1765 with length and offset values. This means that the output is different
1766 for different internal link sizes. Test 2 is skipped for link sizes
1767 other than 2 because of this, bypassing the problem. Unfortunately,
1768 there was also a test in test 3 (the locale tests) that used /B and
1769 failed for link sizes other than 2. Rather than cut the whole test out,
1770 I have added a new /Z option to pcretest that replaces the length and
1771 offset values with spaces. This is now used to make test 3 independent
1772 of link size. (Test 2 will be tidied up later.)
1773
1774 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1775 segmentation fault instead of returning the appropriate error message.
1776
1777 15. In multiline mode when the newline sequence was set to "any", the pattern
1778 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1779 This doesn't seem right; it now treats the CRLF combination as the line
1780 ending, and so does not match in that case. It's only a pattern such as ^$
1781 that would hit this one: something like ^ABC$ would have failed after \r
1782 and then tried again after \r\n.
1783
1784 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1785 in an attempt to make files that differ only in their line terminators
1786 compare equal. This works on Linux.
1787
1788 17. Under certain error circumstances pcregrep might try to free random memory
1789 as it exited. This is now fixed, thanks to valgrind.
1790
1791 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1792 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1793 was because its rules for how to advance for /g after matching an empty
1794 string at the end of a line did not allow for this case. They now check for
1795 it specially.
1796
1797 20. pcretest is supposed to handle patterns and data of any length, by
1798 extending its buffers when necessary. It was getting this wrong when the
1799 buffer for a data line had to be extended.
1800
1801 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1802 CRLF as a newline sequence.
1803
1804 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1805 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1806 I have nevertheless tidied it up.
1807
1808 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1809
1810 24. Added a man page for pcre-config.
1811
1812
1813 Version 7.0 19-Dec-06
1814 ---------------------
1815
1816 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1817 moving to gcc 4.1.1.
1818
1819 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1820 sys/time.h, which is documented as needed for this function. It doesn't
1821 seem to matter on Linux, but it showed up on some releases of OS X.
1822
1823 3. It seems that there are systems where bytes whose values are greater than
1824 127 match isprint() in the "C" locale. The "C" locale should be the
1825 default when a C program starts up. In most systems, only ASCII printing
1826 characters match isprint(). This difference caused the output from pcretest
1827 to vary, making some of the tests fail. I have changed pcretest so that:
1828
1829 (a) When it is outputting text in the compiled version of a pattern, bytes
1830 other than 32-126 are always shown as hex escapes.
1831
1832 (b) When it is outputting text that is a matched part of a subject string,
1833 it does the same, unless a different locale has been set for the match
1834 (using the /L modifier). In this case, it uses isprint() to decide.
1835
1836 4. Fixed a major bug that caused incorrect computation of the amount of memory
1837 required for a compiled pattern when options that changed within the
1838 pattern affected the logic of the preliminary scan that determines the
1839 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1840 that the computed length was too small. The symptoms of this bug were
1841 either the PCRE error "internal error: code overflow" from pcre_compile(),
1842 or a glibc crash with a message such as "pcretest: free(): invalid next
1843 size (fast)". Examples of patterns that provoked this bug (shown in
1844 pcretest format) are:
1845
1846 /(?-x: )/x
1847 /(?x)(?-x: \s*#\s*)/
1848 /((?i)[\x{c0}])/8
1849 /(?i:[\x{c0}])/8
1850
1851 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1852 is now done differently.
1853
1854 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1855 wrapper classes; (b) implement a new function in the C++ scanner that is
1856 more efficient than the old way of doing things because it avoids levels of
1857 recursion in the regex matching; (c) add a paragraph to the documentation
1858 for the FullMatch() function.
1859
1860 6. The escape sequence \n was being treated as whatever was defined as
1861 "newline". Not only was this contrary to the documentation, which states
1862 that \n is character 10 (hex 0A), but it also went horribly wrong when
1863 "newline" was defined as CRLF. This has been fixed.
1864
1865 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1866 was being set to -1 for the "end of line" case (supposedly a value that no
1867 character can have). Though this value is never used (the check for end of
1868 line is "zero bytes in current character"), it caused compiler complaints.
1869 I've changed it to 0xffffffff.
1870
1871 8. In pcre_version.c, the version string was being built by a sequence of
1872 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1873 string (as it is for production releases) called a macro with an empty
1874 argument. The C standard says the result of this is undefined. The gcc
1875 compiler treats it as an empty string (which was what was wanted) but it is
1876 reported that Visual C gives an error. The source has been hacked around to
1877 avoid this problem.
1878
1879 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1880 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1881 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1882 of them did).
1883
1884 10. Originally, pcretest opened its input and output without "b"; then I was
1885 told that "b" was needed in some environments, so it was added for release
1886 5.0 to both the input and output. (It makes no difference on Unix-like
1887 systems.) Later I was told that it is wrong for the input on Windows. I've
1888 now abstracted the modes into two macros, to make it easier to fiddle with
1889 them, and removed "b" from the input mode under Windows.
1890
1891 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1892
1893 12. Added -help and --help to pcretest as an official way of being reminded
1894 of the options.
1895
1896 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1897 and pcrecpp.cc because they annoy compilers at high warning levels.
1898
1899 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1900
1901 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1902 scripts are not C programs :-) and which was not noticed because it works
1903 on Linux.
1904
1905 16. pcretest is supposed to handle any length of pattern and data line (as one
1906 line or as a continued sequence of lines) by extending its input buffer if
1907 necessary. This feature was broken for very long pattern lines, leading to
1908 a string of junk being passed to pcre_compile() if the pattern was longer
1909 than about 50K.
1910
1911 17. I have done a major re-factoring of the way pcre_compile() computes the
1912 amount of memory needed for a compiled pattern. Previously, there was code
1913 that made a preliminary scan of the pattern in order to do this. That was
1914 OK when PCRE was new, but as the facilities have expanded, it has become
1915 harder and harder to keep it in step with the real compile phase, and there
1916 have been a number of bugs (see for example, 4 above). I have now found a
1917 cunning way of running the real compile function in a "fake" mode that
1918 enables it to compute how much memory it would need, while actually only
1919 ever using a few hundred bytes of working memory and without too many
1920 tests of the mode. This should make future maintenance and development
1921 easier. A side effect of this work is that the limit of 200 on the nesting
1922 depth of parentheses has been removed (though this was never a serious
1923 limitation, I suspect). However, there is a downside: pcre_compile() now
1924 runs more slowly than before (30% or more, depending on the pattern). I
1925 hope this isn't a big issue. There is no effect on runtime performance.
1926
1927 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1928 newline (only possible for the last line of a file) and it was a
1929 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1930
1931 19. Added additional timing features to pcretest. (1) The -tm option now times
1932 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1933 separate command line item, by a number that specifies the number of
1934 repeats to use when timing. The default is 50000; this gives better
1935 precision, but takes uncomfortably long for very large patterns.
1936
1937 20. Extended pcre_study() to be more clever in cases where a branch of a
1938 subpattern has no definite first character. For example, (a*|b*)[cd] would
1939 previously give no result from pcre_study(). Now it recognizes that the
1940 first character must be a, b, c, or d.
1941
1942 21. There was an incorrect error "recursive call could loop indefinitely" if
1943 a subpattern (or the entire pattern) that was being tested for matching an
1944 empty string contained only one non-empty item after a nested subpattern.
1945 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1946 incorrectly, because the \d was being skipped in the check.
1947
1948 22. The pcretest program now has a new pattern option /B and a command line
1949 option -b, which is equivalent to adding /B to every pattern. This causes
1950 it to show the compiled bytecode, without the additional information that
1951 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1952 is the same as /B/I).
1953
1954 23. A new optimization is now able automatically to treat some sequences such
1955 as a*b as a*+b. More specifically, if something simple (such as a character
1956 or a simple class like \d) has an unlimited quantifier, and is followed by
1957 something that cannot possibly match the quantified thing, the quantifier
1958 is automatically "possessified".
1959
1960 24. A recursive reference to a subpattern whose number was greater than 39
1961 went wrong under certain circumstances in UTF-8 mode. This bug could also
1962 have affected the operation of pcre_study().
1963
1964 25. Realized that a little bit of performance could be had by replacing
1965 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1966
1967 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1968
1969 27. Possessive quantifiers such as a++ were previously implemented by turning
1970 them into atomic groups such as ($>a+). Now they have their own opcodes,
1971 which improves performance. This includes the automatically created ones
1972 from 23 above.
1973
1974 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1975 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1976 the first matched character to be a colon. This applied both to named and
1977 numbered groups.
1978
1979 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1980
1981 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1982 building PCRE on Windows easier, so I have included it in the distribution.
1983
1984 31. There is now a check in pcretest against a ridiculously large number being
1985 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1986 loop, the loop is abandoned.
1987
1988 32. Forward references to subpatterns in conditions such as (?(2)...) where
1989 subpattern 2 is defined later cause pcre_compile() to search forwards in
1990 the pattern for the relevant set of parentheses. This search went wrong
1991 when there were unescaped parentheses in a character class, parentheses
1992 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1993
1994 33. "Subroutine" calls and backreferences were previously restricted to
1995 referencing subpatterns earlier in the regex. This restriction has now
1996 been removed.
1997
1998 34. Added a number of extra features that are going to be in Perl 5.10. On the
1999 whole, these are just syntactic alternatives for features that PCRE had
2000 previously implemented using the Python syntax or my own invention. The
2001 other formats are all retained for compatibility.
2002
2003 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2004 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2005 also .NET compatible.
2006
2007 (b) A recursion or subroutine call to a named group can now be defined as
2008 (?&name) as well as (?P>name).
2009
2010 (c) A backreference to a named group can now be defined as \k<name> or
2011 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2012 5.10, are also .NET compatible.
2013
2014 (d) A conditional reference to a named group can now use the syntax
2015 (?(<name>) or (?('name') as well as (?(name).
2016
2017 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2018 groups (named and numbered) that are never evaluated inline, but can be
2019 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2020 is always false. There may be only one alternative in such a group.
2021
2022 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2023 as the simple (?(R). The condition is true only if the most recent
2024 recursion is that of the given number or name. It does not search out
2025 through the entire recursion stack.
2026
2027 (g) The escape \gN or \g{N} has been added, where N is a positive or
2028 negative number, specifying an absolute or relative reference.
2029
2030 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2031 some "unreachable code" warnings.
2032
2033 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2034 things, this adds five new scripts.
2035
2036 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2037 There were also incompatibilities regarding the handling of \Q..\E inside
2038 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2039 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2040
2041 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2042 matches an empty string, and forcibly breaks the loop. There were bugs in
2043 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2044 against aaaa the result was just "a" rather than "aaaa", for example. Two
2045 separate and independent bugs (that affected different cases) have been
2046 fixed.
2047
2048 39. Refactored the code to abolish the use of different opcodes for small
2049 capturing bracket numbers. This is a tidy that I avoided doing when I
2050 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2051 The new approach is not only tidier, it makes it possible to reduce the
2052 memory needed to fix the previous bug (38).
2053
2054 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2055 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2056 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2057 mode.
2058
2059 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2060 report.
2061
2062 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2063 copy construction and assignment in the C++ wrapper.
2064
2065 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2066 couple of bugs that could have given wrong results in the "--newline=crlf"
2067 case.
2068
2069 44. Added a number of casts and did some reorganization of signed/unsigned int
2070 variables following suggestions from Dair Grant. Also renamed the variable
2071 "this" as "item" because it is a C++ keyword.
2072
2073 45. Arranged for dftables to add
2074
2075 #include "pcre_internal.h"
2076
2077 to pcre_chartables.c because without it, gcc 4.x may remove the array
2078 definition from the final binary if PCRE is built into a static library and
2079 dead code stripping is activated.
2080
2081 46. For an unanchored pattern, if a match attempt fails at the start of a
2082 newline sequence, and the newline setting is CRLF or ANY, and the next two
2083 characters are CRLF, advance by two characters instead of one.
2084
2085
2086 Version 6.7 04-Jul-06
2087 ---------------------
2088
2089 1. In order to handle tests when input lines are enormously long, pcretest has
2090 been re-factored so that it automatically extends its buffers when
2091 necessary. The code is crude, but this _is_ just a test program. The
2092 default size has been increased from 32K to 50K.
2093
2094 2. The code in pcre_study() was using the value of the re argument before
2095 testing it for NULL. (Of course, in any sensible call of the function, it
2096 won't be NULL.)
2097
2098 3. The memmove() emulation function in pcre_internal.h, which is used on
2099 systems that lack both memmove() and bcopy() - that is, hardly ever -
2100 was missing a "static" storage class specifier.
2101
2102 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2103 containing an extended class (one that cannot be represented by a bitmap
2104 because it contains high-valued characters or Unicode property items, e.g.
2105 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2106 pattern, but PCRE should not loop if you do not (it no longer does).
2107 [Detail: two cases were found: (a) a repeated subpattern containing an
2108 extended class; (b) a recursive reference to a subpattern that followed a
2109 previous extended class. It wasn't skipping over the extended class
2110 correctly when UTF-8 mode was not set.]
2111
2112 5. A negated single-character class was not being recognized as fixed-length
2113 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2114 compile error "lookbehind assertion is not fixed length".
2115
2116 6. The RunPerlTest auxiliary script was showing an unexpected difference
2117 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2118 write a Perl script that can interpret lines of an input file either as
2119 byte characters or as UTF-8, which is what "perltest" was being required to
2120 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2121 can't do is switch easily at run time between having the "use utf8;" pragma
2122 or not. In the end, I fudged it by using the RunPerlTest script to insert
2123 "use utf8;" explicitly for the UTF-8 tests.
2124
2125 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2126 the end of the subject string, contrary to the documentation and to what
2127 Perl does. This was true of both matching functions. Now it matches only at
2128 the start of the subject and immediately after *internal* newlines.
2129
2130 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2131 a pointer to an int instead of a pointer to an unsigned long int. This
2132 caused problems on 64-bit systems.
2133
2134 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2135 instance of the 'standard' template library not being so standard".
2136
2137 10. There was no check on the number of named subpatterns nor the maximum
2138 length of a subpattern name. The product of these values is used to compute
2139 the size of the memory block for a compiled pattern. By supplying a very
2140 long subpattern name and a large number of named subpatterns, the size
2141 computation could be caused to overflow. This is now prevented by limiting
2142 the length of names to 32 characters, and the number of named subpatterns
2143 to 10,000.
2144
2145 11. Subpatterns that are repeated with specific counts have to be replicated in
2146 the compiled pattern. The size of memory for this was computed from the
2147 length of the subpattern and the repeat count. The latter is limited to
2148 65535, but there was no limit on the former, meaning that integer overflow
2149 could in principle occur. The compiled length of a repeated subpattern is
2150 now limited to 30,000 bytes in order to prevent this.
2151
2152 12. Added the optional facility to have named substrings with the same name.
2153
2154 13. Added the ability to use a named substring as a condition, using the
2155 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2156 are numbers (not recommended). Forward references are permitted.
2157
2158 14. Added forward references in named backreferences (if you see what I mean).
2159
2160 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2161 pattern could run off the end of the subject. For example, the pattern
2162 "(?s)(.{1,5})"8 did this with the subject "ab".
2163
2164 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2165 PCRE_CASELESS was set when matching characters that were quantified with ?
2166 or *.
2167
2168 17. A character class other than a single negated character that had a minimum
2169 but no maximum quantifier - for example [ab]{6,} - was not handled
2170 correctly by pce_dfa_exec(). It would match only one character.
2171
2172 18. A valid (though odd) pattern that looked like a POSIX character
2173 class but used an invalid character after [ (for example [[,abc,]]) caused
2174 pcre_compile() to give the error "Failed: internal error: code overflow" or
2175 in some cases to crash with a glibc free() error. This could even happen if
2176 the pattern terminated after [[ but there just happened to be a sequence of
2177 letters, a binary zero, and a closing ] in the memory that followed.
2178
2179 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2180 over the years. Originally (before any Unicode support), just the bottom 8
2181 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2182 output from "man perlunicode" includes this:
2183
2184 The regular expression compiler produces polymorphic opcodes. That
2185 is, the pattern adapts to the data and automatically switches to
2186 the Unicode character scheme when presented with Unicode data--or
2187 instead uses a traditional byte scheme when presented with byte
2188 data.
2189
2190 Sadly, a wide octal escape does not cause a switch, and in a string with
2191 no other multibyte characters, these octal escapes are treated as before.
2192 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2193 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2194 Unicode string.
2195
2196 I have not perpetrated such confusion in PCRE. Up till now, it took just
2197 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2198 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2199 translate to the appropriate multibyte character.
2200
2201 29. Applied some refactoring to reduce the number of warnings from Microsoft
2202 and Borland compilers. This has included removing the fudge introduced
2203 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2204 a warning about an unused variable.
2205
2206 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2207 characters since release 4.0, because Perl (from release 5.004) does not.
2208 [Or at least, is documented not to: some releases seem to be in conflict
2209 with the documentation.] However, when a pattern was studied with
2210 pcre_study() and all its branches started with \s, PCRE still included VT
2211 as a possible starting character. Of course, this did no harm; it just
2212 caused an unnecessary match attempt.
2213
2214 22. Removed a now-redundant internal flag bit that recorded the fact that case
2215 dependency changed within the pattern. This was once needed for "required
2216 byte" processing, but is no longer used. This recovers a now-scarce options
2217 bit. Also moved the least significant internal flag bit to the most-
2218 significant bit of the word, which was not previously used (hangover from
2219 the days when it was an int rather than a uint) to free up another bit for
2220 the future.
2221
2222 23. Added support for CRLF line endings as well as CR and LF. As well as the
2223 default being selectable at build time, it can now be changed at runtime
2224 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2225 specify that it is scanning data with non-default line endings.
2226
2227 24. Changed the definition of CXXLINK to make it agree with the definition of
2228 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2229
2230 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2231 recursions. This makes a big different to stack usage for some patterns.
2232
2233 26. If a subpattern containing a named recursion or subroutine reference such
2234 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2235 the space required for the compiled pattern went wrong and gave too small a
2236 value. Depending on the environment, this could lead to "Failed: internal
2237 error: code overflow at offset 49" or "glibc detected double free or
2238 corruption" errors.
2239
2240 27. Applied patches from Google (a) to support the new newline modes and (b) to
2241 advance over multibyte UTF-8 characters in GlobalReplace.
2242
2243 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2244 difference for some implementation of PCRE in some Windows version.
2245
2246 29. Added some extra testing facilities to pcretest:
2247
2248 \q<number> in a data line sets the "match limit" value
2249 \Q<number> in a data line sets the "match recursion limt" value
2250 -S <number> sets the stack size, where <number> is in megabytes
2251
2252 The -S option isn't available for Windows.
2253
2254
2255 Version 6.6 06-Feb-06
2256 ---------------------
2257
2258 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2259 in pcreposix.h. I have copied the definition from pcre.h.
2260
2261 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2262 because pcre.h is no longer a built file.
2263
2264 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2265 not normally included in the compiled code.
2266
2267
2268 Version 6.5 01-Feb-06
2269 ---------------------
2270
2271 1. When using the partial match feature with pcre_dfa_exec(), it was not
2272 anchoring the second and subsequent partial matches at the new starting
2273 point. This could lead to incorrect results. For example, with the pattern
2274 /1234/, partially matching against "123" and then "a4" gave a match.
2275
2276 2. Changes to pcregrep:
2277
2278 (a) All non-match returns from pcre_exec() were being treated as failures
2279 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2280 error message is output. Some extra information is given for the
2281 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2282 probably the only errors that are likely to be caused by users (by
2283 specifying a regex that has nested indefinite repeats, for instance).
2284 If there are more than 20 of these errors, pcregrep is abandoned.
2285
2286 (b) A binary zero was treated as data while matching, but terminated the
2287 output line if it was written out. This has been fixed: binary zeroes
2288 are now no different to any other data bytes.
2289
2290 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2291 used to set a locale for matching. The --locale=xxxx long option has
2292 been added (no short equivalent) to specify a locale explicitly on the
2293 pcregrep command, overriding the environment variables.
2294
2295 (d) When -B was used with -n, some line numbers in the output were one less
2296 than they should have been.
2297
2298 (e) Added the -o (--only-matching) option.
2299
2300 (f) If -A or -C was used with -c (count only), some lines of context were
2301 accidentally printed for the final match.
2302
2303 (g) Added the -H (--with-filename) option.
2304
2305 (h) The combination of options -rh failed to suppress file names for files
2306 that were found from directory arguments.
2307
2308 (i) Added the -D (--devices) and -d (--directories) options.
2309
2310 (j) Added the -F (--fixed-strings) option.
2311
2312 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2313
2314 (l) Added the --colo(u)r option.
2315
2316 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2317 is not present by default.
2318
2319 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2320 items such as (?R) or (?1), when the recursion could match a number of
2321 alternatives. If it matched one of the alternatives, but subsequently,
2322 outside the recursion, there was a failure, the code tried to back up into
2323 the recursion. However, because of the way PCRE is implemented, this is not
2324 possible, and the result was an incorrect result from the match.
2325
2326 In order to prevent this happening, the specification of recursion has
2327 been changed so that all such subpatterns are automatically treated as
2328 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2329
2330 4. I had overlooked the fact that, in some locales, there are characters for
2331 which isalpha() is true but neither isupper() nor islower() are true. In
2332 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2333 and ordfeminine) are like this. This affected the treatment of \w and \W
2334 when they appeared in character classes, but not when they appeared outside
2335 a character class. The bit map for "word" characters is now created
2336 separately from the results of isalnum() instead of just taking it from the
2337 upper, lower, and digit maps. (Plus the underscore character, of course.)
2338
2339 5. The above bug also affected the handling of POSIX character classes such as
2340 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2341 permanent tables. Instead, the bit maps for such a class were previously
2342 created as the appropriate unions of the upper, lower, and digit bitmaps.
2343 Now they are created by subtraction from the [[:word:]] class, which has
2344 its own bitmap.
2345
2346 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2347 It is created by subtracting the vertical space characters (\x09, \x0a,
2348 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2349 subtraction was done in the overall bitmap for a character class, meaning
2350 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2351 be recognized. This bug has been fixed.
2352
2353 7. Patches from the folks at Google:
2354
2355 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2356 real life, but is still worth protecting against".
2357
2358 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2359 regular expressions".
2360
2361 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2362 have it.
2363
2364 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2365 "configure" and the latter not, in order to fix a problem somebody had
2366 with compiling the Arg class on HP-UX.
2367
2368 (e) Improve the error-handling of the C++ wrapper a little bit.
2369
2370 (f) New tests for checking recursion limiting.
2371
2372 8. The pcre_memmove() function, which is used only if the environment does not
2373 have a standard memmove() function (and is therefore rarely compiled),
2374 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2375 returning a result (though PCRE never actually uses the result).
2376
2377 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2378 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2379 returned instead of calling malloc() with an overflowing number that would
2380 most likely cause subsequent chaos.
2381
2382 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2383
2384 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2385 with this option is matched, the nmatch and pmatch options of regexec() are
2386 ignored.
2387
2388 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2389 provided in case anyone wants to the the POSIX interface with UTF-8
2390 strings.
2391
2392 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2393 C++ linking (needed for some HP-UX environments).
2394
2395 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2396 (unused parameter) and in the pcre_printint() function (omitted "default"
2397 switch label when the default is to do nothing).
2398
2399 15. Added some code to make it possible, when PCRE is compiled as a C++
2400 library, to replace subject pointers for pcre_exec() with a smart pointer
2401 class, thus making it possible to process discontinuous strings.
2402
2403 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2404 much the same function. They were added by different people who were trying
2405 to make PCRE easy to compile on non-Unix systems. It has been suggested
2406 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2407 for compiling on Windows systems. I have therefore replaced it with
2408 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2409 defaults to "extern" for C or "extern C" for C++, which works fine on
2410 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2411 SCOPE with something explicit in config.h. In addition:
2412
2413 (a) pcreposix.h still had just "extern" instead of either of these macros;
2414 I have replaced it with PCRE_DATA_SCOPE.
2415
2416 (b) Functions such as _pcre_xclass(), which are internal to the library,
2417 but external in the C sense, all had PCRE_EXPORT in their definitions.
2418 This is apparently wrong for the Windows case, so I have removed it.
2419 (It makes no difference on Unix-like systems.)
2420
2421 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2422 of recursive calls to match(). This is different to MATCH_LIMIT because
2423 that limits the total number of calls to match(), not all of which increase
2424 the depth of recursion. Limiting the recursion depth limits the amount of
2425 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2426 when PCRE is compiled, and changed at run time. A patch from Google adds
2427 this functionality to the C++ interface.
2428
2429 18. Changes to the handling of Unicode character properties:
2430
2431 (a) Updated the table to Unicode 4.1.0.
2432
2433 (b) Recognize characters that are not in the table as "Cn" (undefined).
2434
2435 (c) I revised the way the table is implemented to a much improved format
2436 which includes recognition of ranges. It now supports the ranges that
2437 are defined in UnicodeData.txt, and it also amalgamates other
2438 characters into ranges. This has reduced the number of entries in the
2439 table from around 16,000 to around 3,000, thus reducing its size
2440 considerably. I realized I did not need to use a tree structure after
2441 all - a binary chop search is just as efficient. Having reduced the
2442 number of entries, I extended their size from 6 bytes to 8 bytes to
2443 allow for more data.
2444
2445 (d) Added support for Unicode script names via properties such as \p{Han}.
2446
2447 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2448 matching that character.
2449
2450 20. When matching a repeated Unicode property with a minimum greater than zero,
2451 (for example \pL{2,}), PCRE could look past the end of the subject if it
2452 reached it while seeking the minimum number of characters. This could
2453 happen only if some of the characters were more than one byte long, because
2454 there is a check for at least the minimum number of bytes.
2455
2456 21. Refactored the implementation of \p and \P so as to be more general, to
2457 allow for more different types of property in future. This has changed the
2458 compiled form incompatibly. Anybody with saved compiled patterns that use
2459 \p or \P will have to recompile them.
2460
2461 22. Added "Any" and "L&" to the supported property types.
2462
2463 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2464 but give a compile time error if the value is greater than 0xff.
2465
2466 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2467 accidentally not being installed or uninstalled.
2468
2469 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2470 made were to insert the current release number. This seemed silly, because
2471 it made things harder for people building PCRE on systems that don't run
2472 "configure". I have turned pcre.h into a distributed file, no longer built
2473 by "configure", with the version identification directly included. There is
2474 no longer a pcre.h.in file.
2475
2476 However, this change necessitated a change to the pcre-config script as
2477 well. It is built from pcre-config.in, and one of the substitutions was the
2478 release number. I have updated configure.ac so that ./configure now finds
2479 the release number by grepping pcre.h.
2480
2481 26. Added the ability to run the tests under valgrind.
2482
2483
2484 Version 6.4 05-Sep-05
2485 ---------------------
2486
2487 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2488 "--" to be printed when multiple files were scanned, even when none of the
2489 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2490 consider it to be a bug, and have restored the previous behaviour.
2491
2492 2. A couple of code tidies to get rid of compiler warnings.
2493
2494 3. The pcretest program used to cheat by referring to symbols in the library
2495 whose names begin with _pcre_. These are internal symbols that are not
2496 really supposed to be visible externally, and in some environments it is
2497 possible to suppress them. The cheating is now confined to including
2498 certain files from the library's source, which is a bit cleaner.
2499
2500 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2501 file's purpose clearer.
2502
2503 5. Reorganized pcre_ucp_findchar().
2504
2505
2506 Version 6.3 15-Aug-05
2507 ---------------------
2508
2509 1. The file libpcre.pc.in did not have general read permission in the tarball.
2510
2511 2. There were some problems when building without C++ support:
2512
2513 (a) If C++ support was not built, "make install" and "make test" still
2514 tried to test it.
2515
2516 (b) There were problems when the value of CXX was explicitly set. Some
2517 changes have been made to try to fix these, and ...
2518
2519 (c) --disable-cpp can now be used to explicitly disable C++ support.
2520
2521 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2522 backslash in a target when C++ was disabled. This confuses some
2523 versions of "make", apparently. Using an intermediate variable solves
2524 this. (Same for CPP_LOBJ.)
2525
2526 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2527 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2528 necessary on certain architectures.
2529
2530 4. Added a setting of -export-symbols-regex to the link command to remove
2531 those symbols that are exported in the C sense, but actually are local
2532 within the library, and not documented. Their names all begin with
2533 "_pcre_". This is not a perfect job, because (a) we have to except some
2534 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2535 available (and never for static libraries). I have made a note to try to
2536 find a way round (a) in the future.
2537
2538
2539 Version 6.2 01-Aug-05
2540 ---------------------
2541
2542 1. There was no test for integer overflow of quantifier values. A construction
2543 such as {1111111111111111} would give undefined results. What is worse, if
2544 a minimum quantifier for a parenthesized subpattern overflowed and became
2545 negative, the calculation of the memory size went wrong. This could have
2546 led to memory overwriting.
2547
2548 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2549
2550 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2551 operating environments where this matters.
2552
2553 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2554 PCRE options from within the C++ wrapper.
2555
2556 5. Named capturing subpatterns were not being correctly counted when a pattern
2557 was compiled. This caused two problems: (a) If there were more than 100
2558 such subpatterns, the calculation of the memory needed for the whole
2559 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2560 back references of the form \12, where the number was greater than 9, were
2561 not recognized as back references, even though there were sufficient
2562 previous subpatterns.
2563
2564 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2565 versions of gcc, e.g. 2.95.4.
2566
2567
2568 Version 6.1 21-Jun-05
2569 ---------------------
2570
2571 1. There was one reference to the variable "posix" in pcretest.c that was not
2572 surrounded by "#if !defined NOPOSIX".
2573
2574 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2575 the cross-check on the old pcre_info() function, for the benefit of the
2576 cut-down version of PCRE that is currently imported into Exim.
2577
2578 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2579 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2580 patterns that start (?i) but I don't think that matters. The use of (?i) is
2581 just an example; this all applies to the other options as well.
2582
2583 4. Since libtool seems to echo the compile commands it is issuing, the output
2584 from "make" can be reduced a bit by putting "@" in front of each libtool
2585 compile command.
2586
2587 5. Patch from the folks at Google for configure.in to be a bit more thorough
2588 in checking for a suitable C++ installation before trying to compile the
2589 C++ stuff. This should fix a reported problem when a compiler was present,
2590 but no suitable headers.
2591
2592 6. The man pages all had just "PCRE" as their title. I have changed them to
2593 be the relevant file name. I have also arranged that these names are
2594 retained in the file doc/pcre.txt, which is a concatenation in text format
2595 of all the man pages except the little individual ones for each function.
2596
2597 7. The NON-UNIX-USE file had not been updated for the different set of source
2598 files that come with release 6. I also added a few comments about the C++
2599 wrapper.
2600
2601
2602 Version 6.0 07-Jun-05
2603 ---------------------
2604
2605 1. Some minor internal re-organization to help with my DFA experiments.
2606
2607 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2608 didn't matter for the library itself when fully configured, but did matter
2609 when compiling without UCP support, or within Exim, where the ucp files are
2610 not imported.
2611
2612 3. Refactoring of the library code to split up the various functions into
2613 different source modules. The addition of the new DFA matching code (see
2614 below) to a single monolithic source would have made it really too
2615 unwieldy, quite apart from causing all the code to be include in a
2616 statically linked application, when only some functions are used. This is
2617 relevant even without the DFA addition now that patterns can be compiled in
2618 one application and matched in another.
2619
2620 The downside of splitting up is that there have to be some external
2621 functions and data tables that are used internally in different modules of
2622 the library but which are not part of the API. These have all had their
2623 names changed to start with "_pcre_" so that they are unlikely to clash
2624 with other external names.
2625
2626 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2627 a different (DFA) algorithm. Although it is slower than the original
2628 function, it does have some advantages for certain types of matching
2629 problem.
2630
2631 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2632 including restarting after a partial match.
2633
2634 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2635 defined when compiling for Windows was sent to me. I have put it into the
2636 code, though I have no means of testing or verifying it.
2637
2638 7. Added the pcre_refcount() auxiliary function.
2639
2640 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2641 match before or at the first newline in the subject string. In pcretest,
2642 the /f option on a pattern can be used to set this.
2643
2644 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2645 would behave wrongly. This has been present in PCRE since release 4.0.
2646
2647 10. A number of changes to the pcregrep command:
2648
2649 (a) Refactored how -x works; insert ^(...)$ instead of setting
2650 PCRE_ANCHORED and checking the length, in preparation for adding
2651 something similar for -w.
2652
2653 (b) Added the -w (match as a word) option.
2654
2655 (c) Refactored the way lines are read and buffered so as to have more
2656 than one at a time available.
2657
2658 (d) Implemented a pcregrep test script.
2659
2660 (e) Added the -M (multiline match) option. This allows patterns to match
2661 over several lines of the subject. The buffering ensures that at least
2662 8K, or the rest of the document (whichever is the shorter) is available
2663 for matching (and similarly the previous 8K for lookbehind assertions).
2664
2665 (f) Changed the --help output so that it now says
2666
2667 -w, --word-regex(p)
2668
2669 instead of two lines, one with "regex" and the other with "regexp"
2670 because that confused at least one person since the short forms are the
2671 same. (This required a bit of code, as the output is generated
2672 automatically from a table. It wasn't just a text change.)
2673
2674 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2675 option but starts with a hyphen. Could be a pattern or a path name
2676 starting with a hyphen, for instance.
2677
2678 (h) "-" can be given as a file name to represent stdin.
2679
2680 (i) When file names are being printed, "(standard input)" is used for
2681 the standard input, for compatibility with GNU grep. Previously
2682 "<stdin>" was used.
2683
2684 (j) The option --label=xxx can be used to supply a name to be used for
2685 stdin when file names are being printed. There is no short form.
2686
2687 (k) Re-factored the options decoding logic because we are going to add
2688 two more options that take data. Such options can now be given in four
2689 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2690
2691 (l) Added the -A, -B, and -C options for requesting that lines of context
2692 around matches be printed.
2693
2694 (m) Added the -L option to print the names of files that do not contain
2695 any matching lines, that is, the complement of -l.
2696
2697 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2698 continue to scan other files.
2699
2700 (o) The -s option was incorrectly implemented. For compatibility with other
2701 greps, it now suppresses the error message for a non-existent or non-
2702 accessible file (but not the return code). There is a new option called
2703 -q that suppresses the output of matching lines, which was what -s was
2704 previously doing.
2705
2706 (p) Added --include and --exclude options to specify files for inclusion
2707 and exclusion when recursing.
2708
2709 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2710 Hopefully, it now does.
2711
2712 12. Missing cast in pcre_study().
2713
2714 13. Added an "uninstall" target to the makefile.
2715
2716 14. Replaced "extern" in the function prototypes in Makefile.in with
2717 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2718 world, but is set differently for Windows.
2719
2720 15. Added a second compiling function called pcre_compile2(). The only
2721 difference is that it has an extra argument, which is a pointer to an
2722 integer error code. When there is a compile-time failure, this is set
2723 non-zero, in addition to the error test pointer being set to point to an
2724 error message. The new argument may be NULL if no error number is required
2725 (but then you may as well call pcre_compile(), which is now just a
2726 wrapper). This facility is provided because some applications need a
2727 numeric error indication, but it has also enabled me to tidy up the way
2728 compile-time errors are handled in the POSIX wrapper.
2729
2730 16. Added VPATH=.libs to the makefile; this should help when building with one
2731 prefix path and installing with another. (Or so I'm told by someone who
2732 knows more about this stuff than I do.)
2733
2734 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2735 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2736 match everything, including newlines. This is not POSIX-compatible, but
2737 somebody wanted the feature. From pcretest it can be activated by using
2738 both the P and the s flags.
2739
2740 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2741
2742 19. libpcre.pc was being incorrectly installed as executable.
2743
2744 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2745 it now also looks for '\r' so that it will work unmodified on Windows.
2746
2747 21. Added Google's contributed C++ wrapper to the distribution.
2748
2749 22. Added some untidy missing memory free() calls in pcretest, to keep
2750 Electric Fence happy when testing.
2751
2752
2753
2754 Version 5.0 13-Sep-04
2755 ---------------------
2756
2757 1. Internal change: literal characters are no longer packed up into items
2758 containing multiple characters in a single byte-string. Each character
2759 is now matched using a separate opcode. However, there may be more than one
2760 byte in the character in UTF-8 mode.
2761
2762 2. The pcre_callout_block structure has two new fields: pattern_position and
2763 next_item_length. These contain the offset in the pattern to the next match
2764 item, and its length, respectively.
2765
2766 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2767 insertion of callouts before each pattern item. Added the /C option to
2768 pcretest to make use of this.
2769
2770 4. On the advice of a Windows user, the lines
2771
2772 #if defined(_WIN32) || defined(WIN32)
2773 _setmode( _fileno( stdout ), 0x8000 );
2774 #endif /* defined(_WIN32) || defined(WIN32) */
2775
2776 have been added to the source of pcretest. This apparently does useful
2777 magic in relation to line terminators.
2778
2779 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2780 for the benefit of those environments where the "b" makes a difference.
2781
2782 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2783 to know about it. I have put a hack into configure.in that adds in code
2784 to set GCC=yes if CC=icc. This seems to end up at a point in the
2785 generated configure script that is early enough to affect the setting of
2786 compiler options, which is what is needed, but I have no means of testing
2787 whether it really works. (The user who reported this had patched the
2788 generated configure script, which of course I cannot do.)
2789
2790 LATER: After change 22 below (new libtool files), the configure script
2791 seems to know about icc (and also ecc). Therefore, I have commented out
2792 this hack in configure.in.
2793
2794 7. Added support for pkg-config (2 patches were sent in).
2795
2796 8. Negated POSIX character classes that used a combination of internal tables
2797 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2798 [[:^ascii]]. Typically, they would match almost any characters. The other
2799 POSIX classes were not broken in this way.
2800
2801 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2802 to find the match, as PCRE was deluded into thinking that the match had to
2803 start at the start point or following a newline. The same bug applied to
2804 patterns with negative forward assertions or any backward assertions
2805 preceding ".*" at the start, unless the pattern required a fixed first
2806 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2807
2808 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2809 starting at the last subject character, bytes beyond the end of the subject
2810 string were read.
2811
2812 11. Renamed the variable "class" as "classbits" to make life easier for C++
2813 users. (Previously there was a macro definition, but it apparently wasn't
2814 enough.)
2815
2816 12. Added the new field "tables" to the extra data so that tables can be passed
2817 in at exec time, or the internal tables can be re-selected. This allows
2818 a compiled regex to be saved and re-used at a later time by a different
2819 program that might have everything at different addresses.
2820
2821 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2822 -R library as well as a -L library.
2823
2824 14. The debugging options of pcretest (-d on the command line or D on a
2825 pattern) showed incorrect output for anything following an extended class
2826 that contained multibyte characters and which was followed by a quantifier.
2827
2828 15. Added optional support for general category Unicode character properties
2829 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2830 support. It adds about 90K to the size of the library. The meanings of the
2831 inbuilt class escapes such as \d and \s have NOT been changed.
2832
2833 16. Updated pcredemo.c to include calls to free() to release the memory for the
2834 compiled pattern.
2835
2836 17. The generated file chartables.c was being created in the source directory
2837 instead of in the building directory. This caused the build to fail if the
2838 source directory was different from the building directory, and was
2839 read-only.
2840
2841 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2842 file. No doubt somebody will tell me if they don't make sense... Also added
2843 Dan Mooney's comments about building on OpenVMS.
2844
2845 19. Added support for partial matching via the PCRE_PARTIAL option for
2846 pcre_exec() and the \P data escape in pcretest.
2847
2848 20. Extended pcretest with 3 new pattern features:
2849
2850 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2851 write the compiled pattern to the file whose name is "rest-of-line".
2852 This is a straight binary dump of the data, with the saved pointer to
2853 the character tables forced to be NULL. The study data, if any, is
2854 written too. After writing, pcretest reads a new pattern.
2855
2856 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2857 compiled pattern from the given file. There must not be any
2858 occurrences of "<" in the file name (pretty unlikely); if there are,
2859 pcretest will instead treat the initial "<" as a pattern delimiter.
2860 After reading in the pattern, pcretest goes on to read data lines as
2861 usual.
2862
2863 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2864 and 16-bit fields in a compiled pattern, to simulate a pattern that
2865 was compiled on a host of opposite endianness.
2866
2867 21. The pcre-exec() function can now cope with patterns that were compiled on
2868 hosts of opposite endianness, with this restriction:
2869
2870 As for any compiled expression that is saved and used later, the tables
2871 pointer field cannot be preserved; the extra_data field in the arguments
2872 to pcre_exec() should be used to pass in a tables address if a value
2873 other than the default internal tables were used at compile time.
2874
2875 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2876 now diagnosed as an error. Previously, most of the time, a negative number
2877 would have been treated as zero, but if in addition "ovector" was passed as
2878 NULL, a crash could occur.
2879
2880 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2881 new versions from the libtool 1.5 distribution (the last one is a copy of
2882 a file called libtool.m4). This seems to have fixed the need to patch
2883 "configure" to support Darwin 1.3 (which I used to do). However, I still
2884 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2885 workstation).
2886
2887 24. Changed the PCRE licence to be the more standard "BSD" licence.
2888
2889
2890 Version 4.5 01-Dec-03
2891 ---------------------
2892
2893 1. There has been some re-arrangement of the code for the match() function so
2894 that it can be compiled in a version that does not call itself recursively.
2895 Instead, it keeps those local variables that need separate instances for
2896 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2897 needs to "recurse". Keeping track of where control must go is done by means
2898 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2899 hide most of the details from the main code, and operates only if
2900 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2901 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2902 operating.
2903
2904 To make it easier for callers to provide specially tailored get/free
2905 functions for this usage, two new functions, pcre_stack_malloc, and
2906 pcre_stack_free, are used. They are always called in strict stacking order,
2907 and the size of block requested is always the same.
2908
2909 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2910 PCRE has been compiled to use the stack or the heap for recursion. The
2911 -C option of pcretest uses this to show which version is compiled.
2912
2913 A new data escape \S, is added to pcretest; it causes the amounts of store
2914 obtained and freed by both kinds of malloc/free at match time to be added
2915 to the output.
2916
2917 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2918 what's available on my current Linux desktop machine.
2919
2920 3. When matching a UTF-8 string, the test for a valid string at the start has
2921 been extended. If start_offset is not zero, PCRE now checks that it points
2922 to a byte that is the start of a UTF-8 character. If not, it returns
2923 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2924 this is necessary because there may be backward assertions in the pattern.
2925 When matching the same subject several times, it may save resources to use
2926 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2927
2928 4. The code for checking the validity of UTF-8 strings has been tightened so
2929 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2930 containing "overlong sequences".
2931
2932 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2933 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2934 should have been "&", but it just so happened that all the cases this let
2935 through by mistake were picked up later in the function.
2936
2937 6. I had used a variable called "isblank" - this is a C99 function, causing
2938 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2939
2940 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2941 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2942 so the version is shown; (c) stop "make test" echoing "./RunTest".
2943
2944 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2945
2946 9. The prototype for memmove() for systems that don't have it was using
2947 size_t, but the inclusion of the header that defines size_t was later. I've
2948 moved the #includes for the C headers earlier to avoid this.
2949
2950 10. Added some adjustments to the code to make it easier to compiler on certain
2951 special systems:
2952
2953 (a) Some "const" qualifiers were missing.
2954 (b) Added the macro EXPORT before all exported functions; by default this
2955 is defined to be empty.
2956 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2957 that it reads its output file name as an argument instead of writing
2958 to the standard output and assuming this can be redirected.
2959
2960 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2961 class containing characters with values greater than 255, PCRE compilation
2962 went into a loop.
2963
2964 12. A recursive reference to a subpattern that was within another subpattern
2965 that had a minimum quantifier of zero caused PCRE to crash. For example,
2966 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2967 recursion. If the recursively-called subpattern itself had a zero repeat,
2968 that was OK.
2969
2970 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2971 buffer into which it was copied (for escape processing) was still set at
2972 1024, so long lines caused crashes.
2973
2974 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2975 "internal error: code overflow...". This applied to any character class
2976 that was followed by a possessive quantifier.
2977
2978 15. Modified the Makefile to add libpcre.la as a prerequisite for
2979 libpcreposix.la because I was told this is needed for a parallel build to
2980 work.
2981
2982 16. If a pattern that contained .* following optional items at the start was
2983 studied, the wrong optimizing data was generated, leading to matching
2984 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2985 matching string must start with a or b or c. The correct conclusion for
2986 this pattern is that a match can start with any character.
2987
2988
2989 Version 4.4 13-Aug-03
2990 ---------------------
2991
2992 1. In UTF-8 mode, a character class containing characters with values between
2993 127 and 255 was not handled correctly if the compiled pattern was studied.
2994 In fixing this, I have also improved the studying algorithm for such
2995 classes (slightly).
2996
2997 2. Three internal functions had redundant arguments passed to them. Removal
2998 might give a very teeny performance improvement.
2999
3000 3. Documentation bug: the value of the capture_top field in a callout is *one
3001 more than* the number of the hightest numbered captured substring.
3002
3003 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3004 in incorrectly linking with a previously installed version. They now link
3005 explicitly with libpcre.la.
3006
3007 5. configure.in no longer needs to recognize Cygwin specially.
3008
3009 6. A problem in pcre.in for Windows platforms is fixed.
3010
3011 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3012 pcretest, it used to include the size of the study block as part of its
3013 output. Unfortunately, the structure contains a field that has a different
3014 size on different hardware architectures. This meant that the tests that
3015 showed this size failed. As the block is currently always of a fixed size,
3016 this information isn't actually particularly useful in pcretest output, so
3017 I have just removed it.
3018
3019 8. Three pre-processor statements accidentally did not start in column 1.
3020 Sadly, there are *still* compilers around that complain, even though
3021 standard C has not required this for well over a decade. Sigh.
3022
3023 9. In pcretest, the code for checking callouts passed small integers in the
3024 callout_data field, which is a void * field. However, some picky compilers
3025 complained about the casts involved for this on 64-bit systems. Now
3026 pcretest passes the address of the small integer instead, which should get
3027 rid of the warnings.
3028
3029 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3030 both compile and run time, and gives an error if an invalid UTF-8 sequence
3031 is found. There is a option for disabling this check in cases where the
3032 string is known to be correct and/or the maximum performance is wanted.
3033
3034 11. In response to a bug report, I changed one line in Makefile.in from
3035
3036 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3037 to
3038 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3039
3040 to look similar to other lines, but I have no way of telling whether this
3041 is the right thing to do, as I do not use Windows. No doubt I'll get told
3042 if it's wrong...
3043
3044
3045 Version 4.3 21-May-03
3046 ---------------------
3047
3048 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3049 Makefile.
3050
3051 2. Some refactoring to improve the quality of the code:
3052
3053 (i) The utf8_table... variables are now declared "const".
3054
3055 (ii) The code for \cx, which used the "case flipping" table to upper case
3056 lower case letters, now just substracts 32. This is ASCII-specific,
3057 but the whole concept of \cx is ASCII-specific, so it seems
3058 reasonable.
3059
3060 (iii) PCRE was using its character types table to recognize decimal and
3061 hexadecimal digits in the pattern. This is silly, because it handles
3062 only 0-9, a-f, and A-F, but the character types table is locale-
3063 specific, which means strange things might happen. A private
3064 table is now used for this - though it costs 256 bytes, a table is
3065 much faster than multiple explicit tests. Of course, the standard
3066 character types table is still used for matching digits in subject
3067 strings against \d.
3068
3069 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3070 ending in _t are). So I've renamed it as ESC_tee.
3071
3072 3. The first argument for regexec() in the POSIX wrapper should have been
3073 defined as "const".
3074
3075 4. Changed pcretest to use malloc() for its buffers so that they can be
3076 Electric Fenced for debugging.
3077
3078 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3079 to read one or more bytes before the start of the subject string. Often this
3080 had no effect on PCRE's behaviour, but in some circumstances it could
3081 provoke a segmentation fault.
3082
3083 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3084 to try to read one or more bytes before the start of the subject string.
3085
3086 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3087 UTF-8 support could misbehave in various ways if the subject string
3088 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3089 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3090 back over UTF-8 characters.)
3091
3092
3093 Version 4.2 14-Apr-03
3094 ---------------------
3095
3096 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3097
3098 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3099 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3100 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3101 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3102 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3103 and BUILD_EXEEXT
3104 Note: automatic setting of the BUILD variables is not yet working
3105 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3106 compile-time but not at link-time
3107 [LINK]: use for linking executables only
3108 make different versions for Windows and non-Windows
3109 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3110 libraries
3111 [LINK_FOR_BUILD]: new variable
3112 [OBJEXT]: use throughout
3113 [EXEEXT]: use throughout
3114 <winshared>: new target
3115 <wininstall>: new target
3116 <dftables.o>: use native compiler
3117 <dftables>: use native linker
3118 <install>: handle Windows platform correctly
3119 <clean>: ditto
3120 <check>: ditto
3121 copy DLL to top builddir before testing
3122
3123 As part of these changes, -no-undefined was removed again. This was reported
3124 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3125 in any case.
3126
3127 3. Some tidies to get rid of compiler warnings:
3128
3129 . In the match_data structure, match_limit was an unsigned long int, whereas
3130 match_call_count was an int. I've made them both unsigned long ints.
3131
3132 . In pcretest the fact that a const uschar * doesn't automatically cast to
3133 a void * provoked a warning.
3134
3135 . Turning on some more compiler warnings threw up some "shadow" variables
3136 and a few more missing casts.
3137
3138 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3139 option, a class that contained a single character with a value between 128
3140 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3141
3142 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3143 option, a class that contained several characters, but with at least one
3144 whose value was between 128 and 255 caused PCRE to crash.
3145
3146
3147 Version 4.1 12-Mar-03
3148 ---------------------
3149
3150 1. Compiling with gcc -pedantic found a couple of places where casts were
3151 needed, and a string in dftables.c that was longer than standard compilers are
3152 required to support.
3153
3154 2. Compiling with Sun's compiler found a few more places where the code could
3155 be tidied up in order to avoid warnings.
3156
3157 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3158 first of these names is deprecated in the latest Autoconf in favour of the name
3159 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3160 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3161 analogy I have changed it to CFLAGS_FOR_BUILD.
3162
3163 4. Added -no-undefined to the linking command in the Makefile, because this is
3164 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3165 linking step for the pcreposix library.
3166
3167 5. PCRE was failing to diagnose the case of two named groups with the same
3168 name.
3169
3170 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3171 literal character that is needed in the subject for a match, and scans along to
3172 ensure that it is present before embarking on the full matching process. This
3173 saves time in cases of nested unlimited repeats that are never going to match.
3174 Problem: the scan can take a lot of time if the subject is very long (e.g.
3175 megabytes), thus penalizing straightforward matches. It is now done only if the
3176 amount of subject to be scanned is less than 1000 bytes.
3177
3178 7. A lesser problem with the same optimization is that it was recording the
3179 first character of an anchored pattern as "needed", thus provoking a search
3180 right along the subject, even when the first match of the pattern was going to
3181 fail. The "needed" character is now not set for anchored patterns, unless it
3182 follows something in the pattern that is of non-fixed length. Thus, it still
3183 fulfils its original purpose of finding quick non-matches in cases of nested
3184 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3185
3186
3187 Version 4.0 17-Feb-03
3188 ---------------------
3189
3190 1. If a comment in an extended regex that started immediately after a meta-item
3191 extended to the end of string, PCRE compiled incorrect data. This could lead to
3192 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3193
3194 2. Moved to autoconf 2.53 and libtool 1.4.2.
3195
3196 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3197 the special perltest8 script is no longer needed - all the tests can be run
3198 from a single perltest script.
3199
3200 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3201 by \s. It has now been removed in PCRE. This means it isn't recognized as
3202 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3203 class [:space:] *does* include VT, thereby creating a mess.
3204
3205 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3206 space and tab.
3207
3208 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3209 its new features into the main test script, reducing the number of scripts.
3210
3211 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3212 were backward compatible, and made the (?i) apply to the whole pattern, as if
3213 /i were given. Now it behaves more logically, and applies the option setting
3214 only to what follows. PCRE has been changed to follow suit. However, if it
3215 finds options settings right at the start of the pattern, it extracts them into
3216 the global options, as before. Thus, they show up in the info data.
3217
3218 8. Added support for the \Q...\E escape sequence. Characters in between are
3219 treated as literals. This is slightly different from Perl in that $ and @ are
3220 also handled as literals inside the quotes. In Perl, they will cause variable
3221 interpolation. Note the following examples:
3222
3223 Pattern PCRE matches Perl matches
3224
3225 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3226 \Qabc\$xyz\E abc\$xyz abc\$xyz
3227 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3228
3229 For compatibility with Perl, \Q...\E sequences are recognized inside character
3230 classes as well as outside them.
3231
3232 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3233 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3234 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3235 signed/unsigned warnings.
3236
3237 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3238 option for pcretest, so I've replaced it by a simple function that does just
3239 that job.
3240
3241 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3242 "pcregrep -".
3243
3244 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3245 Java package. This provides some syntactic sugar for simple cases of what my
3246 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3247 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3248 item, you can use this simplified notation. Note that only makes sense with
3249 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3250 greediness, whatever the setting of the PCRE_UNGREEDY option.
3251
3252 13. A change of greediness default within a pattern was not taking effect at
3253 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3254 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3255 was abstracted outside.
3256
3257 14. PCRE now supports the \G assertion. It is true when the current matching
3258 position is at the start point of the match. This differs from \A when the
3259 starting offset is non-zero. Used with the /g option of pcretest (or similar
3260 code), it works in the same way as it does for Perl's /g option. If all
3261 alternatives of a regex begin with \G, the expression is anchored to the start
3262 match position, and the "anchored" flag is set in the compiled expression.
3263
3264 15. Some bugs concerning the handling of certain option changes within patterns
3265 have been fixed. These applied to options other than (?ims). For example,
3266 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3267 been the other way round. Some of this was related to change 7 above.
3268
3269 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3270 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3271 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3272 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3273
3274 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3275 mode. Unlike ".", it always matches newline, whatever the setting of
3276 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3277 assertions. Perl allows it, but it doesn't (in general) work because it can't
3278 calculate the length of the lookbehind. At least, that's the case for Perl
3279 5.8.0 - I've been told they are going to document that it doesn't work in
3280 future.
3281
3282 18. Added an error diagnosis for escapes that PCRE does not support: these are
3283 \L, \l, \N, \P, \p, \U, \u, and \X.
3284
3285 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3286 reading past the end of the pattern in cases such as /[abcd/.
3287
3288 20. PCRE was getting more memory than necessary for patterns with classes that
3289 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3290
3291 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3292 compiling PCRE for use with Virtual Pascal.
3293
3294 22. Small fix to the Makefile to make it work properly if the build is done
3295 outside the source tree.
3296
3297 23. Added a new extension: a condition to go with recursion. If a conditional
3298 subpattern starts with (?(R) the "true" branch is used if recursion has
3299 happened, whereas the "false" branch is used only at the top level.
3300
3301 24. When there was a very long string of literal characters (over 255 bytes
3302 without UTF support, over 250 bytes with UTF support), the computation of how
3303 much memory was required could be incorrect, leading to segfaults or other
3304 strange effects.
3305
3306 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3307 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3308 there was a subsequent back reference to those brackets. This meant that, for
3309 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3310 possible to check for precisely this case. All we can do is abandon the
3311 optimization if .* occurs inside capturing brackets when there are any back
3312 references whatsoever. (See below for a better fix that came later.)
3313
3314 26. The handling of the optimization for finding the first character of a
3315 non-anchored pattern, and for finding a character that is required later in the
3316 match were failing in some cases. This didn't break the matching; it just
3317 failed to optimize when it could. The way this is done has been re-implemented.
3318
3319 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3320
3321 28. Added a new feature that provides some of the functionality that Perl
3322 provides with (?{...}). The facility is termed a "callout". The way it is done
3323 in PCRE is for the caller to provide an optional function, by setting
3324 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3325 global variable. By default it is unset, which disables all calling out. To get
3326 the function called, the regex must include (?C) at appropriate points. This
3327 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3328 This provides a means of identifying different callout points. When PCRE
3329 reaches such a point in the regex, if pcre_callout has been set, the external
3330 function is called. It is provided with data in a structure called
3331 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3332 matching continues; if it returns a non-zero value, the match at the current
3333 point fails. However, backtracking will occur if possible. [This was changed
3334 later and other features added - see item 49 below.]
3335
3336 29. pcretest is upgraded to test the callout functionality. It provides a
3337 callout function that displays information. By default, it shows the start of
3338 the match and the current position in the text. There are some new data escapes
3339 to vary what happens:
3340
3341 \C+ in addition, show current contents of captured substrings
3342 \C- do not supply a callout function
3343 \C!n return 1 when callout number n is reached
3344 \C!n!m return 1 when callout number n is reached for the mth time
3345
3346 30. If pcregrep was called with the -l option and just a single file name, it
3347 output "<stdin>" if a match was found, instead of the file name.
3348
3349 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3350 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3351 pcre_exec(). This saves a malloc/free per call. The default value of
3352 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3353 when configuring.
3354
3355 32. The default maximum size of a compiled pattern is 64K. There have been a
3356 few cases of people hitting this limit. The code now uses macros to handle the
3357 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3358 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3359 configuring. Tests 2 and 5 work only with 2-byte links because they output
3360 debugging information about compiled patterns.
3361
3362 33. Internal code re-arrangements:
3363
3364 (a) Moved the debugging function for printing out a compiled regex into
3365 its own source file (printint.c) and used #include to pull it into
3366 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3367 separate copies.
3368
3369 (b) Defined the list of op-code names for debugging as a macro in
3370 internal.h so that it is next to the definition of the opcodes.
3371
3372 (c) Defined a table of op-code lengths for simpler skipping along compiled
3373 code. This is again a macro in internal.h so that it is next to the
3374 definition of the opcodes.
3375
3376 34. Added support for recursive calls to individual subpatterns, along the
3377 lines of Robin Houston's patch (but implemented somewhat differently).
3378
3379 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3380 allow it to read and process whole directories in Win32. This code was
3381 contributed by Lionel Fourquaux; it has not been tested by me.
3382
3383 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3384 used to name a group. Names consist of alphanumerics and underscores, and must
3385 be unique. Back references use the syntax (?P=name) and recursive calls use
3386 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3387 numbers. The function pcre_fullinfo() can be used after compilation to extract
3388 a name/number map. There are three relevant calls:
3389
3390 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3391 PCRE_INFO_NAMECOUNT yields the number of entries
3392 PCRE_INFO_NAMETABLE yields a pointer to the map.
3393
3394 The map is a vector of fixed-size entries. The size of each entry depends on
3395 the length of the longest name used. The first two bytes of each entry are the
3396 group number, most significant byte first. There follows the corresponding
3397 name, zero terminated. The names are in alphabetical order.
3398
3399 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3400 case instead of 255. Making it the same both with and without UTF-8 support
3401 means that the same test output works with both.
3402
3403 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3404 calling malloc() with a zero argument.
3405
3406 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3407 optimization. I've improved things by keeping a bitmap of backreferences with
3408 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3409 fact referenced, the optimization can be applied. It is unlikely that a
3410 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3411 the match to follow \n) will appear inside brackets with a number greater than
3412 31, but if it does, any back reference > 31 suppresses the optimization.
3413
3414 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3415 of disabling numbered capturing parentheses. Any opening parenthesis that is
3416 not followed by ? behaves as if it were followed by ?: but named parentheses
3417 can still be used for capturing (and they will acquire numbers in the usual
3418 way).
3419
3420 41. Redesigned the return codes from the match() function into yes/no/error so
3421 that errors can be passed back from deep inside the nested calls. A malloc
3422 failure while inside a recursive subpattern call now causes the
3423 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3424
3425 42. It is now possible to set a limit on the number of times the match()
3426 function is called in a call to pcre_exec(). This facility makes it possible to
3427 limit the amount of recursion and backtracking, though not in a directly
3428 obvious way, because the match() function is used in a number of different
3429 circumstances. The count starts from zero for each position in the subject
3430 string (for non-anchored patterns). The default limit is, for compatibility, a
3431 large number, namely 10 000 000. You can change this in two ways:
3432
3433 (a) When configuring PCRE before making, you can use --with-match-limit=n
3434 to set a default value for the compiled library.
3435
3436 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3437 a different value is set. See 45 below.
3438
3439 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3440
3441 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3442 of things that can be changed at compile time. The first argument specifies
3443 what is wanted and the second points to where the information is to be placed.
3444 The current list of available information is:
3445
3446 PCRE_CONFIG_UTF8
3447
3448 The output is an integer that is set to one if UTF-8 support is available;
3449 otherwise it is set to zero.
3450
3451 PCRE_CONFIG_NEWLINE
3452
3453 The output is an integer that it set to the value of the code that is used for
3454 newline. It is either LF (10) or CR (13).
3455
3456 PCRE_CONFIG_LINK_SIZE
3457
3458 The output is an integer that contains the number of bytes used for internal
3459 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3460
3461 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3462
3463 The output is an integer that contains the threshold above which the POSIX
3464 interface uses malloc() for output vectors. See item 31 above.
3465
3466 PCRE_CONFIG_MATCH_LIMIT
3467
3468 The output is an unsigned integer that contains the default limit of the number
3469 of match() calls in a pcre_exec() execution. See 42 above.
3470
3471 44. pcretest has been upgraded by the addition of the -C option. This causes it
3472 to extract all the available output from the new pcre_config() function, and to
3473 output it. The program then exits immediately.
3474
3475 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3476 order to support additional features. One way would have been to define
3477 pcre_exec2() (for example) with extra arguments, but this would not have been
3478 extensible, and would also have required all calls to the original function to
3479 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3480 is used for passing in "extra" data from pcre_study().
3481
3482 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3483 contains the following fields:
3484
3485 flags a bitmap indicating which of the following fields are set
3486 study_data opaque data from pcre_study()
3487 match_limit a way of specifying a limit on match() calls for a specific
3488 call to pcre_exec()
3489 callout_data data for callouts (see 49 below)
3490
3491 The flag bits are also defined in pcre.h, and are
3492
3493 PCRE_EXTRA_STUDY_DATA
3494 PCRE_EXTRA_MATCH_LIMIT
3495 PCRE_EXTRA_CALLOUT_DATA
3496
3497 The pcre_study() function now returns one of these new pcre_extra blocks, with
3498 the actual study data pointed to by the study_data field, and the
3499 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3500 before. That is, this change is entirely upwards-compatible and requires no
3501 change to existing code.
3502
3503 If you want to pass in additional data to pcre_exec(), you can either place it
3504 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3505 block.
3506
3507 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3508 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3509 times with different match limits, until it finds the minimum value needed for
3510 pcre_exec() to complete. The value is then output. This can be instructive; for
3511 most simple matches the number is quite small, but for pathological cases it
3512 gets very large very quickly.
3513
3514 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3515 returns the size of the data block pointed to by the study_data field in a
3516 pcre_extra block, that is, the value that was passed as the argument to
3517 pcre_malloc() when PCRE was getting memory in which to place the information
3518 created by pcre_study(). The fourth argument should point to a size_t variable.
3519 pcretest has been extended so that this information is shown after a successful
3520 pcre_study() call when information about the compiled regex is being displayed.
3521
3522 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3523 because what follows is always an absolute path. (Later: it turns out that this
3524 is more than cosmetic for MinGW, because it doesn't like empty path
3525 components.)
3526
3527 49. Some changes have been made to the callout feature (see 28 above):
3528
3529 (i) A callout function now has three choices for what it returns:
3530
3531 0 => success, carry on matching
3532 > 0 => failure at this point, but backtrack if possible
3533 < 0 => serious error, return this value from pcre_exec()
3534
3535 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3536 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3537 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3538 use by callout functions. It will never be used by PCRE itself.
3539
3540 (ii) The pcre_extra structure (see 45 above) has a void * field called
3541 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3542 pcre_callout_block structure has a field of the same name. The contents of
3543 the field passed in the pcre_extra structure are passed to the callout
3544 function in the corresponding field in the callout block. This makes it
3545 easier to use the same callout-containing regex from multiple threads. For
3546 testing, the pcretest program has a new data escape
3547
3548 \C*n pass the number n (may be negative) as callout_data
3549
3550 If the callout function in pcretest receives a non-zero value as
3551 callout_data, it returns that value.
3552
3553 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3554 there were some redundant $(CFLAGS) in commands that are now specified as
3555 $(LINK), which already includes $(CFLAGS).
3556
3557 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3558 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3559 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3560 one-byte characters throughout. Note that case-insensitive matching applies
3561 only to characters whose values are less than 256. PCRE doesn't support the
3562 notion of cases for higher-valued characters.
3563
3564 (i) A character class whose characters are all within 0-255 is handled as
3565 a bit map, and the map is inverted for negative classes. Previously, a
3566 character > 255 always failed to match such a class; however it should
3567 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3568
3569 (ii) A negated character class with a single character < 255 is coded as
3570 "not this character" (OP_NOT). This wasn't working properly when the test
3571 character was multibyte, either singly or repeated.
3572
3573 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3574 mode, for example: \x{100}{2,3}.
3575
3576 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3577 singly or repeated) now correctly test multibyte characters. However,
3578 PCRE doesn't recognize any characters with values greater than 255 as
3579 digits, spaces, or word characters. Such characters always match \D, \S,
3580 and \W, and never match \d, \s, or \w.
3581
3582 (v) Classes may now contain characters and character ranges with values
3583 greater than 255. For example: [ab\x{100}-\x{400}].
3584
3585 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3586 PCRE in UTF-8 mode.
3587
3588 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3589 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3590 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3591 value.)
3592
3593 53. The single man page has become too large. I have therefore split it up into
3594 a number of separate man pages. These also give rise to individual HTML pages;
3595 these are now put in a separate directory, and there is an index.html page that
3596 lists them all. Some hyperlinking between the pages has been installed.
3597
3598 54. Added convenience functions for handling named capturing parentheses.
3599
3600 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3601 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3602 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3603 are faulted.
3604
3605 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3606 calling configure. These values are used when compiling the dftables.c program
3607 which is run to generate the source of the default character tables. They
3608 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3609 you will need to set these values.
3610
3611 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3612
3613
3614 Version 3.9 02-Jan-02
3615 ---------------------
3616
3617 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3618
3619 2. If --disable-static was given, the building process failed when trying to
3620 build pcretest and pcregrep. (For some reason it was using libtool to compile
3621 them, which is not right, as they aren't part of the library.)
3622
3623
3624 Version 3.8 18-Dec-01
3625 ---------------------
3626
3627 1. The experimental UTF-8 code was completely screwed up. It was packing the
3628 bytes in the wrong order. How dumb can you get?
3629
3630
3631 Version 3.7 29-Oct-01
3632 ---------------------
3633
3634 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3635 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3636 this didn't happen under Solaris 8, where I normally test things.
3637
3638 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3639 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3640 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3641 this fix an hour or so after the initial 3.7 release.)
3642
3643
3644 Version 3.6 23-Oct-01
3645 ---------------------
3646
3647 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3648 offsets passed as NULL with zero offset count.
3649
3650 2. The config.guess and config.sub files had not been updated when I moved to
3651 the latest autoconf.
3652
3653
3654 Version 3.5 15-Aug-01
3655 ---------------------
3656
3657 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3658 had been forgotten.
3659
3660 2. By using declared but undefined structures, we can avoid using "void"
3661 definitions in pcre.h while keeping the internal definitions of the structures
3662 private.
3663
3664 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3665 user point of view, this means that both static and shared libraries are built
3666 by default, but this can be individually controlled. More of the work of
3667 handling this static/shared cases is now inside libtool instead of PCRE's make
3668 file.
3669
3670 4. The pcretest utility is now installed along with pcregrep because it is
3671 useful for users (to test regexs) and by doing this, it automatically gets
3672 relinked by libtool. The documentation has been turned into a man page, so
3673 there are now .1, .txt, and .html versions in /doc.
3674
3675 5. Upgrades to pcregrep:
3676 (i) Added long-form option names like gnu grep.
3677 (ii) Added --help to list all options with an explanatory phrase.
3678 (iii) Added -r, --recursive to recurse into sub-directories.
3679 (iv) Added -f, --file to read patterns from a file.
3680
3681 6. pcre_exec() was referring to its "code" argument before testing that
3682 argument for NULL (and giving an error if it was NULL).
3683
3684 7. Upgraded Makefile.in to allow for compiling in a different directory from
3685 the source directory.
3686
3687 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3688 options bits, the pointer it was passed was to an int instead of to an unsigned
3689 long int. This mattered only on 64-bit systems.
3690
3691 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3692 generated) instead of pcre.in, which it its source. Also made the same change
3693 in several of the .c files.
3694
3695 10. A new release of gcc defines printf() as a macro, which broke pcretest
3696 because it had an ifdef in the middle of a string argument for printf(). Fixed
3697 by using separate calls to printf().
3698
3699 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3700 script, to force use of CR or LF instead of \n in the source. On non-Unix
3701 systems, the value can be set in config.h.
3702
3703 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3704 absolute limit. Changed the text of the error message to make this clear, and
3705 likewise updated the man page.
3706
3707 13. The limit of 99 on the number of capturing subpatterns has been removed.
3708 The new limit is 65535, which I hope will not be a "real" limit.
3709
3710
3711 Version 3.4 22-Aug-00
3712 ---------------------
3713
3714 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3715
3716 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3717
3718
3719 Version 3.3 01-Aug-00
3720 ---------------------
3721
3722 1. If an octal character was given, but the value was greater than \377, it
3723 was not getting masked to the least significant bits, as documented. This could
3724 lead to crashes in some systems.
3725
3726 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3727 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3728
3729 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3730 These just pass their arguments on to (pcre_free)(), but they are provided
3731 because some uses of PCRE bind it to non-C systems that can call its functions,
3732 but cannot call free() or pcre_free() directly.
3733
3734 4. Add "make test" as a synonym for "make check". Corrected some comments in
3735 the Makefile.
3736
3737 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3738 Makefile.
3739
3740 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3741 command called pgrep for grepping around the active processes.
3742
3743 7. Added the beginnings of support for UTF-8 character strings.
3744
3745 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3746 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3747 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3748 out for the ar command.)
3749
3750
3751 Version 3.2 12-May-00
3752 ---------------------
3753
3754 This is purely a bug fixing release.
3755
3756 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3757 of ZA. This was just one example of several cases that could provoke this bug,
3758 which was introduced by change 9 of version 2.00. The code for breaking
3759 infinite loops after an iteration that matches an empty string was't working
3760 correctly.
3761
3762 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3763 when matched against abbab (for example). After matching an empty string, it
3764 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3765 caused it to match further down the string than it should.
3766
3767 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3768 was there because it doesn't seem to be needed, and it causes trouble on some
3769 systems, as it is not a Standard C header. It has been removed.
3770
3771 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3772 were reported on the Macintosh. The changes were from
3773
3774 while ((c = *(++ptr)) != 0 && c != '\n');
3775 to
3776 while ((c = *(++ptr)) != 0 && c != '\n') ;
3777
3778 Totally extraordinary, but if that's what it takes...
3779
3780 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3781 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3782 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3783 assumes the way PCRE uses memmove() (always moving upwards).
3784
3785 6. PCRE is being used in one environment where strchr() is not available. There
3786 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3787 faster code anyway.
3788
3789
3790 Version 3.1 09-Feb-00
3791 ---------------------
3792
3793 The only change in this release is the fixing of some bugs in Makefile.in for
3794 the "install" target:
3795
3796 (1) It was failing to install pcreposix.h.
3797
3798 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3799
3800
3801 Version 3.0 01-Feb-00
3802 ---------------------
3803
3804 1. Add support for the /+ modifier to perltest (to output $` like it does in
3805 pcretest).
3806
3807 2. Add support for the /g modifier to perltest.
3808
3809 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3810 matches null strings.
3811
3812 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3813 pattern. Perl treats empty patterns specially - it reuses the most recent
3814 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3815 effect.
3816
3817 5. The POSIX interface was broken in that it was just handing over the POSIX
3818 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3819 required a bigger vector, with some working space on the end. This means that
3820 the POSIX wrapper now has to get and free some memory, and copy the results.
3821
3822 6. Added some simple autoconf support, placing the test data and the
3823 documentation in separate directories, re-organizing some of the
3824 information files, and making it build pcre-config (a GNU standard). Also added
3825 libtool support for building PCRE as a shared library, which is now the
3826 default.
3827
3828 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3829 09 are not valid octal constants. Single digits will be used for minor values
3830 less than 10.
3831
3832 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3833 existing programs that set these in the POSIX interface can use PCRE without
3834 modification.
3835
3836 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3837 return all that pcre_info() returns, plus additional data. The pcre_info()
3838 function is retained for compatibility, but is considered to be obsolete.
3839
3840 10. Added experimental recursion feature (?R) to handle one common case that
3841 Perl 5.6 will be able to do with (?p{...}).
3842
3843 11. Added support for POSIX character classes like [:alpha:], which Perl is
3844 adopting.
3845
3846
3847 Version 2.08 31-Aug-99
3848 ----------------------
3849
3850 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3851 trying to match at the startoffset position, but instead was moving forward to
3852 the next newline as if a previous match had failed.
3853
3854 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3855 and could get into a loop if a null string was matched other than at the start
3856 of the subject.
3857
3858 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3859 be distinguished at compile time, and for completeness also added PCRE_DATE.
3860
3861 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3862 in GnuWin32 environments.
3863
3864
3865 Version 2.07 29-Jul-99
3866 ----------------------
3867
3868 1. The documentation is now supplied in plain text form and HTML as well as in
3869 the form of man page sources.
3870
3871 2. C++ compilers don't like assigning (void *) values to other pointer types.
3872 In particular this affects malloc(). Although there is no problem in Standard
3873 C, I've put in casts to keep C++ compilers happy.
3874
3875 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3876 should be (const char *).
3877
3878 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3879 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3880 However, I haven't made this a standard facility. The documentation doesn't
3881 mention it, and the Makefile doesn't support it.
3882
3883 5. The Makefile now contains an "install" target, with editable destinations at
3884 the top of the file. The pcretest program is not installed.
3885
3886 6. pgrep -V now gives the PCRE version number and date.
3887
3888 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3889 causing the entire string to be ignored, instead of just the last character.
3890
3891 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3892 non-matching string, it can take a very, very long time, even for strings of
3893 quite modest length, because of the nested recursion. PCRE now does better in
3894 some of these cases. It does this by remembering the last required literal
3895 character in the pattern, and pre-searching the subject to ensure it is present
3896 before running the real match. In other words, it applies a heuristic to detect
3897 some types of certain failure quickly, and in the above example, if presented
3898 with a string that has no trailing " it gives "no match" very quickly.
3899
3900 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3901 other alternatives are tried instead.
3902
3903
3904 Version 2.06 09-Jun-99
3905 ----------------------
3906
3907 1. Change pcretest's output for amount of store used to show just the code
3908 space, because the remainder (the data block) varies in size between 32-bit and
3909 64-bit systems.
3910
3911 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3912 start matching at. This allows lookbehinds to work when searching for multiple
3913 occurrences in a string.
3914
3915 3. Added additional options to pcretest for testing multiple occurrences:
3916
3917 /+ outputs the rest of the string that follows a match
3918 /g loops for multiple occurrences, using the new startoffset argument
3919 /G loops for multiple occurrences by passing an incremented pointer
3920
3921 4. PCRE wasn't doing the "first character" optimization for patterns starting
3922 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3923 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3924 the letter 'x'. On long subject strings, this gives a significant speed-up.
3925
3926
3927 Version 2.05 21-Apr-99
3928 ----------------------
3929
3930 1. Changed the type of magic_number from int to long int so that it works
3931 properly on 16-bit systems.
3932
3933 2. Fixed a bug which caused patterns starting with .* not to work correctly
3934 when the subject string contained newline characters. PCRE was assuming
3935 anchoring for such patterns in all cases, which is not correct because .* will
3936 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3937 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3938 must be retried after every newline in the subject.
3939
3940
3941 Version 2.04 18-Feb-99
3942 ----------------------
3943
3944 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3945 computation of the store needed to hold the pattern was incorrect (too large).
3946 If such patterns were nested a few deep, this could multiply and become a real
3947 problem.
3948
3949 2. Added /M option to pcretest to show the memory requirement of a specific
3950 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3951
3952 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3953 compiled in such a way that the backtracking after subsequent failure was
3954 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3955 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3956
3957
3958 Version 2.03 02-Feb-99
3959 ----------------------
3960
3961 1. Fixed typo and small mistake in man page.
3962
3963 2. Added 4th condition (GPL supersedes if conflict) and created separate
3964 LICENCE file containing the conditions.
3965
3966 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3967 Perl, that is the internal \ allows the delimiter to be included in the
3968 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3969 the final delimiter, add \ to the end of the pattern (to test the error).
3970
3971 4. Added the convenience functions for extracting substrings after a successful
3972 match. Updated pcretest to make it able to test these functions.
3973
3974
3975 Version 2.02 14-Jan-99
3976 ----------------------
3977
3978 1. Initialized the working variables associated with each extraction so that
3979 their saving and restoring doesn't refer to uninitialized store.
3980
3981 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3982 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3983 fix the problem.
3984
3985 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3986 calls, and wasn't printing the correct value for compiling calls. Increased the
3987 default value of LOOPREPEAT, and the number of significant figures in the
3988 times.
3989
3990 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3991
3992 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3993 a building problem on Windows NT with a FAT file system.
3994
3995
3996 Version 2.01 21-Oct-98
3997 ----------------------
3998
3999 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4000 to character tables built by pcre_maketables() in the current locale. If NULL
4001 is passed, the default tables are used.
4002
4003
4004 Version 2.00 24-Sep-98
4005 ----------------------
4006
4007 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4008 it any more.
4009
4010 2. Allow quantification of (?>) groups, and make it work correctly.
4011
4012 3. The first character computation wasn't working for (?>) groups.
4013
4014 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4015 end of the subject) and add 5.005's \z, which really does match only at the
4016 very end of the subject.
4017
4018 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4019
4020 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4021 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4022 localized options. All options to pcre_study() were also removed.
4023
4024 7. Add other new features from 5.005:
4025
4026 $(?<= positive lookbehind
4027 $(?<! negative lookbehind
4028 (?imsx-imsx) added the unsetting capability
4029 such a setting is global if at outer level; local otherwise
4030 (?imsx-imsx:) non-capturing groups with option setting
4031 (?(cond)re|re) conditional pattern matching
4032
4033 A backreference to itself in a repeated group matches the previous
4034 captured string.
4035
4036 8. General tidying up of studying (both automatic and via "study")
4037 consequential on the addition of new assertions.
4038
4039 9. As in 5.005, unlimited repeated groups that could match an empty substring
4040 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4041 runtime if any iteration does actually match an empty substring.
4042
4043 10. Include the RunTest script in the distribution.
4044
4045 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4046 discrepancies, some of which were old and were also with respect to 5.004. They
4047 have now been fixed.
4048
4049
4050 Version 1.09 28-Apr-98
4051 ----------------------
4052
4053 1. A negated single character class followed by a quantifier with a minimum
4054 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4055 program crashes, or just wrong answers. This did not apply to negated classes
4056 containing more than one character, or to minima other than one.
4057
4058
4059 Version 1.08 27-Mar-98
4060 ----------------------
4061
4062 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4063
4064 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4065 latter must appear before anything that relies on it in the pattern.
4066
4067
4068 Version 1.07 16-Feb-98
4069 ----------------------
4070
4071 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4072 repeat of a potentially empty string).
4073
4074
4075 Version 1.06 23-Jan-98
4076 ----------------------
4077
4078 1. Added Markus Oberhumer's little patches for C++.
4079
4080 2. Literal strings longer than 255 characters were broken.
4081
4082
4083 Version 1.05 23-Dec-97
4084 ----------------------
4085
4086 1. Negated character classes containing more than one character were failing if
4087 PCRE_CASELESS was set at run time.
4088
4089
4090 Version 1.04 19-Dec-97
4091 ----------------------
4092
4093 1. Corrected the man page, where some "const" qualifiers had been omitted.
4094
4095 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4096 input syntax.
4097
4098 3. Fixed memory leak which occurred when a regex with back references was
4099 matched with an offsets vector that wasn't big enough. The temporary memory
4100 that is used in this case wasn't being freed if the match failed.
4101
4102 4. Tidied pcretest to ensure it frees memory that it gets.
4103
4104 5. Temporary memory was being obtained in the case where the passed offsets
4105 vector was exactly big enough.
4106
4107 6. Corrected definition of offsetof() from change 5 below.
4108
4109 7. I had screwed up change 6 below and broken the rules for the use of
4110 setjmp(). Now fixed.
4111
4112
4113 Version 1.03 18-Dec-97
4114 ----------------------
4115
4116 1. A erroneous regex with a missing opening parenthesis was correctly
4117 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4118 on some systems.
4119
4120 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4121 it was reported that one broken compiler failed on the former because "code" is
4122 also an independent variable.
4123
4124 3. The erroneous regex a[]b caused an array overrun reference.
4125
4126 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4127 fail on data ending with that character. (It was going on too far, and checking
4128 the next character, typically a binary zero.) This was specific to the
4129 optimized code for single-character negative classes.
4130
4131 5. Added a contributed patch from the TIN world which does the following:
4132
4133 + Add an undef for memmove, in case the the system defines a macro for it.
4134
4135 + Add a definition of offsetof(), in case there isn't one. (I don't know
4136 the reason behind this - offsetof() is part of the ANSI standard - but
4137 it does no harm).
4138
4139 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4140 most of the places where whitespace preceded '#'. I have given up and
4141 allowed the remaining 2 cases to be at the margin.
4142
4143 + Rename some variables in pcre to eliminate shadowing. This seems very
4144 pedantic, but does no harm, of course.
4145
4146 6. Moved the call to setjmp() into its own function, to get rid of warnings
4147 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4148
4149 7. Constructs such as \d{8,} were compiling into the equivalent of
4150 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4151 outcome, but in this particular case used more store than had been allocated,
4152 which caused the bug to be discovered because it threw up an internal error.
4153
4154 8. The debugging code in both pcre and pcretest for outputting the compiled
4155 form of a regex was going wrong in the case of back references followed by
4156 curly-bracketed repeats.
4157
4158
4159 Version 1.02 12-Dec-97
4160 ----------------------
4161
4162 1. Typos in pcre.3 and comments in the source fixed.
4163
4164 2. Applied a contributed patch to get rid of places where it used to remove
4165 'const' from variables, and fixed some signed/unsigned and uninitialized
4166 variable warnings.
4167
4168 3. Added the "runtest" target to Makefile.
4169
4170 4. Set default compiler flag to -O2 rather than just -O.
4171
4172
4173 Version 1.01 19-Nov-97
4174 ----------------------
4175
4176 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4177 like /([ab]*)*/, that is, for classes with more than one character in them.
4178
4179 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4180 as /((?>a*))*/ (a PCRE_EXTRA facility).
4181
4182
4183 Version 1.00 18-Nov-97
4184 ----------------------
4185
4186 1. Added compile-time macros to support systems such as SunOS4 which don't have
4187 memmove() or strerror() but have other things that can be used instead.
4188
4189 2. Arranged that "make clean" removes the executables.
4190
4191
4192 Version 0.99 27-Oct-97
4193 ----------------------
4194
4195 1. Fixed bug in code for optimizing classes with only one character. It was
4196 initializing a 32-byte map regardless, which could cause it to run off the end
4197 of the memory it had got.
4198
4199 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4200
4201
4202 Version 0.98 22-Oct-97
4203 ----------------------
4204
4205 1. Fixed bug in code for handling temporary memory usage when there are more
4206 back references than supplied space in the ovector. This could cause segfaults.
4207
4208
4209 Version 0.97 21-Oct-97
4210 ----------------------
4211
4212 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4213
4214 2. Optimized negated single characters not to use a bit map.
4215
4216 3. Brought error texts together as macro definitions; clarified some of them;
4217 fixed one that was wrong - it said "range out of order" when it meant "invalid
4218 escape sequence".
4219
4220 4. Changed some char * arguments to const char *.
4221
4222 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4223
4224 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4225 pcretest.
4226
4227
4228 Version 0.96 16-Oct-97
4229 ----------------------
4230
4231 1. Added a simple "pgrep" utility to the distribution.
4232
4233 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4234 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4235 where "ddd" means "one or more decimal digits".
4236
4237 3. Fixed serious bug. If a pattern had a back reference, but the call to
4238 pcre_exec() didn't supply a large enough ovector to record the related
4239 identifying subpattern, the match always failed. PCRE now remembers the number
4240 of the largest back reference, and gets some temporary memory in which to save
4241 the offsets during matching if necessary, in order to ensure that
4242 backreferences always work.
4243
4244 4. Increased the compatibility with Perl in a number of ways:
4245
4246 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4247 to request this handling. The option can be set at compile or exec time.
4248
4249 (b) $ matches before a terminating newline by default; an option
4250 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4251 mode). The option can be set at compile or exec time.
4252
4253 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4254 the same as Perl's. If the decimal number it represents is less than 10
4255 or there aren't that many previous left capturing parentheses, an octal
4256 escape is read. Inside a character class, it's always an octal escape,
4257 even if it is a single digit.
4258
4259 (d) An escaped but undefined alphabetic character is taken as a literal,
4260 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4261 escapes.
4262
4263 (e) {0} is now permitted. (The previous item is removed from the compiled
4264 pattern).
4265
4266 5. Changed all the names of code files so that the basic parts are no longer
4267 than 10 characters, and abolished the teeny "globals.c" file.
4268
4269 6. Changed the handling of character classes; they are now done with a 32-byte
4270 bit map always.
4271
4272 7. Added the -d and /D options to pcretest to make it possible to look at the
4273 internals of compilation without having to recompile pcre.
4274
4275
4276 Version 0.95 23-Sep-97
4277 ----------------------
4278
4279 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4280 \x20 at the start of a run of normal characters. These were being treated as
4281 real characters, instead of the source characters being re-checked.
4282
4283
4284 Version 0.94 18-Sep-97
4285 ----------------------
4286
4287 1. The functions are now thread-safe, with the caveat that the global variables
4288 containing pointers to malloc() and free() or alternative functions are the
4289 same for all threads.
4290
4291 2. Get pcre_study() to generate a bitmap of initial characters for non-
4292 anchored patterns when this is possible, and use it if passed to pcre_exec().
4293
4294
4295 Version 0.93 15-Sep-97
4296 ----------------------
4297
4298 1. /(b)|(:+)/ was computing an incorrect first character.
4299
4300 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4301 but not actually doing anything yet.
4302
4303 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4304 as Perl does (e.g. [-az] or [az-]).
4305
4306 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4307 all possible positions.
4308
4309 5. Split up into different modules to avoid including unneeded functions in a
4310 compiled binary. However, compile and exec are still in one module. The "study"
4311 function is split off.
4312
4313 6. The character tables are now in a separate module whose source is generated
4314 by an auxiliary program - but can then be edited by hand if required. There are
4315 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4316 toupper() in the code.
4317
4318 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4319 make them global. Abolish the function for setting them, as the caller can now
4320 set them directly.
4321
4322
4323 Version 0.92 11-Sep-97
4324 ----------------------
4325
4326 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4327 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4328
4329 2. Caseless matching was not working in character classes if the characters in
4330 the pattern were in upper case.
4331
4332 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4333
4334 4. Make PCRE_ANCHORED public and accept as a compile option.
4335
4336 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4337 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4338 pass them.
4339
4340 6. Give an error if bad option bits passed at compile or run time.
4341
4342 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4343 pcretest to cause it to pass that flag.
4344
4345 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4346 options, and the first character, if set.
4347
4348 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4349
4350
4351 Version 0.91 10-Sep-97
4352 ----------------------
4353
4354 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4355 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4356
4357 2. PCRE was looping on encountering an indefinitely repeated back reference to
4358 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4359 Perl does - treats the match as successful.
4360
4361 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5