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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 609 - (show annotations)
Wed Jun 15 18:09:23 2011 UTC (8 years, 5 months ago) by ph10
File size: 203628 byte(s)
Fix bug with /\A.*?(?:a|b(*THEN)c)/ by removing the tail recursion optimization 
for the final branch. Also fix a similar bug for conditional subpatterns.
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 collection of tests with study enabled,
83 and thereby test pcre_study() more easily.
84
85 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
86 restoration of the capturing data to the outer values was not happening
87 correctly.
88
89 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
90 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
91 pattern had matched an empty string, and so incorrectly returned a no
92 match.
93
94 17. There was optimizing code for the last branch of non-capturing parentheses,
95 and also for the obeyed branch of a conditional subexpression, which used
96 tail recursion to cut down on stack usage. Unfortunately, not that there is
97 the possibility of (*THEN) occurring in these branches, tail recursion is
98 no longer possible because the return has to be checked for (*THEN). These
99 two optimizations have therefore been removed.
100
101
102 Version 8.12 15-Jan-2011
103 ------------------------
104
105 1. Fixed some typos in the markup of the man pages, and wrote a script that
106 checks for such things as part of the documentation building process.
107
108 2. On a big-endian 64-bit system, pcregrep did not correctly process the
109 --match-limit and --recursion-limit options (added for 8.11). In
110 particular, this made one of the standard tests fail. (The integer value
111 went into the wrong half of a long int.)
112
113 3. If the --colour option was given to pcregrep with -v (invert match), it
114 did strange things, either producing crazy output, or crashing. It should,
115 of course, ignore a request for colour when reporting lines that do not
116 match.
117
118 4. Another pcregrep bug caused similar problems if --colour was specified with
119 -M (multiline) and the pattern match finished with a line ending.
120
121 5. In pcregrep, when a pattern that ended with a literal newline sequence was
122 matched in multiline mode, the following line was shown as part of the
123 match. This seems wrong, so I have changed it.
124
125 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
126 the check for further matches in the same line (so they could be coloured)
127 to overrun the end of the current line. If another match was found, it was
128 incorrectly shown (and then shown again when found in the next line).
129
130 7. If pcregrep was compiled under Windows, there was a reference to the
131 function pcregrep_exit() before it was defined. I am assuming this was
132 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
133 reported by a user. I've moved the definition above the reference.
134
135
136 Version 8.11 10-Dec-2010
137 ------------------------
138
139 1. (*THEN) was not working properly if there were untried alternatives prior
140 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
141 backtracked to try for "b" instead of moving to the next alternative branch
142 at the same level (in this case, to look for "c"). The Perl documentation
143 is clear that when (*THEN) is backtracked onto, it goes to the "next
144 alternative in the innermost enclosing group".
145
146 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
147 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
148 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
149 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
150 (*THEN).
151
152 3. If \s appeared in a character class, it removed the VT character from
153 the class, even if it had been included by some previous item, for example
154 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
155 of \s, but is part of the POSIX "space" class.)
156
157 4. A partial match never returns an empty string (because you can always
158 match an empty string at the end of the subject); however the checking for
159 an empty string was starting at the "start of match" point. This has been
160 changed to the "earliest inspected character" point, because the returned
161 data for a partial match starts at this character. This means that, for
162 example, /(?<=abc)def/ gives a partial match for the subject "abc"
163 (previously it gave "no match").
164
165 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
166 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
167 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
168 has an implication that the given string is incomplete (because a partial
169 match is preferred over a full match). For this reason, these items now
170 give a partial match in this situation. [Aside: previously, the one case
171 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
172 match rather than a full match, which was wrong by the old rules, but is
173 now correct.]
174
175 6. There was a bug in the handling of #-introduced comments, recognized when
176 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
177 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
178 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
179 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
180 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
181 places in pcre_compile().
182
183 7. Related to (6) above, when pcre_compile() was skipping #-introduced
184 comments when looking ahead for named forward references to subpatterns,
185 the only newline sequence it recognized was NL. It now handles newlines
186 according to the set newline convention.
187
188 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
189 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
190 cater for a lack of strerror(). These oversights have been fixed.
191
192 9. Added --match-limit and --recursion-limit to pcregrep.
193
194 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
195
196 11. When the -o option was used, pcregrep was setting a return code of 1, even
197 when matches were found, and --line-buffered was not being honoured.
198
199 12. Added an optional parentheses number to the -o and --only-matching options
200 of pcregrep.
201
202 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
203 can match an empty string. The code to do it in pcretest and pcredemo
204 needed fixing:
205
206 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
207 only one byte after an empty string match just before CRLF (this case
208 just got forgotten; "any" and "anycrlf" were OK).
209
210 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
211 mode when an empty string match preceded an ASCII character followed by
212 a non-ASCII character. (The code for advancing by one character rather
213 than one byte was nonsense.)
214
215 (c) The pcredemo.c sample program did not have any code at all to handle
216 the cases when CRLF is a valid newline sequence.
217
218 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
219 as a starting offset was within the subject string. There is now a new
220 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
221 negative or greater than the length of the string. In order to test this,
222 pcretest is extended to allow the setting of negative starting offsets.
223
224 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
225 starting offset points to the beginning of a UTF-8 character was
226 unnecessarily clumsy. I tidied it up.
227
228 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
229 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
230
231 17. Nobody had reported that the --include_dir option, which was added in
232 release 7.7 should have been called --include-dir (hyphen, not underscore)
233 for compatibility with GNU grep. I have changed it to --include-dir, but
234 left --include_dir as an undocumented synonym, and the same for
235 --exclude-dir, though that is not available in GNU grep, at least as of
236 release 2.5.4.
237
238 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
239 characters from a string of bytes) have been redefined so as not to use
240 loops, in order to improve performance in some environments. At the same
241 time, I abstracted some of the common code into auxiliary macros to save
242 repetition (this should not affect the compiled code).
243
244 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
245 compile-time error is now given if \c is not followed by an ASCII
246 character, that is, a byte less than 128. (In EBCDIC mode, the code is
247 different, and any byte value is allowed.)
248
249 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
250 START_OPTIMIZE option, which is now allowed at compile time - but just
251 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
252 to pcregrep and other applications that have no direct access to PCRE
253 options. The new /Y option in pcretest sets this option when calling
254 pcre_compile().
255
256 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
257 back references. Groups containing recursive back references were forced to
258 be atomic by that change, but in the case of named groups, the amount of
259 memory required was incorrectly computed, leading to "Failed: internal
260 error: code overflow". This has been fixed.
261
262 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
263 pcretest.c, to avoid build problems in some Borland environments.
264
265
266 Version 8.10 25-Jun-2010
267 ------------------------
268
269 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
270 THEN.
271
272 2. (*ACCEPT) was not working when inside an atomic group.
273
274 3. Inside a character class, \B is treated as a literal by default, but
275 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
276 causes the error). The code is unchanged, but I tidied the documentation.
277
278 4. Inside a character class, PCRE always treated \R and \X as literals,
279 whereas Perl faults them if its -w option is set. I have changed PCRE so
280 that it faults them when PCRE_EXTRA is set.
281
282 5. Added support for \N, which always matches any character other than
283 newline. (It is the same as "." when PCRE_DOTALL is not set.)
284
285 6. When compiling pcregrep with newer versions of gcc which may have
286 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
287 declared with attribute warn_unused_result" were given. Just casting the
288 result to (void) does not stop the warnings; a more elaborate fudge is
289 needed. I've used a macro to implement this.
290
291 7. Minor change to pcretest.c to avoid a compiler warning.
292
293 8. Added four artifical Unicode properties to help with an option to make
294 \s etc use properties (see next item). The new properties are: Xan
295 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
296
297 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
298 use Unicode properties. (*UCP) at the start of a pattern can be used to set
299 this option. Modified pcretest to add /W to test this facility. Added
300 REG_UCP to make it available via the POSIX interface.
301
302 10. Added --line-buffered to pcregrep.
303
304 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
305 studied, and the match started with a letter with a code point greater than
306 127 whose first byte was different to the first byte of the other case of
307 the letter, the other case of this starting letter was not recognized
308 (#976).
309
310 12. If a pattern that was studied started with a repeated Unicode property
311 test, for example, \p{Nd}+, there was the theoretical possibility of
312 setting up an incorrect bitmap of starting bytes, but fortunately it could
313 not have actually happened in practice until change 8 above was made (it
314 added property types that matched character-matching opcodes).
315
316 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
317 possible starting bytes for non-anchored patterns.
318
319 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
320 \R, and also a number of cases that involve Unicode properties, both
321 explicit and implicit when PCRE_UCP is set.
322
323 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
324 input, it could crash or give wrong results if characters with values
325 greater than 0xc0 were present in the subject string. (Detail: it assumed
326 UTF-8 input when processing these items.)
327
328 16. Added a lot of (int) casts to avoid compiler warnings in systems where
329 size_t is 64-bit (#991).
330
331 17. Added a check for running out of memory when PCRE is compiled with
332 --disable-stack-for-recursion (#990).
333
334 18. If the last data line in a file for pcretest does not have a newline on
335 the end, a newline was missing in the output.
336
337 19. The default pcre_chartables.c file recognizes only ASCII characters (values
338 less than 128) in its various bitmaps. However, there is a facility for
339 generating tables according to the current locale when PCRE is compiled. It
340 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
341 characters, are recognized by isspace() and therefore were getting set in
342 these tables, and indeed these tables seem to approximate to ISO 8859. This
343 caused a problem in UTF-8 mode when pcre_study() was used to create a list
344 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
345 which of course cannot start UTF-8 characters. I have changed the code so
346 that only real ASCII characters (less than 128) and the correct starting
347 bytes for UTF-8 encodings are set for characters greater than 127 when in
348 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
349 altogether.)
350
351 20. Added the /T option to pcretest so as to be able to run tests with non-
352 standard character tables, thus making it possible to include the tests
353 used for 19 above in the standard set of tests.
354
355 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
356 reference to a subpattern the other side of a comment that contains an
357 opening parenthesis caused either an internal compiling error, or a
358 reference to the wrong subpattern.
359
360
361 Version 8.02 19-Mar-2010
362 ------------------------
363
364 1. The Unicode data tables have been updated to Unicode 5.2.0.
365
366 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
367 configured.
368
369 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
370 original author of that file, following a query about its status.
371
372 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
373 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
374
375 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
376 quantifier applied to a forward-referencing subroutine call, could compile
377 incorrect code or give the error "internal error: previously-checked
378 referenced subpattern not found".
379
380 6. Both MS Visual Studio and Symbian OS have problems with initializing
381 variables to point to external functions. For these systems, therefore,
382 pcre_malloc etc. are now initialized to local functions that call the
383 relevant global functions.
384
385 7. There were two entries missing in the vectors called coptable and poptable
386 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
387 I've fixed the data, and added a kludgy way of testing at compile time that
388 the lengths are correct (equal to the number of opcodes).
389
390 8. Following on from 7, I added a similar kludge to check the length of the
391 eint vector in pcreposix.c.
392
393 9. Error texts for pcre_compile() are held as one long string to avoid too
394 much relocation at load time. To find a text, the string is searched,
395 counting zeros. There was no check for running off the end of the string,
396 which could happen if a new error number was added without updating the
397 string.
398
399 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
400
401 11. \K was not working if it appeared in an atomic group or in a group that
402 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
403 \K is "not well defined" if used in an assertion. PCRE now accepts it if
404 the assertion is positive, but not if it is negative.
405
406 12. Change 11 fortuitously reduced the size of the stack frame used in the
407 "match()" function of pcre_exec.c by one pointer. Forthcoming
408 implementation of support for (*MARK) will need an extra pointer on the
409 stack; I have reserved it now, so that the stack frame size does not
410 decrease.
411
412 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
413 item in branch that calls a recursion is a subroutine call - as in the
414 second branch in the above example - was incorrectly given the compile-
415 time error "recursive call could loop indefinitely" because pcre_compile()
416 was not correctly checking the subroutine for matching a non-empty string.
417
418 14. The checks for overrunning compiling workspace could trigger after an
419 overrun had occurred. This is a "should never occur" error, but it can be
420 triggered by pathological patterns such as hundreds of nested parentheses.
421 The checks now trigger 100 bytes before the end of the workspace.
422
423 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
424
425
426 Version 8.01 19-Jan-2010
427 ------------------------
428
429 1. If a pattern contained a conditional subpattern with only one branch (in
430 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
431 computed the wrong minimum data length (which is of course zero for such
432 subpatterns). This could cause incorrect "no match" results.
433
434 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
435 the pattern is reset in the first branch, pcre_compile() failed with
436 "internal error: code overflow at offset...". This happened only when
437 the reset was to the original external option setting. (An optimization
438 abstracts leading options settings into an external setting, which was the
439 cause of this.)
440
441 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
442 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
443 assertion pattern did not match (meaning that the assertion was true), it
444 was incorrectly treated as false if the SKIP had been reached during the
445 matching. This also applied to assertions used as conditions.
446
447 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
448 assertion subpattern, including such a pattern used as a condition,
449 unpredictable results occurred, instead of the error return
450 PCRE_ERROR_DFA_UITEM.
451
452 5. The C++ GlobalReplace function was not working like Perl for the special
453 situation when an empty string is matched. It now does the fancy magic
454 stuff that is necessary.
455
456 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
457 removed. (These were left over from very, very early versions of PCRE.)
458
459 7. Some cosmetic changes to the code to make life easier when compiling it
460 as part of something else:
461
462 (a) Change DEBUG to PCRE_DEBUG.
463
464 (b) In pcre_compile(), rename the member of the "branch_chain" structure
465 called "current" as "current_branch", to prevent a collision with the
466 Linux macro when compiled as a kernel module.
467
468 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
469 prevent a collision with the Linux macro when compiled as a kernel
470 module.
471
472 8. In pcre_compile() there are some checks for integer overflows that used to
473 cast potentially large values to (double). This has been changed to that
474 when building, a check for int64_t is made, and if it is found, it is used
475 instead, thus avoiding the use of floating point arithmetic. (There is no
476 other use of FP in PCRE.) If int64_t is not found, the fallback is to
477 double.
478
479 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
480 2005 (difference between two addresses compared to an unsigned value).
481
482 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
483 custom one, because of the following reported problem in Windows:
484
485 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
486 under Win32.
487 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
488 therefore missing the function definition.
489 - The compiler thus generates a "C" signature for the test function.
490 - The linker fails to find the "C" function.
491 - PCRE fails to configure if asked to do so against libbz2.
492
493 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
494 messages were output:
495
496 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
497 rerunning libtoolize, to keep the correct libtool macros in-tree.
498 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
499
500 I have done both of these things.
501
502 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
503 most of the time, it *can* run out if it is given a pattern that contains a
504 runaway infinite recursion. I updated the discussion in the pcrestack man
505 page.
506
507 13. Now that we have gone to the x.xx style of version numbers, the minor
508 version may start with zero. Using 08 or 09 is a bad idea because users
509 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
510 interpreted as invalid octal numbers. I've updated the previous comment in
511 configure.ac, and also added a check that gives an error if 08 or 09 are
512 used.
513
514 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
515 causing partial matching to fail when the end of the subject matched \W
516 in a UTF-8 pattern where \W was quantified with a minimum of 3.
517
518 15. There were some discrepancies between the declarations in pcre_internal.h
519 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
520 their definitions. The declarations used "const uschar *" and the
521 definitions used USPTR. Even though USPTR is normally defined as "const
522 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
523 reported that: "This difference in casting confuses some C++ compilers, for
524 example, SunCC recognizes above declarations as different functions and
525 generates broken code for hbpcre." I have changed the declarations to use
526 USPTR.
527
528 16. GNU libtool is named differently on some systems. The autogen.sh script now
529 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
530 (FreeBSD).
531
532 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
533 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
534 comment: "Figure out how to create a longlong from a string: strtoll and
535 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
536 instance, but it only takes 2 args instead of 3!"
537
538 18. A subtle bug concerned with back references has been fixed by a change of
539 specification, with a corresponding code fix. A pattern such as
540 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
541 refers, was giving matches when it shouldn't. For example, xa=xaaa would
542 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
543 same bug. Such groups have to be quantified to be useful, or contained
544 inside another quantified group. (If there's no repetition, the reference
545 can never match.) The problem arises because, having left the group and
546 moved on to the rest of the pattern, a later failure that backtracks into
547 the group uses the captured value from the final iteration of the group
548 rather than the correct earlier one. I have fixed this in PCRE by forcing
549 any group that contains a reference to itself to be an atomic group; that
550 is, there cannot be any backtracking into it once it has completed. This is
551 similar to recursive and subroutine calls.
552
553
554 Version 8.00 19-Oct-09
555 ----------------------
556
557 1. The table for translating pcre_compile() error codes into POSIX error codes
558 was out-of-date, and there was no check on the pcre_compile() error code
559 being within the table. This could lead to an OK return being given in
560 error.
561
562 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
563 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
564 in a Windows environment.
565
566 3. The pcregrep --count option prints the count for each file even when it is
567 zero, as does GNU grep. However, pcregrep was also printing all files when
568 --files-with-matches was added. Now, when both options are given, it prints
569 counts only for those files that have at least one match. (GNU grep just
570 prints the file name in this circumstance, but including the count seems
571 more useful - otherwise, why use --count?) Also ensured that the
572 combination -clh just lists non-zero counts, with no names.
573
574 4. The long form of the pcregrep -F option was incorrectly implemented as
575 --fixed_strings instead of --fixed-strings. This is an incompatible change,
576 but it seems right to fix it, and I didn't think it was worth preserving
577 the old behaviour.
578
579 5. The command line items --regex=pattern and --regexp=pattern were not
580 recognized by pcregrep, which required --regex pattern or --regexp pattern
581 (with a space rather than an '='). The man page documented the '=' forms,
582 which are compatible with GNU grep; these now work.
583
584 6. No libpcreposix.pc file was created for pkg-config; there was just
585 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
586
587 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
588 when UCP support is not needed, by modifying the Python script that
589 generates it from Unicode data files. This should not matter if the module
590 is correctly used as a library, but I received one complaint about 50K of
591 unwanted data. My guess is that the person linked everything into his
592 program rather than using a library. Anyway, it does no harm.
593
594 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
595 was a minimum greater than 1 for a wide character in a possessive
596 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
597 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
598 character. Chaos in the form of incorrect output or a compiling loop could
599 result.
600
601 9. The restrictions on what a pattern can contain when partial matching is
602 requested for pcre_exec() have been removed. All patterns can now be
603 partially matched by this function. In addition, if there are at least two
604 slots in the offset vector, the offset of the earliest inspected character
605 for the match and the offset of the end of the subject are set in them when
606 PCRE_ERROR_PARTIAL is returned.
607
608 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
609 synonymous with PCRE_PARTIAL, for backwards compatibility, and
610 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
611 and may be more useful for multi-segment matching.
612
613 11. Partial matching with pcre_exec() is now more intuitive. A partial match
614 used to be given if ever the end of the subject was reached; now it is
615 given only if matching could not proceed because another character was
616 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
617 string "Z", which now yields "no match" instead of "partial match". In the
618 case of pcre_dfa_exec(), "no match" is given if every matching path for the
619 final character ended with (*FAIL).
620
621 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
622 if the pattern had a "must contain" character that was already found in the
623 earlier partial match, unless partial matching was again requested. For
624 example, with the pattern /dog.(body)?/, the "must contain" character is
625 "g". If the first part-match was for the string "dog", restarting with
626 "sbody" failed. This bug has been fixed.
627
628 13. The string returned by pcre_dfa_exec() after a partial match has been
629 changed so that it starts at the first inspected character rather than the
630 first character of the match. This makes a difference only if the pattern
631 starts with a lookbehind assertion or \b or \B (\K is not supported by
632 pcre_dfa_exec()). It's an incompatible change, but it makes the two
633 matching functions compatible, and I think it's the right thing to do.
634
635 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
636 so that the demonstration program is easily available in environments where
637 PCRE has not been installed from source.
638
639 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
640 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
641 library.
642
643 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
644 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
645 is not the first non-POSIX option to be added. Clearly some people find
646 these options useful.
647
648 17. If a caller to the POSIX matching function regexec() passes a non-zero
649 value for nmatch with a NULL value for pmatch, the value of
650 nmatch is forced to zero.
651
652 18. RunGrepTest did not have a test for the availability of the -u option of
653 the diff command, as RunTest does. It now checks in the same way as
654 RunTest, and also checks for the -b option.
655
656 19. If an odd number of negated classes containing just a single character
657 interposed, within parentheses, between a forward reference to a named
658 subpattern and the definition of the subpattern, compilation crashed with
659 an internal error, complaining that it could not find the referenced
660 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
661 [The bug was that it was starting one character too far in when skipping
662 over the character class, thus treating the ] as data rather than
663 terminating the class. This meant it could skip too much.]
664
665 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
666 /g option in pcretest when the pattern contains \K, which makes it possible
667 to have an empty string match not at the start, even when the pattern is
668 anchored. Updated pcretest and pcredemo to use this option.
669
670 21. If the maximum number of capturing subpatterns in a recursion was greater
671 than the maximum at the outer level, the higher number was returned, but
672 with unset values at the outer level. The correct (outer level) value is
673 now given.
674
675 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
676 PCRE did not set those parentheses (unlike Perl). I have now found a way to
677 make it do so. The string so far is captured, making this feature
678 compatible with Perl.
679
680 23. The tests have been re-organized, adding tests 11 and 12, to make it
681 possible to check the Perl 5.10 features against Perl 5.10.
682
683 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
684 pattern matches a fixed length string. PCRE did not allow this; now it
685 does. Neither allows recursion.
686
687 25. I finally figured out how to implement a request to provide the minimum
688 length of subject string that was needed in order to match a given pattern.
689 (It was back references and recursion that I had previously got hung up
690 on.) This code has now been added to pcre_study(); it finds a lower bound
691 to the length of subject needed. It is not necessarily the greatest lower
692 bound, but using it to avoid searching strings that are too short does give
693 some useful speed-ups. The value is available to calling programs via
694 pcre_fullinfo().
695
696 26. While implementing 25, I discovered to my embarrassment that pcretest had
697 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
698 study optimizations had never been tested with that matching function.
699 Oops. What is worse, even when it was passed study data, there was a bug in
700 pcre_dfa_exec() that meant it never actually used it. Double oops. There
701 were also very few tests of studied patterns with pcre_dfa_exec().
702
703 27. If (?| is used to create subpatterns with duplicate numbers, they are now
704 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
705 on the other side of the coin, they are no longer allowed to have different
706 names, because these cannot be distinguished in PCRE, and this has caused
707 confusion. (This is a difference from Perl.)
708
709 28. When duplicate subpattern names are present (necessarily with different
710 numbers, as required by 27 above), and a test is made by name in a
711 conditional pattern, either for a subpattern having been matched, or for
712 recursion in such a pattern, all the associated numbered subpatterns are
713 tested, and the overall condition is true if the condition is true for any
714 one of them. This is the way Perl works, and is also more like the way
715 testing by number works.
716
717
718 Version 7.9 11-Apr-09
719 ---------------------
720
721 1. When building with support for bzlib/zlib (pcregrep) and/or readline
722 (pcretest), all targets were linked against these libraries. This included
723 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
724 libraries. This caused unwanted dependencies to be created. This problem
725 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
726 pcretest is linked with readline.
727
728 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
729 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
730 moved outside it again, because FALSE and TRUE are already defined in AIX,
731 but BOOL is not.
732
733 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
734 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
735
736 4. The pcregrep documentation said spaces were inserted as well as colons (or
737 hyphens) following file names and line numbers when outputting matching
738 lines. This is not true; no spaces are inserted. I have also clarified the
739 wording for the --colour (or --color) option.
740
741 5. In pcregrep, when --colour was used with -o, the list of matching strings
742 was not coloured; this is different to GNU grep, so I have changed it to be
743 the same.
744
745 6. When --colo(u)r was used in pcregrep, only the first matching substring in
746 each matching line was coloured. Now it goes on to look for further matches
747 of any of the test patterns, which is the same behaviour as GNU grep.
748
749 7. A pattern that could match an empty string could cause pcregrep to loop; it
750 doesn't make sense to accept an empty string match in pcregrep, so I have
751 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
752 seems to be how GNU grep behaves.
753
754 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
755 start or after a newline", because the conditional assertion was not being
756 correctly handled. The rule now is that both the assertion and what follows
757 in the first alternative must satisfy the test.
758
759 9. If auto-callout was enabled in a pattern with a conditional group whose
760 condition was an assertion, PCRE could crash during matching, both with
761 pcre_exec() and pcre_dfa_exec().
762
763 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
764 used for matching.
765
766 11. Unicode property support in character classes was not working for
767 characters (bytes) greater than 127 when not in UTF-8 mode.
768
769 12. Added the -M command line option to pcretest.
770
771 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
772
773 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
774
775 16. Added comments and documentation about mis-use of no_arg in the C++
776 wrapper.
777
778 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
779 from Martin Jerabek that uses macro names for all relevant character and
780 string constants.
781
782 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
783 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
784 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
785 these, but not everybody uses configure.
786
787 19. A conditional group that had only one branch was not being correctly
788 recognized as an item that could match an empty string. This meant that an
789 enclosing group might also not be so recognized, causing infinite looping
790 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
791 with the subject "ab", where knowledge that the repeated group can match
792 nothing is needed in order to break the loop.
793
794 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
795 exec(), but without supplying a callout function, matching went wrong.
796
797 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
798 leak if the size of the offset vector was greater than 30. When the vector
799 is smaller, the saved offsets during recursion go onto a local stack
800 vector, but for larger vectors malloc() is used. It was failing to free
801 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
802 error, in fact).
803
804 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
805 heapframe that is used only when UTF-8 support is enabled. This caused no
806 problem, but was untidy.
807
808 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
809 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
810 included within another project.
811
812 24. Steven Van Ingelgem's patches to add more options to the CMake support,
813 slightly modified by me:
814
815 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
816 not building pcregrep.
817
818 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
819 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
820
821 25. Forward references, both numeric and by name, in patterns that made use of
822 duplicate group numbers, could behave incorrectly or give incorrect errors,
823 because when scanning forward to find the reference group, PCRE was not
824 taking into account the duplicate group numbers. A pattern such as
825 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
826
827 26. Changed a few more instances of "const unsigned char *" to USPTR, making
828 the feature of a custom pointer more persuasive (as requested by a user).
829
830 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
831 pcretest.c, inside #ifndefs, because it seems they are sometimes already
832 pre-defined.
833
834 28. Added support for (*UTF8) at the start of a pattern.
835
836 29. Arrange for flags added by the "release type" setting in CMake to be shown
837 in the configuration summary.
838
839
840 Version 7.8 05-Sep-08
841 ---------------------
842
843 1. Replaced UCP searching code with optimized version as implemented for Ad
844 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
845 stage table and inline lookup instead of a function, giving speed ups of 2
846 to 5 times on some simple patterns that I tested. Permission was given to
847 distribute the MultiStage2.py script that generates the tables (it's not in
848 the tarball, but is in the Subversion repository).
849
850 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
851 scripts.
852
853 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
854 a group with a zero qualifier. The result of the study could be incorrect,
855 or the function might crash, depending on the pattern.
856
857 4. Caseless matching was not working for non-ASCII characters in back
858 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
859 It now works when Unicode Property Support is available.
860
861 5. In pcretest, an escape such as \x{de} in the data was always generating
862 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
863 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
864 truncation.
865
866 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
867
868 7. Added two (int) casts to pcregrep when printing the difference of two
869 pointers, in case they are 64-bit values.
870
871 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
872 test 2 if it fails.
873
874 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
875 and a #define of that name to empty if it is not externally set. This is to
876 allow users of MSVC to set it if necessary.
877
878 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
879 the convenience functions in the pcre_get.c source file.
880
881 11. An option change at the start of a pattern that had top-level alternatives
882 could cause overwriting and/or a crash. This command provoked a crash in
883 some environments:
884
885 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
886
887 This potential security problem was recorded as CVE-2008-2371.
888
889 12. For a pattern where the match had to start at the beginning or immediately
890 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
891 pcre_dfa_exec() could read past the end of the passed subject if there was
892 no match. To help with detecting such bugs (e.g. with valgrind), I modified
893 pcretest so that it places the subject at the end of its malloc-ed buffer.
894
895 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
896 exec() might read past the end of the data buffer in UTF-8 mode.
897
898 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
899 the data contained the byte 0x85 as part of a UTF-8 character within its
900 first line. This applied both to normal and DFA matching.
901
902 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
903 /^[^d]*?$/8 failed to match "abc".
904
905 16. Added a missing copyright notice to pcrecpp_internal.h.
906
907 17. Make it more clear in the documentation that values returned from
908 pcre_exec() in ovector are byte offsets, not character counts.
909
910 18. Tidied a few places to stop certain compilers from issuing warnings.
911
912 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
913 supplied by Stefan Weber. I made a further small update for 7.8 because
914 there is a change of source arrangements: the pcre_searchfuncs.c module is
915 replaced by pcre_ucd.c.
916
917
918 Version 7.7 07-May-08
919 ---------------------
920
921 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
922 a string to a long long, pretend we don't even have a long long." This is
923 done by checking for the strtoq, strtoll, and _strtoi64 functions.
924
925 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
926 pre-7.6 versions, which defined a global no_arg variable instead of putting
927 it in the RE class. (See also #8 below.)
928
929 3. Remove a line of dead code, identified by coverity and reported by Nuno
930 Lopes.
931
932 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
933
934 (1) The include/exclude patterns were being applied to the whole pathnames
935 of files, instead of just to the final components.
936
937 (2) If there was more than one level of directory, the subdirectories were
938 skipped unless they satisfied the include/exclude conditions. This is
939 inconsistent with GNU grep (and could even be seen as contrary to the
940 pcregrep specification - which I improved to make it absolutely clear).
941 The action now is always to scan all levels of directory, and just
942 apply the include/exclude patterns to regular files.
943
944 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
945 --exclude_dir in the tests to avoid scanning .svn directories.
946
947 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
948 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
949 doesn't support NULs in patterns.
950
951 7. Added some missing "const"s to declarations of static tables in
952 pcre_compile.c and pcre_dfa_exec.c.
953
954 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
955 caused by fix #2 above. (Subsequently also a second patch to fix the
956 first patch. And a third patch - this was a messy problem.)
957
958 9. Applied Craig's patch to remove the use of push_back().
959
960 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
961 matching function regexec().
962
963 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
964 which, however, unlike Perl's \g{...}, are subroutine calls, not back
965 references. PCRE supports relative numbers with this syntax (I don't think
966 Oniguruma does).
967
968 12. Previously, a group with a zero repeat such as (...){0} was completely
969 omitted from the compiled regex. However, this means that if the group
970 was called as a subroutine from elsewhere in the pattern, things went wrong
971 (an internal error was given). Such groups are now left in the compiled
972 pattern, with a new opcode that causes them to be skipped at execution
973 time.
974
975 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
976 to the way PCRE behaves:
977
978 (a) A lone ] character is dis-allowed (Perl treats it as data).
979
980 (b) A back reference to an unmatched subpattern matches an empty string
981 (Perl fails the current match path).
982
983 (c) A data ] in a character class must be notated as \] because if the
984 first data character in a class is ], it defines an empty class. (In
985 Perl it is not possible to have an empty class.) The empty class []
986 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
987 The negative empty class [^] matches any one character, independently
988 of the DOTALL setting.
989
990 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
991 non-existent subpattern following a character class starting with ']' and
992 containing () gave an internal compiling error instead of "reference to
993 non-existent subpattern". Fortunately, when the pattern did exist, the
994 compiled code was correct. (When scanning forwards to check for the
995 existencd of the subpattern, it was treating the data ']' as terminating
996 the class, so got the count wrong. When actually compiling, the reference
997 was subsequently set up correctly.)
998
999 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1000 it was being rejected as not supported by pcre_dfa_exec(), even though
1001 other assertions are supported. I have made pcre_dfa_exec() support
1002 (*FAIL).
1003
1004 16. The implementation of 13c above involved the invention of a new opcode,
1005 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1006 cannot be changed at match time, I realized I could make a small
1007 improvement to matching performance by compiling OP_ALLANY instead of
1008 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1009 on the OP_ANY path.
1010
1011 17. Compiling pcretest on Windows with readline support failed without the
1012 following two fixes: (1) Make the unistd.h include conditional on
1013 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1014
1015 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1016 ncurses library to be included for pcretest when ReadLine support is
1017 requested, but also to allow for it to be overridden. This patch came from
1018 Daniel Bergström.
1019
1020 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1021 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1022 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1023 spotting this.
1024
1025
1026 Version 7.6 28-Jan-08
1027 ---------------------
1028
1029 1. A character class containing a very large number of characters with
1030 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1031 overflow.
1032
1033 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1034 HAVE_LONG_LONG is not defined.
1035
1036 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1037 bring them up to date and include new features. This patch includes:
1038
1039 - Fixed PH's badly added libz and libbz2 support.
1040 - Fixed a problem with static linking.
1041 - Added pcredemo. [But later removed - see 7 below.]
1042 - Fixed dftables problem and added an option.
1043 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1044 HAVE_LONG_LONG.
1045 - Added readline support for pcretest.
1046 - Added an listing of the option settings after cmake has run.
1047
1048 4. A user submitted a patch to Makefile that makes it easy to create
1049 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1050 Makefile.am that cause it to include this special target, without
1051 affecting anything else. Note that the same mingw target plus all
1052 the other distribution libraries and programs are now supported
1053 when configuring with CMake (see 6 below) instead of with
1054 Configure/Make.
1055
1056 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1057 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1058 exported in the Windows port". It has not yet been confirmed that the patch
1059 solves the problem, but it does no harm.
1060
1061 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1062 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1063 with CMake, and also correct the comment about stack recursion.
1064
1065 7. Remove the automatic building of pcredemo from the ./configure system and
1066 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1067 of a program that users should build themselves after PCRE is installed, so
1068 building it automatically is not really right. What is more, it gave
1069 trouble in some build environments.
1070
1071 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1072
1073
1074 Version 7.5 10-Jan-08
1075 ---------------------
1076
1077 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1078 values in parens when parsing an RE using the C++ wrapper."
1079
1080 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1081 Characters greater than 255 were excluded from the class instead of being
1082 included.
1083
1084 3. The same bug as (2) above applied to negated POSIX classes such as
1085 [:^space:].
1086
1087 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1088 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1089 I have changed it.
1090
1091 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1092 first named subpattern) and a construct such as (?&a) would reference the
1093 first named subpattern whose name started with "a" (in other words, the
1094 length check was missing). Both these problems are fixed. "Subpattern name
1095 expected" is now given for (?&) (a zero-length name), and this patch also
1096 makes it give the same error for \k'' (previously it complained that that
1097 was a reference to a non-existent subpattern).
1098
1099 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1100 this is right because (?- can be followed by option settings as well as by
1101 digits. I have, however, made the messages clearer.
1102
1103 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1104 than the number used in the conditional) now cause a compile-time error.
1105 This is actually not compatible with Perl, which accepts such patterns, but
1106 treats the conditional as always being FALSE (as PCRE used to), but it
1107 seems to me that giving a diagnostic is better.
1108
1109 8. Change "alphameric" to the more common word "alphanumeric" in comments
1110 and messages.
1111
1112 9. Fix two occurrences of "backslash" in comments that should have been
1113 "backspace".
1114
1115 10. Remove two redundant lines of code that can never be obeyed (their function
1116 was moved elsewhere).
1117
1118 11. The program that makes PCRE's Unicode character property table had a bug
1119 which caused it to generate incorrect table entries for sequences of
1120 characters that have the same character type, but are in different scripts.
1121 It amalgamated them into a single range, with the script of the first of
1122 them. In other words, some characters were in the wrong script. There were
1123 thirteen such cases, affecting characters in the following ranges:
1124
1125 U+002b0 - U+002c1
1126 U+0060c - U+0060d
1127 U+0061e - U+00612
1128 U+0064b - U+0065e
1129 U+0074d - U+0076d
1130 U+01800 - U+01805
1131 U+01d00 - U+01d77
1132 U+01d9b - U+01dbf
1133 U+0200b - U+0200f
1134 U+030fc - U+030fe
1135 U+03260 - U+0327f
1136 U+0fb46 - U+0fbb1
1137 U+10450 - U+1049d
1138
1139 12. The -o option (show only the matching part of a line) for pcregrep was not
1140 compatible with GNU grep in that, if there was more than one match in a
1141 line, it showed only the first of them. It now behaves in the same way as
1142 GNU grep.
1143
1144 13. If the -o and -v options were combined for pcregrep, it printed a blank
1145 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1146 does the same. The return code can be used to tell if there were any
1147 non-matching lines.
1148
1149 14. Added --file-offsets and --line-offsets to pcregrep.
1150
1151 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1152 infinitely looping recursion. The bug was that positive lookaheads were not
1153 being skipped when checking for a possible empty match (negative lookaheads
1154 and both kinds of lookbehind were skipped).
1155
1156 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1157 inclusion of <windows.h> to before rather than after the definition of
1158 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1159
1160 17. Specifying a possessive quantifier with a specific limit for a Unicode
1161 character property caused pcre_compile() to compile bad code, which led at
1162 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1163 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1164 caused the error; without that there was no problem.
1165
1166 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1167
1168 19. Added --enable-pcretest-libreadline.
1169
1170 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1171 RE::GlobalReplace(). As a result, the number of replacements returned was
1172 double what it should be. I removed one of the increments, but Craig sent a
1173 later patch that removed the other one (the right fix) and added unit tests
1174 that check the return values (which was not done before).
1175
1176 21. Several CMake things:
1177
1178 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1179 the names libpcre and libpcreposix, not just pcre and pcreposix.
1180
1181 (2) The above change means that pcretest and pcregrep are now correctly
1182 linked with the newly-built libraries, not previously installed ones.
1183
1184 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1185
1186 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1187 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1188 UTF-8 newline character). The key issue is that the pattern starts .*;
1189 this means that the match must be either at the beginning, or after a
1190 newline. The bug was in the code for advancing after a failed match and
1191 checking that the new position followed a newline. It was not taking
1192 account of UTF-8 characters correctly.
1193
1194 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1195 character classes. PCRE was not treating the sequence [:...:] as a
1196 character class unless the ... were all letters. Perl, however, seems to
1197 allow any characters between [: and :], though of course it rejects as
1198 unknown any "names" that contain non-letters, because all the known class
1199 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1200 for example, whereas PCRE did not - it did not recognize a POSIX character
1201 class. This seemed a bit dangerous, so the code has been changed to be
1202 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1203 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1204 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1205 Perl does, and where it didn't before.
1206
1207 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1208 Windows environments %n is disabled by default.
1209
1210
1211 Version 7.4 21-Sep-07
1212 ---------------------
1213
1214 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1215 means that a class such as [\s] counted as "explicit reference to CR or
1216 LF". That isn't really right - the whole point of the change was to try to
1217 help when there was an actual mention of one of the two characters. So now
1218 the change happens only if \r or \n (or a literal CR or LF) character is
1219 encountered.
1220
1221 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1222 of both had grown to the point where there were only 3 bits left.
1223 Fortunately, there was spare space in the data structure, and so I have
1224 moved the internal flags into a new 16-bit field to free up more option
1225 bits.
1226
1227 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1228 but did not set the internal JCHANGED flag - either of these is enough to
1229 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1230 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1231 start sets both bits.
1232
1233 4. Added options (at build time, compile time, exec time) to change \R from
1234 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1235
1236 5. doc/pcresyntax.html was missing from the distribution.
1237
1238 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1239 compatibility, even though it is no longer used.
1240
1241 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1242 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1243 windows.h file is present (where different names are used). [This was
1244 reversed later after testing - see 16 below.]
1245
1246 8. Changed all #include <config.h> to #include "config.h". There were also
1247 some further <pcre.h> cases that I changed to "pcre.h".
1248
1249 9. When pcregrep was used with the --colour option, it missed the line ending
1250 sequence off the lines that it output.
1251
1252 10. It was pointed out to me that arrays of string pointers cause lots of
1253 relocations when a shared library is dynamically loaded. A technique of
1254 using a single long string with a table of offsets can drastically reduce
1255 these. I have refactored PCRE in four places to do this. The result is
1256 dramatic:
1257
1258 Originally: 290
1259 After changing UCP table: 187
1260 After changing error message table: 43
1261 After changing table of "verbs" 36
1262 After changing table of Posix names 22
1263
1264 Thanks to the folks working on Gregex for glib for this insight.
1265
1266 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1267 unicode-properties was also set.
1268
1269 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1270
1271 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1272 checked only for CRLF.
1273
1274 14. Added casts to pcretest.c to avoid compiler warnings.
1275
1276 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1277
1278 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1279 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1280 entirely. This removes changes made in 7 above.
1281
1282 17. The CMake files have been updated, and there is now more information about
1283 building with CMake in the NON-UNIX-USE document.
1284
1285
1286 Version 7.3 28-Aug-07
1287 ---------------------
1288
1289 1. In the rejigging of the build system that eventually resulted in 7.1, the
1290 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1291 brackets there is not right, since it causes compilers to look for an
1292 installed pcre.h, not the version that is in the source that is being
1293 compiled (which of course may be different). I have changed it back to:
1294
1295 #include "pcre.h"
1296
1297 I have a vague recollection that the change was concerned with compiling in
1298 different directories, but in the new build system, that is taken care of
1299 by the VPATH setting the Makefile.
1300
1301 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1302 when the subject happened to end in the byte 0x85 (e.g. if the last
1303 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1304 characters but of course it shouldn't be taken as a newline when it is part
1305 of another character. The bug was that, for an unlimited repeat of . in
1306 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1307 characters when looking for a newline.
1308
1309 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1310
1311 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1312 in debug output.
1313
1314 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1315 long printing in the pcrecpp unittest when running under MinGW.
1316
1317 6. ESC_K was left out of the EBCDIC table.
1318
1319 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1320 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1321 limit also applies to "virtual nesting" when a pattern is recursive, and in
1322 this case 1000 isn't so big. I have been able to remove this limit at the
1323 expense of backing off one optimization in certain circumstances. Normally,
1324 when pcre_exec() would call its internal match() function recursively and
1325 immediately return the result unconditionally, it uses a "tail recursion"
1326 feature to save stack. However, when a subpattern that can match an empty
1327 string has an unlimited repetition quantifier, it no longer makes this
1328 optimization. That gives it a stack frame in which to save the data for
1329 checking that an empty string has been matched. Previously this was taken
1330 from the 1000-entry workspace that had been reserved. So now there is no
1331 explicit limit, but more stack is used.
1332
1333 8. Applied Daniel's patches to solve problems with the import/export magic
1334 syntax that is required for Windows, and which was going wrong for the
1335 pcreposix and pcrecpp parts of the library. These were overlooked when this
1336 problem was solved for the main library.
1337
1338 9. There were some crude static tests to avoid integer overflow when computing
1339 the size of patterns that contain repeated groups with explicit upper
1340 limits. As the maximum quantifier is 65535, the maximum group length was
1341 set at 30,000 so that the product of these two numbers did not overflow a
1342 32-bit integer. However, it turns out that people want to use groups that
1343 are longer than 30,000 bytes (though not repeat them that many times).
1344 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1345 made it possible to implement the integer overflow checks in a much more
1346 dynamic way, which I have now done. The artificial limitation on group
1347 length has been removed - we now have only the limit on the total length of
1348 the compiled pattern, which depends on the LINK_SIZE setting.
1349
1350 10. Fixed a bug in the documentation for get/copy named substring when
1351 duplicate names are permitted. If none of the named substrings are set, the
1352 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1353 empty string.
1354
1355 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1356 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1357 because the ] is interpreted as the first data character and the
1358 terminating ] is not found. PCRE has been made compatible with Perl in this
1359 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1360 cause memory overwriting.
1361
1362 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1363 string has been matched (to stop an infinite loop). It was not recognizing
1364 a conditional subpattern that could match an empty string if that
1365 subpattern was within another subpattern. For example, it looped when
1366 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1367 condition was not nested. This bug has been fixed.
1368
1369 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1370 past the start of the subject in the presence of bytes with the top bit
1371 set, for example "\x8aBCD".
1372
1373 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1374 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1375
1376 14. Optimized (?!) to (*FAIL).
1377
1378 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1379 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1380 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1381 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1382 does: it's just the validity check that is more restrictive.
1383
1384 16. Inserted checks for integer overflows during escape sequence (backslash)
1385 processing, and also fixed erroneous offset values for syntax errors during
1386 backslash processing.
1387
1388 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1389 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1390
1391 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1392 caused an overrun.
1393
1394 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1395 something other than just ASCII characters) inside a group that had an
1396 unlimited repeat caused a loop at compile time (while checking to see
1397 whether the group could match an empty string).
1398
1399 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1400 [\P{Any}] did so. (Error in the code for printing property names.)
1401
1402 21. An orphan \E inside a character class could cause a crash.
1403
1404 22. A repeated capturing bracket such as (A)? could cause a wild memory
1405 reference during compilation.
1406
1407 23. There are several functions in pcre_compile() that scan along a compiled
1408 expression for various reasons (e.g. to see if it's fixed length for look
1409 behind). There were bugs in these functions when a repeated \p or \P was
1410 present in the pattern. These operators have additional parameters compared
1411 with \d, etc, and these were not being taken into account when moving along
1412 the compiled data. Specifically:
1413
1414 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1415 length.
1416
1417 (b) An item such as \pL+ within a repeated group could cause crashes or
1418 loops.
1419
1420 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1421 "reference to non-existent subpattern" error.
1422
1423 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1424
1425 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1426 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1427
1428 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1429
1430 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1431 character were causing crashes (broken optimization).
1432
1433 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1434 \p or \P) caused a compile-time loop.
1435
1436 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1437 break. For example, the unstudied pattern [\r\n]A does not match the string
1438 "\r\nA" because change 7.0/46 below moves the current point on by two
1439 characters after failing to match at the start. However, the pattern \nA
1440 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1441 the same is true. There doesn't seem any very clean way out of this, but
1442 what I have chosen to do makes the common cases work: PCRE now takes note
1443 of whether there can be an explicit match for \r or \n anywhere in the
1444 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1445 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1446 pattern has explicit CR or LF references.
1447
1448 29. Added (*CR) etc for changing newline setting at start of pattern.
1449
1450
1451 Version 7.2 19-Jun-07
1452 ---------------------
1453
1454 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1455 which is apparently normally available under Windows.
1456
1457 2. Re-jig the pcregrep tests with different newline settings in an attempt
1458 to make them independent of the local environment's newline setting.
1459
1460 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1461
1462 4. Some of the "internals" tests were previously cut out when the link size
1463 was not 2, because the output contained actual offsets. The recent new
1464 "Z" feature of pcretest means that these can be cut out, making the tests
1465 usable with all link sizes.
1466
1467 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1468 stack recursion. This gives a massive performance boost under BSD, but just
1469 a small improvement under Linux. However, it saves one field in the frame
1470 in all cases.
1471
1472 6. Added more features from the forthcoming Perl 5.10:
1473
1474 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1475 recursion call. It refers to the nth most recently opened parentheses.
1476
1477 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1478 to be opened parentheses.
1479
1480 (c) Conditions that refer to capturing parentheses can be specified
1481 relatively, for example, (?(-2)... or (?(+3)...
1482
1483 (d) \K resets the start of the current match so that everything before
1484 is not part of it.
1485
1486 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1487
1488 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1489 reference syntax.
1490
1491 (g) (?| introduces a group in which the numbering of parentheses in each
1492 alternative starts with the same number.
1493
1494 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1495
1496 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1497 PCRE_INFO_JCHANGED.
1498
1499 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1500 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1501 for detecting groups that can match an empty string.
1502
1503 9. A pattern with a very large number of alternatives (more than several
1504 hundred) was running out of internal workspace during the pre-compile
1505 phase, where pcre_compile() figures out how much memory will be needed. A
1506 bit of new cunning has reduced the workspace needed for groups with
1507 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1508 workspace instead of running out of the 4096 that are available.
1509
1510 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1511
1512 11. Applied patch from Google to remove an optimization that didn't quite work.
1513 The report of the bug said:
1514
1515 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1516 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1517 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1518
1519 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1520 it matched the wrong number of bytes.
1521
1522
1523 Version 7.1 24-Apr-07
1524 ---------------------
1525
1526 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1527 that is more "standard", making use of automake and other Autotools. There
1528 is some re-arrangement of the files and adjustment of comments consequent
1529 on this.
1530
1531 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1532 for recursive directory scanning broke on some systems because the files
1533 are not scanned in any specific order and on different systems the order
1534 was different. A call to "sort" has been inserted into RunGrepTest for the
1535 approprate test as a short-term fix. In the longer term there may be an
1536 alternative.
1537
1538 3. I had an email from Eric Raymond about problems translating some of PCRE's
1539 man pages to HTML (despite the fact that I distribute HTML pages, some
1540 people do their own conversions for various reasons). The problems
1541 concerned the use of low-level troff macros .br and .in. I have therefore
1542 removed all such uses from the man pages (some were redundant, some could
1543 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1544 HTML has been updated to handle .nf/.fi and to complain if it encounters
1545 .br or .in.
1546
1547 4. Updated comments in configure.ac that get placed in config.h.in and also
1548 arranged for config.h to be included in the distribution, with the name
1549 config.h.generic, for the benefit of those who have to compile without
1550 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1551
1552 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1553 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1554 makevp.bat for the current PCRE, using the additional files
1555 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1556
1557 6. A Windows user reported a minor discrepancy with test 2, which turned out
1558 to be caused by a trailing space on an input line that had got lost in his
1559 copy. The trailing space was an accident, so I've just removed it.
1560
1561 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1562 that is needed.
1563
1564 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1565 as "const" (a) because they are and (b) because it helps the PHP
1566 maintainers who have recently made a script to detect big data structures
1567 in the php code that should be moved to the .rodata section. I remembered
1568 to update Builducptable as well, so it won't revert if ucptable.h is ever
1569 re-created.
1570
1571 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1572 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1573 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1574 support is not required. This saves 1.5-2K of code, which is important in
1575 some applications.
1576
1577 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1578 so as not to refer to the tables, even though these functions will never be
1579 called when UTF-8 support is disabled. Otherwise there are problems with a
1580 shared library.
1581
1582 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1583
1584 (a) It was defining its arguments as char * instead of void *.
1585
1586 (b) It was assuming that all moves were upwards in memory; this was true
1587 a long time ago when I wrote it, but is no longer the case.
1588
1589 The emulated memove() is provided for those environments that have neither
1590 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1591 is clearly not the case, as these two bugs were recently reported.
1592
1593 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1594 and Detrail to create the HTML documentation, the .txt form of the man
1595 pages, and it removes trailing spaces from listed files. It also creates
1596 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1597 case, it wraps all the #defines with #ifndefs. This script should be run
1598 before "make dist".
1599
1600 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1601 with Unicode property support.
1602
1603 (a) For a maximizing quantifier, if the two different cases of the
1604 character were of different lengths in their UTF-8 codings (there are
1605 some cases like this - I found 11), and the matching function had to
1606 back up over a mixture of the two cases, it incorrectly assumed they
1607 were both the same length.
1608
1609 (b) When PCRE was configured to use the heap rather than the stack for
1610 recursion during matching, it was not correctly preserving the data for
1611 the other case of a UTF-8 character when checking ahead for a match
1612 while processing a minimizing repeat. If the check also involved
1613 matching a wide character, but failed, corruption could cause an
1614 erroneous result when trying to check for a repeat of the original
1615 character.
1616
1617 13. Some tidying changes to the testing mechanism:
1618
1619 (a) The RunTest script now detects the internal link size and whether there
1620 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1621 values substituted by "configure". (The RunGrepTest script already did
1622 this for UTF-8.) The configure.ac script no longer substitutes the
1623 relevant variables.
1624
1625 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1626 with length and offset values. This means that the output is different
1627 for different internal link sizes. Test 2 is skipped for link sizes
1628 other than 2 because of this, bypassing the problem. Unfortunately,
1629 there was also a test in test 3 (the locale tests) that used /B and
1630 failed for link sizes other than 2. Rather than cut the whole test out,
1631 I have added a new /Z option to pcretest that replaces the length and
1632 offset values with spaces. This is now used to make test 3 independent
1633 of link size. (Test 2 will be tidied up later.)
1634
1635 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1636 segmentation fault instead of returning the appropriate error message.
1637
1638 15. In multiline mode when the newline sequence was set to "any", the pattern
1639 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1640 This doesn't seem right; it now treats the CRLF combination as the line
1641 ending, and so does not match in that case. It's only a pattern such as ^$
1642 that would hit this one: something like ^ABC$ would have failed after \r
1643 and then tried again after \r\n.
1644
1645 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1646 in an attempt to make files that differ only in their line terminators
1647 compare equal. This works on Linux.
1648
1649 17. Under certain error circumstances pcregrep might try to free random memory
1650 as it exited. This is now fixed, thanks to valgrind.
1651
1652 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1653 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1654 was because its rules for how to advance for /g after matching an empty
1655 string at the end of a line did not allow for this case. They now check for
1656 it specially.
1657
1658 20. pcretest is supposed to handle patterns and data of any length, by
1659 extending its buffers when necessary. It was getting this wrong when the
1660 buffer for a data line had to be extended.
1661
1662 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1663 CRLF as a newline sequence.
1664
1665 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1666 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1667 I have nevertheless tidied it up.
1668
1669 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1670
1671 24. Added a man page for pcre-config.
1672
1673
1674 Version 7.0 19-Dec-06
1675 ---------------------
1676
1677 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1678 moving to gcc 4.1.1.
1679
1680 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1681 sys/time.h, which is documented as needed for this function. It doesn't
1682 seem to matter on Linux, but it showed up on some releases of OS X.
1683
1684 3. It seems that there are systems where bytes whose values are greater than
1685 127 match isprint() in the "C" locale. The "C" locale should be the
1686 default when a C program starts up. In most systems, only ASCII printing
1687 characters match isprint(). This difference caused the output from pcretest
1688 to vary, making some of the tests fail. I have changed pcretest so that:
1689
1690 (a) When it is outputting text in the compiled version of a pattern, bytes
1691 other than 32-126 are always shown as hex escapes.
1692
1693 (b) When it is outputting text that is a matched part of a subject string,
1694 it does the same, unless a different locale has been set for the match
1695 (using the /L modifier). In this case, it uses isprint() to decide.
1696
1697 4. Fixed a major bug that caused incorrect computation of the amount of memory
1698 required for a compiled pattern when options that changed within the
1699 pattern affected the logic of the preliminary scan that determines the
1700 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1701 that the computed length was too small. The symptoms of this bug were
1702 either the PCRE error "internal error: code overflow" from pcre_compile(),
1703 or a glibc crash with a message such as "pcretest: free(): invalid next
1704 size (fast)". Examples of patterns that provoked this bug (shown in
1705 pcretest format) are:
1706
1707 /(?-x: )/x
1708 /(?x)(?-x: \s*#\s*)/
1709 /((?i)[\x{c0}])/8
1710 /(?i:[\x{c0}])/8
1711
1712 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1713 is now done differently.
1714
1715 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1716 wrapper classes; (b) implement a new function in the C++ scanner that is
1717 more efficient than the old way of doing things because it avoids levels of
1718 recursion in the regex matching; (c) add a paragraph to the documentation
1719 for the FullMatch() function.
1720
1721 6. The escape sequence \n was being treated as whatever was defined as
1722 "newline". Not only was this contrary to the documentation, which states
1723 that \n is character 10 (hex 0A), but it also went horribly wrong when
1724 "newline" was defined as CRLF. This has been fixed.
1725
1726 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1727 was being set to -1 for the "end of line" case (supposedly a value that no
1728 character can have). Though this value is never used (the check for end of
1729 line is "zero bytes in current character"), it caused compiler complaints.
1730 I've changed it to 0xffffffff.
1731
1732 8. In pcre_version.c, the version string was being built by a sequence of
1733 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1734 string (as it is for production releases) called a macro with an empty
1735 argument. The C standard says the result of this is undefined. The gcc
1736 compiler treats it as an empty string (which was what was wanted) but it is
1737 reported that Visual C gives an error. The source has been hacked around to
1738 avoid this problem.
1739
1740 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1741 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1742 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1743 of them did).
1744
1745 10. Originally, pcretest opened its input and output without "b"; then I was
1746 told that "b" was needed in some environments, so it was added for release
1747 5.0 to both the input and output. (It makes no difference on Unix-like
1748 systems.) Later I was told that it is wrong for the input on Windows. I've
1749 now abstracted the modes into two macros, to make it easier to fiddle with
1750 them, and removed "b" from the input mode under Windows.
1751
1752 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1753
1754 12. Added -help and --help to pcretest as an official way of being reminded
1755 of the options.
1756
1757 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1758 and pcrecpp.cc because they annoy compilers at high warning levels.
1759
1760 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1761
1762 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1763 scripts are not C programs :-) and which was not noticed because it works
1764 on Linux.
1765
1766 16. pcretest is supposed to handle any length of pattern and data line (as one
1767 line or as a continued sequence of lines) by extending its input buffer if
1768 necessary. This feature was broken for very long pattern lines, leading to
1769 a string of junk being passed to pcre_compile() if the pattern was longer
1770 than about 50K.
1771
1772 17. I have done a major re-factoring of the way pcre_compile() computes the
1773 amount of memory needed for a compiled pattern. Previously, there was code
1774 that made a preliminary scan of the pattern in order to do this. That was
1775 OK when PCRE was new, but as the facilities have expanded, it has become
1776 harder and harder to keep it in step with the real compile phase, and there
1777 have been a number of bugs (see for example, 4 above). I have now found a
1778 cunning way of running the real compile function in a "fake" mode that
1779 enables it to compute how much memory it would need, while actually only
1780 ever using a few hundred bytes of working memory and without too many
1781 tests of the mode. This should make future maintenance and development
1782 easier. A side effect of this work is that the limit of 200 on the nesting
1783 depth of parentheses has been removed (though this was never a serious
1784 limitation, I suspect). However, there is a downside: pcre_compile() now
1785 runs more slowly than before (30% or more, depending on the pattern). I
1786 hope this isn't a big issue. There is no effect on runtime performance.
1787
1788 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1789 newline (only possible for the last line of a file) and it was a
1790 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1791
1792 19. Added additional timing features to pcretest. (1) The -tm option now times
1793 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1794 separate command line item, by a number that specifies the number of
1795 repeats to use when timing. The default is 50000; this gives better
1796 precision, but takes uncomfortably long for very large patterns.
1797
1798 20. Extended pcre_study() to be more clever in cases where a branch of a
1799 subpattern has no definite first character. For example, (a*|b*)[cd] would
1800 previously give no result from pcre_study(). Now it recognizes that the
1801 first character must be a, b, c, or d.
1802
1803 21. There was an incorrect error "recursive call could loop indefinitely" if
1804 a subpattern (or the entire pattern) that was being tested for matching an
1805 empty string contained only one non-empty item after a nested subpattern.
1806 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1807 incorrectly, because the \d was being skipped in the check.
1808
1809 22. The pcretest program now has a new pattern option /B and a command line
1810 option -b, which is equivalent to adding /B to every pattern. This causes
1811 it to show the compiled bytecode, without the additional information that
1812 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1813 is the same as /B/I).
1814
1815 23. A new optimization is now able automatically to treat some sequences such
1816 as a*b as a*+b. More specifically, if something simple (such as a character
1817 or a simple class like \d) has an unlimited quantifier, and is followed by
1818 something that cannot possibly match the quantified thing, the quantifier
1819 is automatically "possessified".
1820
1821 24. A recursive reference to a subpattern whose number was greater than 39
1822 went wrong under certain circumstances in UTF-8 mode. This bug could also
1823 have affected the operation of pcre_study().
1824
1825 25. Realized that a little bit of performance could be had by replacing
1826 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1827
1828 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1829
1830 27. Possessive quantifiers such as a++ were previously implemented by turning
1831 them into atomic groups such as ($>a+). Now they have their own opcodes,
1832 which improves performance. This includes the automatically created ones
1833 from 23 above.
1834
1835 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1836 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1837 the first matched character to be a colon. This applied both to named and
1838 numbered groups.
1839
1840 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1841
1842 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1843 building PCRE on Windows easier, so I have included it in the distribution.
1844
1845 31. There is now a check in pcretest against a ridiculously large number being
1846 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1847 loop, the loop is abandoned.
1848
1849 32. Forward references to subpatterns in conditions such as (?(2)...) where
1850 subpattern 2 is defined later cause pcre_compile() to search forwards in
1851 the pattern for the relevant set of parentheses. This search went wrong
1852 when there were unescaped parentheses in a character class, parentheses
1853 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1854
1855 33. "Subroutine" calls and backreferences were previously restricted to
1856 referencing subpatterns earlier in the regex. This restriction has now
1857 been removed.
1858
1859 34. Added a number of extra features that are going to be in Perl 5.10. On the
1860 whole, these are just syntactic alternatives for features that PCRE had
1861 previously implemented using the Python syntax or my own invention. The
1862 other formats are all retained for compatibility.
1863
1864 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1865 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1866 also .NET compatible.
1867
1868 (b) A recursion or subroutine call to a named group can now be defined as
1869 (?&name) as well as (?P>name).
1870
1871 (c) A backreference to a named group can now be defined as \k<name> or
1872 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1873 5.10, are also .NET compatible.
1874
1875 (d) A conditional reference to a named group can now use the syntax
1876 (?(<name>) or (?('name') as well as (?(name).
1877
1878 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1879 groups (named and numbered) that are never evaluated inline, but can be
1880 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1881 is always false. There may be only one alternative in such a group.
1882
1883 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1884 as the simple (?(R). The condition is true only if the most recent
1885 recursion is that of the given number or name. It does not search out
1886 through the entire recursion stack.
1887
1888 (g) The escape \gN or \g{N} has been added, where N is a positive or
1889 negative number, specifying an absolute or relative reference.
1890
1891 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1892 some "unreachable code" warnings.
1893
1894 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1895 things, this adds five new scripts.
1896
1897 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1898 There were also incompatibilities regarding the handling of \Q..\E inside
1899 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1900 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1901
1902 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1903 matches an empty string, and forcibly breaks the loop. There were bugs in
1904 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1905 against aaaa the result was just "a" rather than "aaaa", for example. Two
1906 separate and independent bugs (that affected different cases) have been
1907 fixed.
1908
1909 39. Refactored the code to abolish the use of different opcodes for small
1910 capturing bracket numbers. This is a tidy that I avoided doing when I
1911 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1912 The new approach is not only tidier, it makes it possible to reduce the
1913 memory needed to fix the previous bug (38).
1914
1915 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1916 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1917 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1918 mode.
1919
1920 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1921 report.
1922
1923 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1924 copy construction and assignment in the C++ wrapper.
1925
1926 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1927 couple of bugs that could have given wrong results in the "--newline=crlf"
1928 case.
1929
1930 44. Added a number of casts and did some reorganization of signed/unsigned int
1931 variables following suggestions from Dair Grant. Also renamed the variable
1932 "this" as "item" because it is a C++ keyword.
1933
1934 45. Arranged for dftables to add
1935
1936 #include "pcre_internal.h"
1937
1938 to pcre_chartables.c because without it, gcc 4.x may remove the array
1939 definition from the final binary if PCRE is built into a static library and
1940 dead code stripping is activated.
1941
1942 46. For an unanchored pattern, if a match attempt fails at the start of a
1943 newline sequence, and the newline setting is CRLF or ANY, and the next two
1944 characters are CRLF, advance by two characters instead of one.
1945
1946
1947 Version 6.7 04-Jul-06
1948 ---------------------
1949
1950 1. In order to handle tests when input lines are enormously long, pcretest has
1951 been re-factored so that it automatically extends its buffers when
1952 necessary. The code is crude, but this _is_ just a test program. The
1953 default size has been increased from 32K to 50K.
1954
1955 2. The code in pcre_study() was using the value of the re argument before
1956 testing it for NULL. (Of course, in any sensible call of the function, it
1957 won't be NULL.)
1958
1959 3. The memmove() emulation function in pcre_internal.h, which is used on
1960 systems that lack both memmove() and bcopy() - that is, hardly ever -
1961 was missing a "static" storage class specifier.
1962
1963 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1964 containing an extended class (one that cannot be represented by a bitmap
1965 because it contains high-valued characters or Unicode property items, e.g.
1966 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1967 pattern, but PCRE should not loop if you do not (it no longer does).
1968 [Detail: two cases were found: (a) a repeated subpattern containing an
1969 extended class; (b) a recursive reference to a subpattern that followed a
1970 previous extended class. It wasn't skipping over the extended class
1971 correctly when UTF-8 mode was not set.]
1972
1973 5. A negated single-character class was not being recognized as fixed-length
1974 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1975 compile error "lookbehind assertion is not fixed length".
1976
1977 6. The RunPerlTest auxiliary script was showing an unexpected difference
1978 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1979 write a Perl script that can interpret lines of an input file either as
1980 byte characters or as UTF-8, which is what "perltest" was being required to
1981 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1982 can't do is switch easily at run time between having the "use utf8;" pragma
1983 or not. In the end, I fudged it by using the RunPerlTest script to insert
1984 "use utf8;" explicitly for the UTF-8 tests.
1985
1986 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1987 the end of the subject string, contrary to the documentation and to what
1988 Perl does. This was true of both matching functions. Now it matches only at
1989 the start of the subject and immediately after *internal* newlines.
1990
1991 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1992 a pointer to an int instead of a pointer to an unsigned long int. This
1993 caused problems on 64-bit systems.
1994
1995 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1996 instance of the 'standard' template library not being so standard".
1997
1998 10. There was no check on the number of named subpatterns nor the maximum
1999 length of a subpattern name. The product of these values is used to compute
2000 the size of the memory block for a compiled pattern. By supplying a very
2001 long subpattern name and a large number of named subpatterns, the size
2002 computation could be caused to overflow. This is now prevented by limiting
2003 the length of names to 32 characters, and the number of named subpatterns
2004 to 10,000.
2005
2006 11. Subpatterns that are repeated with specific counts have to be replicated in
2007 the compiled pattern. The size of memory for this was computed from the
2008 length of the subpattern and the repeat count. The latter is limited to
2009 65535, but there was no limit on the former, meaning that integer overflow
2010 could in principle occur. The compiled length of a repeated subpattern is
2011 now limited to 30,000 bytes in order to prevent this.
2012
2013 12. Added the optional facility to have named substrings with the same name.
2014
2015 13. Added the ability to use a named substring as a condition, using the
2016 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2017 are numbers (not recommended). Forward references are permitted.
2018
2019 14. Added forward references in named backreferences (if you see what I mean).
2020
2021 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2022 pattern could run off the end of the subject. For example, the pattern
2023 "(?s)(.{1,5})"8 did this with the subject "ab".
2024
2025 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2026 PCRE_CASELESS was set when matching characters that were quantified with ?
2027 or *.
2028
2029 17. A character class other than a single negated character that had a minimum
2030 but no maximum quantifier - for example [ab]{6,} - was not handled
2031 correctly by pce_dfa_exec(). It would match only one character.
2032
2033 18. A valid (though odd) pattern that looked like a POSIX character
2034 class but used an invalid character after [ (for example [[,abc,]]) caused
2035 pcre_compile() to give the error "Failed: internal error: code overflow" or
2036 in some cases to crash with a glibc free() error. This could even happen if
2037 the pattern terminated after [[ but there just happened to be a sequence of
2038 letters, a binary zero, and a closing ] in the memory that followed.
2039
2040 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2041 over the years. Originally (before any Unicode support), just the bottom 8
2042 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2043 output from "man perlunicode" includes this:
2044
2045 The regular expression compiler produces polymorphic opcodes. That
2046 is, the pattern adapts to the data and automatically switches to
2047 the Unicode character scheme when presented with Unicode data--or
2048 instead uses a traditional byte scheme when presented with byte
2049 data.
2050
2051 Sadly, a wide octal escape does not cause a switch, and in a string with
2052 no other multibyte characters, these octal escapes are treated as before.
2053 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2054 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2055 Unicode string.
2056
2057 I have not perpetrated such confusion in PCRE. Up till now, it took just
2058 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2059 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2060 translate to the appropriate multibyte character.
2061
2062 29. Applied some refactoring to reduce the number of warnings from Microsoft
2063 and Borland compilers. This has included removing the fudge introduced
2064 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2065 a warning about an unused variable.
2066
2067 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2068 characters since release 4.0, because Perl (from release 5.004) does not.
2069 [Or at least, is documented not to: some releases seem to be in conflict
2070 with the documentation.] However, when a pattern was studied with
2071 pcre_study() and all its branches started with \s, PCRE still included VT
2072 as a possible starting character. Of course, this did no harm; it just
2073 caused an unnecessary match attempt.
2074
2075 22. Removed a now-redundant internal flag bit that recorded the fact that case
2076 dependency changed within the pattern. This was once needed for "required
2077 byte" processing, but is no longer used. This recovers a now-scarce options
2078 bit. Also moved the least significant internal flag bit to the most-
2079 significant bit of the word, which was not previously used (hangover from
2080 the days when it was an int rather than a uint) to free up another bit for
2081 the future.
2082
2083 23. Added support for CRLF line endings as well as CR and LF. As well as the
2084 default being selectable at build time, it can now be changed at runtime
2085 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2086 specify that it is scanning data with non-default line endings.
2087
2088 24. Changed the definition of CXXLINK to make it agree with the definition of
2089 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2090
2091 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2092 recursions. This makes a big different to stack usage for some patterns.
2093
2094 26. If a subpattern containing a named recursion or subroutine reference such
2095 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2096 the space required for the compiled pattern went wrong and gave too small a
2097 value. Depending on the environment, this could lead to "Failed: internal
2098 error: code overflow at offset 49" or "glibc detected double free or
2099 corruption" errors.
2100
2101 27. Applied patches from Google (a) to support the new newline modes and (b) to
2102 advance over multibyte UTF-8 characters in GlobalReplace.
2103
2104 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2105 difference for some implementation of PCRE in some Windows version.
2106
2107 29. Added some extra testing facilities to pcretest:
2108
2109 \q<number> in a data line sets the "match limit" value
2110 \Q<number> in a data line sets the "match recursion limt" value
2111 -S <number> sets the stack size, where <number> is in megabytes
2112
2113 The -S option isn't available for Windows.
2114
2115
2116 Version 6.6 06-Feb-06
2117 ---------------------
2118
2119 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2120 in pcreposix.h. I have copied the definition from pcre.h.
2121
2122 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2123 because pcre.h is no longer a built file.
2124
2125 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2126 not normally included in the compiled code.
2127
2128
2129 Version 6.5 01-Feb-06
2130 ---------------------
2131
2132 1. When using the partial match feature with pcre_dfa_exec(), it was not
2133 anchoring the second and subsequent partial matches at the new starting
2134 point. This could lead to incorrect results. For example, with the pattern
2135 /1234/, partially matching against "123" and then "a4" gave a match.
2136
2137 2. Changes to pcregrep:
2138
2139 (a) All non-match returns from pcre_exec() were being treated as failures
2140 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2141 error message is output. Some extra information is given for the
2142 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2143 probably the only errors that are likely to be caused by users (by
2144 specifying a regex that has nested indefinite repeats, for instance).
2145 If there are more than 20 of these errors, pcregrep is abandoned.
2146
2147 (b) A binary zero was treated as data while matching, but terminated the
2148 output line if it was written out. This has been fixed: binary zeroes
2149 are now no different to any other data bytes.
2150
2151 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2152 used to set a locale for matching. The --locale=xxxx long option has
2153 been added (no short equivalent) to specify a locale explicitly on the
2154 pcregrep command, overriding the environment variables.
2155
2156 (d) When -B was used with -n, some line numbers in the output were one less
2157 than they should have been.
2158
2159 (e) Added the -o (--only-matching) option.
2160
2161 (f) If -A or -C was used with -c (count only), some lines of context were
2162 accidentally printed for the final match.
2163
2164 (g) Added the -H (--with-filename) option.
2165
2166 (h) The combination of options -rh failed to suppress file names for files
2167 that were found from directory arguments.
2168
2169 (i) Added the -D (--devices) and -d (--directories) options.
2170
2171 (j) Added the -F (--fixed-strings) option.
2172
2173 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2174
2175 (l) Added the --colo(u)r option.
2176
2177 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2178 is not present by default.
2179
2180 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2181 items such as (?R) or (?1), when the recursion could match a number of
2182 alternatives. If it matched one of the alternatives, but subsequently,
2183 outside the recursion, there was a failure, the code tried to back up into
2184 the recursion. However, because of the way PCRE is implemented, this is not
2185 possible, and the result was an incorrect result from the match.
2186
2187 In order to prevent this happening, the specification of recursion has
2188 been changed so that all such subpatterns are automatically treated as
2189 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2190
2191 4. I had overlooked the fact that, in some locales, there are characters for
2192 which isalpha() is true but neither isupper() nor islower() are true. In
2193 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2194 and ordfeminine) are like this. This affected the treatment of \w and \W
2195 when they appeared in character classes, but not when they appeared outside
2196 a character class. The bit map for "word" characters is now created
2197 separately from the results of isalnum() instead of just taking it from the
2198 upper, lower, and digit maps. (Plus the underscore character, of course.)
2199
2200 5. The above bug also affected the handling of POSIX character classes such as
2201 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2202 permanent tables. Instead, the bit maps for such a class were previously
2203 created as the appropriate unions of the upper, lower, and digit bitmaps.
2204 Now they are created by subtraction from the [[:word:]] class, which has
2205 its own bitmap.
2206
2207 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2208 It is created by subtracting the vertical space characters (\x09, \x0a,
2209 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2210 subtraction was done in the overall bitmap for a character class, meaning
2211 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2212 be recognized. This bug has been fixed.
2213
2214 7. Patches from the folks at Google:
2215
2216 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2217 real life, but is still worth protecting against".
2218
2219 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2220 regular expressions".
2221
2222 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2223 have it.
2224
2225 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2226 "configure" and the latter not, in order to fix a problem somebody had
2227 with compiling the Arg class on HP-UX.
2228
2229 (e) Improve the error-handling of the C++ wrapper a little bit.
2230
2231 (f) New tests for checking recursion limiting.
2232
2233 8. The pcre_memmove() function, which is used only if the environment does not
2234 have a standard memmove() function (and is therefore rarely compiled),
2235 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2236 returning a result (though PCRE never actually uses the result).
2237
2238 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2239 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2240 returned instead of calling malloc() with an overflowing number that would
2241 most likely cause subsequent chaos.
2242
2243 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2244
2245 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2246 with this option is matched, the nmatch and pmatch options of regexec() are
2247 ignored.
2248
2249 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2250 provided in case anyone wants to the the POSIX interface with UTF-8
2251 strings.
2252
2253 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2254 C++ linking (needed for some HP-UX environments).
2255
2256 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2257 (unused parameter) and in the pcre_printint() function (omitted "default"
2258 switch label when the default is to do nothing).
2259
2260 15. Added some code to make it possible, when PCRE is compiled as a C++
2261 library, to replace subject pointers for pcre_exec() with a smart pointer
2262 class, thus making it possible to process discontinuous strings.
2263
2264 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2265 much the same function. They were added by different people who were trying
2266 to make PCRE easy to compile on non-Unix systems. It has been suggested
2267 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2268 for compiling on Windows systems. I have therefore replaced it with
2269 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2270 defaults to "extern" for C or "extern C" for C++, which works fine on
2271 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2272 SCOPE with something explicit in config.h. In addition:
2273
2274 (a) pcreposix.h still had just "extern" instead of either of these macros;
2275 I have replaced it with PCRE_DATA_SCOPE.
2276
2277 (b) Functions such as _pcre_xclass(), which are internal to the library,
2278 but external in the C sense, all had PCRE_EXPORT in their definitions.
2279 This is apparently wrong for the Windows case, so I have removed it.
2280 (It makes no difference on Unix-like systems.)
2281
2282 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2283 of recursive calls to match(). This is different to MATCH_LIMIT because
2284 that limits the total number of calls to match(), not all of which increase
2285 the depth of recursion. Limiting the recursion depth limits the amount of
2286 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2287 when PCRE is compiled, and changed at run time. A patch from Google adds
2288 this functionality to the C++ interface.
2289
2290 18. Changes to the handling of Unicode character properties:
2291
2292 (a) Updated the table to Unicode 4.1.0.
2293
2294 (b) Recognize characters that are not in the table as "Cn" (undefined).
2295
2296 (c) I revised the way the table is implemented to a much improved format
2297 which includes recognition of ranges. It now supports the ranges that
2298 are defined in UnicodeData.txt, and it also amalgamates other
2299 characters into ranges. This has reduced the number of entries in the
2300 table from around 16,000 to around 3,000, thus reducing its size
2301 considerably. I realized I did not need to use a tree structure after
2302 all - a binary chop search is just as efficient. Having reduced the
2303 number of entries, I extended their size from 6 bytes to 8 bytes to
2304 allow for more data.
2305
2306 (d) Added support for Unicode script names via properties such as \p{Han}.
2307
2308 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2309 matching that character.
2310
2311 20. When matching a repeated Unicode property with a minimum greater than zero,
2312 (for example \pL{2,}), PCRE could look past the end of the subject if it
2313 reached it while seeking the minimum number of characters. This could
2314 happen only if some of the characters were more than one byte long, because
2315 there is a check for at least the minimum number of bytes.
2316
2317 21. Refactored the implementation of \p and \P so as to be more general, to
2318 allow for more different types of property in future. This has changed the
2319 compiled form incompatibly. Anybody with saved compiled patterns that use
2320 \p or \P will have to recompile them.
2321
2322 22. Added "Any" and "L&" to the supported property types.
2323
2324 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2325 but give a compile time error if the value is greater than 0xff.
2326
2327 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2328 accidentally not being installed or uninstalled.
2329
2330 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2331 made were to insert the current release number. This seemed silly, because
2332 it made things harder for people building PCRE on systems that don't run
2333 "configure". I have turned pcre.h into a distributed file, no longer built
2334 by "configure", with the version identification directly included. There is
2335 no longer a pcre.h.in file.
2336
2337 However, this change necessitated a change to the pcre-config script as
2338 well. It is built from pcre-config.in, and one of the substitutions was the
2339 release number. I have updated configure.ac so that ./configure now finds
2340 the release number by grepping pcre.h.
2341
2342 26. Added the ability to run the tests under valgrind.
2343
2344
2345 Version 6.4 05-Sep-05
2346 ---------------------
2347
2348 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2349 "--" to be printed when multiple files were scanned, even when none of the
2350 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2351 consider it to be a bug, and have restored the previous behaviour.
2352
2353 2. A couple of code tidies to get rid of compiler warnings.
2354
2355 3. The pcretest program used to cheat by referring to symbols in the library
2356 whose names begin with _pcre_. These are internal symbols that are not
2357 really supposed to be visible externally, and in some environments it is
2358 possible to suppress them. The cheating is now confined to including
2359 certain files from the library's source, which is a bit cleaner.
2360
2361 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2362 file's purpose clearer.
2363
2364 5. Reorganized pcre_ucp_findchar().
2365
2366
2367 Version 6.3 15-Aug-05
2368 ---------------------
2369
2370 1. The file libpcre.pc.in did not have general read permission in the tarball.
2371
2372 2. There were some problems when building without C++ support:
2373
2374 (a) If C++ support was not built, "make install" and "make test" still
2375 tried to test it.
2376
2377 (b) There were problems when the value of CXX was explicitly set. Some
2378 changes have been made to try to fix these, and ...
2379
2380 (c) --disable-cpp can now be used to explicitly disable C++ support.
2381
2382 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2383 backslash in a target when C++ was disabled. This confuses some
2384 versions of "make", apparently. Using an intermediate variable solves
2385 this. (Same for CPP_LOBJ.)
2386
2387 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2388 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2389 necessary on certain architectures.
2390
2391 4. Added a setting of -export-symbols-regex to the link command to remove
2392 those symbols that are exported in the C sense, but actually are local
2393 within the library, and not documented. Their names all begin with
2394 "_pcre_". This is not a perfect job, because (a) we have to except some
2395 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2396 available (and never for static libraries). I have made a note to try to
2397 find a way round (a) in the future.
2398
2399
2400 Version 6.2 01-Aug-05
2401 ---------------------
2402
2403 1. There was no test for integer overflow of quantifier values. A construction
2404 such as {1111111111111111} would give undefined results. What is worse, if
2405 a minimum quantifier for a parenthesized subpattern overflowed and became
2406 negative, the calculation of the memory size went wrong. This could have
2407 led to memory overwriting.
2408
2409 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2410
2411 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2412 operating environments where this matters.
2413
2414 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2415 PCRE options from within the C++ wrapper.
2416
2417 5. Named capturing subpatterns were not being correctly counted when a pattern
2418 was compiled. This caused two problems: (a) If there were more than 100
2419 such subpatterns, the calculation of the memory needed for the whole
2420 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2421 back references of the form \12, where the number was greater than 9, were
2422 not recognized as back references, even though there were sufficient
2423 previous subpatterns.
2424
2425 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2426 versions of gcc, e.g. 2.95.4.
2427
2428
2429 Version 6.1 21-Jun-05
2430 ---------------------
2431
2432 1. There was one reference to the variable "posix" in pcretest.c that was not
2433 surrounded by "#if !defined NOPOSIX".
2434
2435 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2436 the cross-check on the old pcre_info() function, for the benefit of the
2437 cut-down version of PCRE that is currently imported into Exim.
2438
2439 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2440 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2441 patterns that start (?i) but I don't think that matters. The use of (?i) is
2442 just an example; this all applies to the other options as well.
2443
2444 4. Since libtool seems to echo the compile commands it is issuing, the output
2445 from "make" can be reduced a bit by putting "@" in front of each libtool
2446 compile command.
2447
2448 5. Patch from the folks at Google for configure.in to be a bit more thorough
2449 in checking for a suitable C++ installation before trying to compile the
2450 C++ stuff. This should fix a reported problem when a compiler was present,
2451 but no suitable headers.
2452
2453 6. The man pages all had just "PCRE" as their title. I have changed them to
2454 be the relevant file name. I have also arranged that these names are
2455 retained in the file doc/pcre.txt, which is a concatenation in text format
2456 of all the man pages except the little individual ones for each function.
2457
2458 7. The NON-UNIX-USE file had not been updated for the different set of source
2459 files that come with release 6. I also added a few comments about the C++
2460 wrapper.
2461
2462
2463 Version 6.0 07-Jun-05
2464 ---------------------
2465
2466 1. Some minor internal re-organization to help with my DFA experiments.
2467
2468 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2469 didn't matter for the library itself when fully configured, but did matter
2470 when compiling without UCP support, or within Exim, where the ucp files are
2471 not imported.
2472
2473 3. Refactoring of the library code to split up the various functions into
2474 different source modules. The addition of the new DFA matching code (see
2475 below) to a single monolithic source would have made it really too
2476 unwieldy, quite apart from causing all the code to be include in a
2477 statically linked application, when only some functions are used. This is
2478 relevant even without the DFA addition now that patterns can be compiled in
2479 one application and matched in another.
2480
2481 The downside of splitting up is that there have to be some external
2482 functions and data tables that are used internally in different modules of
2483 the library but which are not part of the API. These have all had their
2484 names changed to start with "_pcre_" so that they are unlikely to clash
2485 with other external names.
2486
2487 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2488 a different (DFA) algorithm. Although it is slower than the original
2489 function, it does have some advantages for certain types of matching
2490 problem.
2491
2492 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2493 including restarting after a partial match.
2494
2495 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2496 defined when compiling for Windows was sent to me. I have put it into the
2497 code, though I have no means of testing or verifying it.
2498
2499 7. Added the pcre_refcount() auxiliary function.
2500
2501 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2502 match before or at the first newline in the subject string. In pcretest,
2503 the /f option on a pattern can be used to set this.
2504
2505 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2506 would behave wrongly. This has been present in PCRE since release 4.0.
2507
2508 10. A number of changes to the pcregrep command:
2509
2510 (a) Refactored how -x works; insert ^(...)$ instead of setting
2511 PCRE_ANCHORED and checking the length, in preparation for adding
2512 something similar for -w.
2513
2514 (b) Added the -w (match as a word) option.
2515
2516 (c) Refactored the way lines are read and buffered so as to have more
2517 than one at a time available.
2518
2519 (d) Implemented a pcregrep test script.
2520
2521 (e) Added the -M (multiline match) option. This allows patterns to match
2522 over several lines of the subject. The buffering ensures that at least
2523 8K, or the rest of the document (whichever is the shorter) is available
2524 for matching (and similarly the previous 8K for lookbehind assertions).
2525
2526 (f) Changed the --help output so that it now says
2527
2528 -w, --word-regex(p)
2529
2530 instead of two lines, one with "regex" and the other with "regexp"
2531 because that confused at least one person since the short forms are the
2532 same. (This required a bit of code, as the output is generated
2533 automatically from a table. It wasn't just a text change.)
2534
2535 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2536 option but starts with a hyphen. Could be a pattern or a path name
2537 starting with a hyphen, for instance.
2538
2539 (h) "-" can be given as a file name to represent stdin.
2540
2541 (i) When file names are being printed, "(standard input)" is used for
2542 the standard input, for compatibility with GNU grep. Previously
2543 "<stdin>" was used.
2544
2545 (j) The option --label=xxx can be used to supply a name to be used for
2546 stdin when file names are being printed. There is no short form.
2547
2548 (k) Re-factored the options decoding logic because we are going to add
2549 two more options that take data. Such options can now be given in four
2550 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2551
2552 (l) Added the -A, -B, and -C options for requesting that lines of context
2553 around matches be printed.
2554
2555 (m) Added the -L option to print the names of files that do not contain
2556 any matching lines, that is, the complement of -l.
2557
2558 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2559 continue to scan other files.
2560
2561 (o) The -s option was incorrectly implemented. For compatibility with other
2562 greps, it now suppresses the error message for a non-existent or non-
2563 accessible file (but not the return code). There is a new option called
2564 -q that suppresses the output of matching lines, which was what -s was
2565 previously doing.
2566
2567 (p) Added --include and --exclude options to specify files for inclusion
2568 and exclusion when recursing.
2569
2570 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2571 Hopefully, it now does.
2572
2573 12. Missing cast in pcre_study().
2574
2575 13. Added an "uninstall" target to the makefile.
2576
2577 14. Replaced "extern" in the function prototypes in Makefile.in with
2578 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2579 world, but is set differently for Windows.
2580
2581 15. Added a second compiling function called pcre_compile2(). The only
2582 difference is that it has an extra argument, which is a pointer to an
2583 integer error code. When there is a compile-time failure, this is set
2584 non-zero, in addition to the error test pointer being set to point to an
2585 error message. The new argument may be NULL if no error number is required
2586 (but then you may as well call pcre_compile(), which is now just a
2587 wrapper). This facility is provided because some applications need a
2588 numeric error indication, but it has also enabled me to tidy up the way
2589 compile-time errors are handled in the POSIX wrapper.
2590
2591 16. Added VPATH=.libs to the makefile; this should help when building with one
2592 prefix path and installing with another. (Or so I'm told by someone who
2593 knows more about this stuff than I do.)
2594
2595 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2596 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2597 match everything, including newlines. This is not POSIX-compatible, but
2598 somebody wanted the feature. From pcretest it can be activated by using
2599 both the P and the s flags.
2600
2601 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2602
2603 19. libpcre.pc was being incorrectly installed as executable.
2604
2605 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2606 it now also looks for '\r' so that it will work unmodified on Windows.
2607
2608 21. Added Google's contributed C++ wrapper to the distribution.
2609
2610 22. Added some untidy missing memory free() calls in pcretest, to keep
2611 Electric Fence happy when testing.
2612
2613
2614
2615 Version 5.0 13-Sep-04
2616 ---------------------
2617
2618 1. Internal change: literal characters are no longer packed up into items
2619 containing multiple characters in a single byte-string. Each character
2620 is now matched using a separate opcode. However, there may be more than one
2621 byte in the character in UTF-8 mode.
2622
2623 2. The pcre_callout_block structure has two new fields: pattern_position and
2624 next_item_length. These contain the offset in the pattern to the next match
2625 item, and its length, respectively.
2626
2627 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2628 insertion of callouts before each pattern item. Added the /C option to
2629 pcretest to make use of this.
2630
2631 4. On the advice of a Windows user, the lines
2632
2633 #if defined(_WIN32) || defined(WIN32)
2634 _setmode( _fileno( stdout ), 0x8000 );
2635 #endif /* defined(_WIN32) || defined(WIN32) */
2636
2637 have been added to the source of pcretest. This apparently does useful
2638 magic in relation to line terminators.
2639
2640 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2641 for the benefit of those environments where the "b" makes a difference.
2642
2643 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2644 to know about it. I have put a hack into configure.in that adds in code
2645 to set GCC=yes if CC=icc. This seems to end up at a point in the
2646 generated configure script that is early enough to affect the setting of
2647 compiler options, which is what is needed, but I have no means of testing
2648 whether it really works. (The user who reported this had patched the
2649 generated configure script, which of course I cannot do.)
2650
2651 LATER: After change 22 below (new libtool files), the configure script
2652 seems to know about icc (and also ecc). Therefore, I have commented out
2653 this hack in configure.in.
2654
2655 7. Added support for pkg-config (2 patches were sent in).
2656
2657 8. Negated POSIX character classes that used a combination of internal tables
2658 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2659 [[:^ascii]]. Typically, they would match almost any characters. The other
2660 POSIX classes were not broken in this way.
2661
2662 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2663 to find the match, as PCRE was deluded into thinking that the match had to
2664 start at the start point or following a newline. The same bug applied to
2665 patterns with negative forward assertions or any backward assertions
2666 preceding ".*" at the start, unless the pattern required a fixed first
2667 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2668
2669 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2670 starting at the last subject character, bytes beyond the end of the subject
2671 string were read.
2672
2673 11. Renamed the variable "class" as "classbits" to make life easier for C++
2674 users. (Previously there was a macro definition, but it apparently wasn't
2675 enough.)
2676
2677 12. Added the new field "tables" to the extra data so that tables can be passed
2678 in at exec time, or the internal tables can be re-selected. This allows
2679 a compiled regex to be saved and re-used at a later time by a different
2680 program that might have everything at different addresses.
2681
2682 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2683 -R library as well as a -L library.
2684
2685 14. The debugging options of pcretest (-d on the command line or D on a
2686 pattern) showed incorrect output for anything following an extended class
2687 that contained multibyte characters and which was followed by a quantifier.
2688
2689 15. Added optional support for general category Unicode character properties
2690 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2691 support. It adds about 90K to the size of the library. The meanings of the
2692 inbuilt class escapes such as \d and \s have NOT been changed.
2693
2694 16. Updated pcredemo.c to include calls to free() to release the memory for the
2695 compiled pattern.
2696
2697 17. The generated file chartables.c was being created in the source directory
2698 instead of in the building directory. This caused the build to fail if the
2699 source directory was different from the building directory, and was
2700 read-only.
2701
2702 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2703 file. No doubt somebody will tell me if they don't make sense... Also added
2704 Dan Mooney's comments about building on OpenVMS.
2705
2706 19. Added support for partial matching via the PCRE_PARTIAL option for
2707 pcre_exec() and the \P data escape in pcretest.
2708
2709 20. Extended pcretest with 3 new pattern features:
2710
2711 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2712 write the compiled pattern to the file whose name is "rest-of-line".
2713 This is a straight binary dump of the data, with the saved pointer to
2714 the character tables forced to be NULL. The study data, if any, is
2715 written too. After writing, pcretest reads a new pattern.
2716
2717 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2718 compiled pattern from the given file. There must not be any
2719 occurrences of "<" in the file name (pretty unlikely); if there are,
2720 pcretest will instead treat the initial "<" as a pattern delimiter.
2721 After reading in the pattern, pcretest goes on to read data lines as
2722 usual.
2723
2724 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2725 and 16-bit fields in a compiled pattern, to simulate a pattern that
2726 was compiled on a host of opposite endianness.
2727
2728 21. The pcre-exec() function can now cope with patterns that were compiled on
2729 hosts of opposite endianness, with this restriction:
2730
2731 As for any compiled expression that is saved and used later, the tables
2732 pointer field cannot be preserved; the extra_data field in the arguments
2733 to pcre_exec() should be used to pass in a tables address if a value
2734 other than the default internal tables were used at compile time.
2735
2736 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2737 now diagnosed as an error. Previously, most of the time, a negative number
2738 would have been treated as zero, but if in addition "ovector" was passed as
2739 NULL, a crash could occur.
2740
2741 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2742 new versions from the libtool 1.5 distribution (the last one is a copy of
2743 a file called libtool.m4). This seems to have fixed the need to patch
2744 "configure" to support Darwin 1.3 (which I used to do). However, I still
2745 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2746 workstation).
2747
2748 24. Changed the PCRE licence to be the more standard "BSD" licence.
2749
2750
2751 Version 4.5 01-Dec-03
2752 ---------------------
2753
2754 1. There has been some re-arrangement of the code for the match() function so
2755 that it can be compiled in a version that does not call itself recursively.
2756 Instead, it keeps those local variables that need separate instances for
2757 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2758 needs to "recurse". Keeping track of where control must go is done by means
2759 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2760 hide most of the details from the main code, and operates only if
2761 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2762 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2763 operating.
2764
2765 To make it easier for callers to provide specially tailored get/free
2766 functions for this usage, two new functions, pcre_stack_malloc, and
2767 pcre_stack_free, are used. They are always called in strict stacking order,
2768 and the size of block requested is always the same.
2769
2770 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2771 PCRE has been compiled to use the stack or the heap for recursion. The
2772 -C option of pcretest uses this to show which version is compiled.
2773
2774 A new data escape \S, is added to pcretest; it causes the amounts of store
2775 obtained and freed by both kinds of malloc/free at match time to be added
2776 to the output.
2777
2778 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2779 what's available on my current Linux desktop machine.
2780
2781 3. When matching a UTF-8 string, the test for a valid string at the start has
2782 been extended. If start_offset is not zero, PCRE now checks that it points
2783 to a byte that is the start of a UTF-8 character. If not, it returns
2784 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2785 this is necessary because there may be backward assertions in the pattern.
2786 When matching the same subject several times, it may save resources to use
2787 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2788
2789 4. The code for checking the validity of UTF-8 strings has been tightened so
2790 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2791 containing "overlong sequences".
2792
2793 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2794 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2795 should have been "&", but it just so happened that all the cases this let
2796 through by mistake were picked up later in the function.
2797
2798 6. I had used a variable called "isblank" - this is a C99 function, causing
2799 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2800
2801 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2802 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2803 so the version is shown; (c) stop "make test" echoing "./RunTest".
2804
2805 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2806
2807 9. The prototype for memmove() for systems that don't have it was using
2808 size_t, but the inclusion of the header that defines size_t was later. I've
2809 moved the #includes for the C headers earlier to avoid this.
2810
2811 10. Added some adjustments to the code to make it easier to compiler on certain
2812 special systems:
2813
2814 (a) Some "const" qualifiers were missing.
2815 (b) Added the macro EXPORT before all exported functions; by default this
2816 is defined to be empty.
2817 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2818 that it reads its output file name as an argument instead of writing
2819 to the standard output and assuming this can be redirected.
2820
2821 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2822 class containing characters with values greater than 255, PCRE compilation
2823 went into a loop.
2824
2825 12. A recursive reference to a subpattern that was within another subpattern
2826 that had a minimum quantifier of zero caused PCRE to crash. For example,
2827 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2828 recursion. If the recursively-called subpattern itself had a zero repeat,
2829 that was OK.
2830
2831 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2832 buffer into which it was copied (for escape processing) was still set at
2833 1024, so long lines caused crashes.
2834
2835 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2836 "internal error: code overflow...". This applied to any character class
2837 that was followed by a possessive quantifier.
2838
2839 15. Modified the Makefile to add libpcre.la as a prerequisite for
2840 libpcreposix.la because I was told this is needed for a parallel build to
2841 work.
2842
2843 16. If a pattern that contained .* following optional items at the start was
2844 studied, the wrong optimizing data was generated, leading to matching
2845 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2846 matching string must start with a or b or c. The correct conclusion for
2847 this pattern is that a match can start with any character.
2848
2849
2850 Version 4.4 13-Aug-03
2851 ---------------------
2852
2853 1. In UTF-8 mode, a character class containing characters with values between
2854 127 and 255 was not handled correctly if the compiled pattern was studied.
2855 In fixing this, I have also improved the studying algorithm for such
2856 classes (slightly).
2857
2858 2. Three internal functions had redundant arguments passed to them. Removal
2859 might give a very teeny performance improvement.
2860
2861 3. Documentation bug: the value of the capture_top field in a callout is *one
2862 more than* the number of the hightest numbered captured substring.
2863
2864 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2865 in incorrectly linking with a previously installed version. They now link
2866 explicitly with libpcre.la.
2867
2868 5. configure.in no longer needs to recognize Cygwin specially.
2869
2870 6. A problem in pcre.in for Windows platforms is fixed.
2871
2872 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2873 pcretest, it used to include the size of the study block as part of its
2874 output. Unfortunately, the structure contains a field that has a different
2875 size on different hardware architectures. This meant that the tests that
2876 showed this size failed. As the block is currently always of a fixed size,
2877 this information isn't actually particularly useful in pcretest output, so
2878 I have just removed it.
2879
2880 8. Three pre-processor statements accidentally did not start in column 1.
2881 Sadly, there are *still* compilers around that complain, even though
2882 standard C has not required this for well over a decade. Sigh.
2883
2884 9. In pcretest, the code for checking callouts passed small integers in the
2885 callout_data field, which is a void * field. However, some picky compilers
2886 complained about the casts involved for this on 64-bit systems. Now
2887 pcretest passes the address of the small integer instead, which should get
2888 rid of the warnings.
2889
2890 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2891 both compile and run time, and gives an error if an invalid UTF-8 sequence
2892 is found. There is a option for disabling this check in cases where the
2893 string is known to be correct and/or the maximum performance is wanted.
2894
2895 11. In response to a bug report, I changed one line in Makefile.in from
2896
2897 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2898 to
2899 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2900
2901 to look similar to other lines, but I have no way of telling whether this
2902 is the right thing to do, as I do not use Windows. No doubt I'll get told
2903 if it's wrong...
2904
2905
2906 Version 4.3 21-May-03
2907 ---------------------
2908
2909 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2910 Makefile.
2911
2912 2. Some refactoring to improve the quality of the code:
2913
2914 (i) The utf8_table... variables are now declared "const".
2915
2916 (ii) The code for \cx, which used the "case flipping" table to upper case
2917 lower case letters, now just substracts 32. This is ASCII-specific,
2918 but the whole concept of \cx is ASCII-specific, so it seems
2919 reasonable.
2920
2921 (iii) PCRE was using its character types table to recognize decimal and
2922 hexadecimal digits in the pattern. This is silly, because it handles
2923 only 0-9, a-f, and A-F, but the character types table is locale-
2924 specific, which means strange things might happen. A private
2925 table is now used for this - though it costs 256 bytes, a table is
2926 much faster than multiple explicit tests. Of course, the standard
2927 character types table is still used for matching digits in subject
2928 strings against \d.
2929
2930 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2931 ending in _t are). So I've renamed it as ESC_tee.
2932
2933 3. The first argument for regexec() in the POSIX wrapper should have been
2934 defined as "const".
2935
2936 4. Changed pcretest to use malloc() for its buffers so that they can be
2937 Electric Fenced for debugging.
2938
2939 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2940 to read one or more bytes before the start of the subject string. Often this
2941 had no effect on PCRE's behaviour, but in some circumstances it could
2942 provoke a segmentation fault.
2943
2944 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2945 to try to read one or more bytes before the start of the subject string.
2946
2947 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2948 UTF-8 support could misbehave in various ways if the subject string
2949 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2950 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2951 back over UTF-8 characters.)
2952
2953
2954 Version 4.2 14-Apr-03
2955 ---------------------
2956
2957 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2958
2959 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2960 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2961 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2962 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2963 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2964 and BUILD_EXEEXT
2965 Note: automatic setting of the BUILD variables is not yet working
2966 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2967 compile-time but not at link-time
2968 [LINK]: use for linking executables only
2969 make different versions for Windows and non-Windows
2970 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2971 libraries
2972 [LINK_FOR_BUILD]: new variable
2973 [OBJEXT]: use throughout
2974 [EXEEXT]: use throughout
2975 <winshared>: new target
2976 <wininstall>: new target
2977 <dftables.o>: use native compiler
2978 <dftables>: use native linker
2979 <install>: handle Windows platform correctly
2980 <clean>: ditto
2981 <check>: ditto
2982 copy DLL to top builddir before testing
2983
2984 As part of these changes, -no-undefined was removed again. This was reported
2985 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2986 in any case.
2987
2988 3. Some tidies to get rid of compiler warnings:
2989
2990 . In the match_data structure, match_limit was an unsigned long int, whereas
2991 match_call_count was an int. I've made them both unsigned long ints.
2992
2993 . In pcretest the fact that a const uschar * doesn't automatically cast to
2994 a void * provoked a warning.
2995
2996 . Turning on some more compiler warnings threw up some "shadow" variables
2997 and a few more missing casts.
2998
2999 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3000 option, a class that contained a single character with a value between 128
3001 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3002
3003 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3004 option, a class that contained several characters, but with at least one
3005 whose value was between 128 and 255 caused PCRE to crash.
3006
3007
3008 Version 4.1 12-Mar-03
3009 ---------------------
3010
3011 1. Compiling with gcc -pedantic found a couple of places where casts were
3012 needed, and a string in dftables.c that was longer than standard compilers are
3013 required to support.
3014
3015 2. Compiling with Sun's compiler found a few more places where the code could
3016 be tidied up in order to avoid warnings.
3017
3018 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3019 first of these names is deprecated in the latest Autoconf in favour of the name
3020 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3021 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3022 analogy I have changed it to CFLAGS_FOR_BUILD.
3023
3024 4. Added -no-undefined to the linking command in the Makefile, because this is
3025 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3026 linking step for the pcreposix library.
3027
3028 5. PCRE was failing to diagnose the case of two named groups with the same
3029 name.
3030
3031 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3032 literal character that is needed in the subject for a match, and scans along to
3033 ensure that it is present before embarking on the full matching process. This
3034 saves time in cases of nested unlimited repeats that are never going to match.
3035 Problem: the scan can take a lot of time if the subject is very long (e.g.
3036 megabytes), thus penalizing straightforward matches. It is now done only if the
3037 amount of subject to be scanned is less than 1000 bytes.
3038
3039 7. A lesser problem with the same optimization is that it was recording the
3040 first character of an anchored pattern as "needed", thus provoking a search
3041 right along the subject, even when the first match of the pattern was going to
3042 fail. The "needed" character is now not set for anchored patterns, unless it
3043 follows something in the pattern that is of non-fixed length. Thus, it still
3044 fulfils its original purpose of finding quick non-matches in cases of nested
3045 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3046
3047
3048 Version 4.0 17-Feb-03
3049 ---------------------
3050
3051 1. If a comment in an extended regex that started immediately after a meta-item
3052 extended to the end of string, PCRE compiled incorrect data. This could lead to
3053 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3054
3055 2. Moved to autoconf 2.53 and libtool 1.4.2.
3056
3057 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3058 the special perltest8 script is no longer needed - all the tests can be run
3059 from a single perltest script.
3060
3061 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3062 by \s. It has now been removed in PCRE. This means it isn't recognized as
3063 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3064 class [:space:] *does* include VT, thereby creating a mess.
3065
3066 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3067 space and tab.
3068
3069 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3070 its new features into the main test script, reducing the number of scripts.
3071
3072 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3073 were backward compatible, and made the (?i) apply to the whole pattern, as if
3074 /i were given. Now it behaves more logically, and applies the option setting
3075 only to what follows. PCRE has been changed to follow suit. However, if it
3076 finds options settings right at the start of the pattern, it extracts them into
3077 the global options, as before. Thus, they show up in the info data.
3078
3079 8. Added support for the \Q...\E escape sequence. Characters in between are
3080 treated as literals. This is slightly different from Perl in that $ and @ are
3081 also handled as literals inside the quotes. In Perl, they will cause variable
3082 interpolation. Note the following examples:
3083
3084 Pattern PCRE matches Perl matches
3085
3086 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3087 \Qabc\$xyz\E abc\$xyz abc\$xyz
3088 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3089
3090 For compatibility with Perl, \Q...\E sequences are recognized inside character
3091 classes as well as outside them.
3092
3093 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3094 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3095 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3096 signed/unsigned warnings.
3097
3098 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3099 option for pcretest, so I've replaced it by a simple function that does just
3100 that job.
3101
3102 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3103 "pcregrep -".
3104
3105 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3106 Java package. This provides some syntactic sugar for simple cases of what my
3107 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3108 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3109 item, you can use this simplified notation. Note that only makes sense with
3110 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3111 greediness, whatever the setting of the PCRE_UNGREEDY option.
3112
3113 13. A change of greediness default within a pattern was not taking effect at
3114 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3115 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3116 was abstracted outside.
3117
3118 14. PCRE now supports the \G assertion. It is true when the current matching
3119 position is at the start point of the match. This differs from \A when the
3120 starting offset is non-zero. Used with the /g option of pcretest (or similar
3121 code), it works in the same way as it does for Perl's /g option. If all
3122 alternatives of a regex begin with \G, the expression is anchored to the start
3123 match position, and the "anchored" flag is set in the compiled expression.
3124
3125 15. Some bugs concerning the handling of certain option changes within patterns
3126 have been fixed. These applied to options other than (?ims). For example,
3127 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3128 been the other way round. Some of this was related to change 7 above.
3129
3130 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3131 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3132 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3133 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3134
3135 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3136 mode. Unlike ".", it always matches newline, whatever the setting of
3137 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3138 assertions. Perl allows it, but it doesn't (in general) work because it can't
3139 calculate the length of the lookbehind. At least, that's the case for Perl
3140 5.8.0 - I've been told they are going to document that it doesn't work in
3141 future.
3142
3143 18. Added an error diagnosis for escapes that PCRE does not support: these are
3144 \L, \l, \N, \P, \p, \U, \u, and \X.
3145
3146 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3147 reading past the end of the pattern in cases such as /[abcd/.
3148
3149 20. PCRE was getting more memory than necessary for patterns with classes that
3150 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3151
3152 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3153 compiling PCRE for use with Virtual Pascal.
3154
3155 22. Small fix to the Makefile to make it work properly if the build is done
3156 outside the source tree.
3157
3158 23. Added a new extension: a condition to go with recursion. If a conditional
3159 subpattern starts with (?(R) the "true" branch is used if recursion has
3160 happened, whereas the "false" branch is used only at the top level.
3161
3162 24. When there was a very long string of literal characters (over 255 bytes
3163 without UTF support, over 250 bytes with UTF support), the computation of how
3164 much memory was required could be incorrect, leading to segfaults or other
3165 strange effects.
3166
3167 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3168 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3169 there was a subsequent back reference to those brackets. This meant that, for
3170 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3171 possible to check for precisely this case. All we can do is abandon the
3172 optimization if .* occurs inside capturing brackets when there are any back
3173 references whatsoever. (See below for a better fix that came later.)
3174
3175 26. The handling of the optimization for finding the first character of a
3176 non-anchored pattern, and for finding a character that is required later in the
3177 match were failing in some cases. This didn't break the matching; it just
3178 failed to optimize when it could. The way this is done has been re-implemented.
3179
3180 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3181
3182 28. Added a new feature that provides some of the functionality that Perl
3183 provides with (?{...}). The facility is termed a "callout". The way it is done
3184 in PCRE is for the caller to provide an optional function, by setting
3185 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3186 global variable. By default it is unset, which disables all calling out. To get
3187 the function called, the regex must include (?C) at appropriate points. This
3188 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3189 This provides a means of identifying different callout points. When PCRE
3190 reaches such a point in the regex, if pcre_callout has been set, the external
3191 function is called. It is provided with data in a structure called
3192 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3193 matching continues; if it returns a non-zero value, the match at the current
3194 point fails. However, backtracking will occur if possible. [This was changed
3195 later and other features added - see item 49 below.]
3196
3197 29. pcretest is upgraded to test the callout functionality. It provides a
3198 callout function that displays information. By default, it shows the start of
3199 the match and the current position in the text. There are some new data escapes
3200 to vary what happens:
3201
3202 \C+ in addition, show current contents of captured substrings
3203 \C- do not supply a callout function
3204 \C!n return 1 when callout number n is reached
3205 \C!n!m return 1 when callout number n is reached for the mth time
3206
3207 30. If pcregrep was called with the -l option and just a single file name, it
3208 output "<stdin>" if a match was found, instead of the file name.
3209
3210 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3211 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3212 pcre_exec(). This saves a malloc/free per call. The default value of
3213 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3214 when configuring.
3215
3216 32. The default maximum size of a compiled pattern is 64K. There have been a
3217 few cases of people hitting this limit. The code now uses macros to handle the
3218 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3219 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3220 configuring. Tests 2 and 5 work only with 2-byte links because they output
3221 debugging information about compiled patterns.
3222
3223 33. Internal code re-arrangements:
3224
3225 (a) Moved the debugging function for printing out a compiled regex into
3226 its own source file (printint.c) and used #include to pull it into
3227 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3228 separate copies.
3229
3230 (b) Defined the list of op-code names for debugging as a macro in
3231 internal.h so that it is next to the definition of the opcodes.
3232
3233 (c) Defined a table of op-code lengths for simpler skipping along compiled
3234 code. This is again a macro in internal.h so that it is next to the
3235 definition of the opcodes.
3236
3237 34. Added support for recursive calls to individual subpatterns, along the
3238 lines of Robin Houston's patch (but implemented somewhat differently).
3239
3240 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3241 allow it to read and process whole directories in Win32. This code was
3242 contributed by Lionel Fourquaux; it has not been tested by me.
3243
3244 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3245 used to name a group. Names consist of alphanumerics and underscores, and must
3246 be unique. Back references use the syntax (?P=name) and recursive calls use
3247 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3248 numbers. The function pcre_fullinfo() can be used after compilation to extract
3249 a name/number map. There are three relevant calls:
3250
3251 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3252 PCRE_INFO_NAMECOUNT yields the number of entries
3253 PCRE_INFO_NAMETABLE yields a pointer to the map.
3254
3255 The map is a vector of fixed-size entries. The size of each entry depends on
3256 the length of the longest name used. The first two bytes of each entry are the
3257 group number, most significant byte first. There follows the corresponding
3258 name, zero terminated. The names are in alphabetical order.
3259
3260 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3261 case instead of 255. Making it the same both with and without UTF-8 support
3262 means that the same test output works with both.
3263
3264 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3265 calling malloc() with a zero argument.
3266
3267 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3268 optimization. I've improved things by keeping a bitmap of backreferences with
3269 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3270 fact referenced, the optimization can be applied. It is unlikely that a
3271 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3272 the match to follow \n) will appear inside brackets with a number greater than
3273 31, but if it does, any back reference > 31 suppresses the optimization.
3274
3275 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3276 of disabling numbered capturing parentheses. Any opening parenthesis that is
3277 not followed by ? behaves as if it were followed by ?: but named parentheses
3278 can still be used for capturing (and they will acquire numbers in the usual
3279 way).
3280
3281 41. Redesigned the return codes from the match() function into yes/no/error so
3282 that errors can be passed back from deep inside the nested calls. A malloc
3283 failure while inside a recursive subpattern call now causes the
3284 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3285
3286 42. It is now possible to set a limit on the number of times the match()
3287 function is called in a call to pcre_exec(). This facility makes it possible to
3288 limit the amount of recursion and backtracking, though not in a directly
3289 obvious way, because the match() function is used in a number of different
3290 circumstances. The count starts from zero for each position in the subject
3291 string (for non-anchored patterns). The default limit is, for compatibility, a
3292 large number, namely 10 000 000. You can change this in two ways:
3293
3294 (a) When configuring PCRE before making, you can use --with-match-limit=n
3295 to set a default value for the compiled library.
3296
3297 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3298 a different value is set. See 45 below.
3299
3300 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3301
3302 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3303 of things that can be changed at compile time. The first argument specifies
3304 what is wanted and the second points to where the information is to be placed.
3305 The current list of available information is:
3306
3307 PCRE_CONFIG_UTF8
3308
3309 The output is an integer that is set to one if UTF-8 support is available;
3310 otherwise it is set to zero.
3311
3312 PCRE_CONFIG_NEWLINE
3313
3314 The output is an integer that it set to the value of the code that is used for
3315 newline. It is either LF (10) or CR (13).
3316
3317 PCRE_CONFIG_LINK_SIZE
3318
3319 The output is an integer that contains the number of bytes used for internal
3320 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3321
3322 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3323
3324 The output is an integer that contains the threshold above which the POSIX
3325 interface uses malloc() for output vectors. See item 31 above.
3326
3327 PCRE_CONFIG_MATCH_LIMIT
3328
3329 The output is an unsigned integer that contains the default limit of the number
3330 of match() calls in a pcre_exec() execution. See 42 above.
3331
3332 44. pcretest has been upgraded by the addition of the -C option. This causes it
3333 to extract all the available output from the new pcre_config() function, and to
3334 output it. The program then exits immediately.
3335
3336 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3337 order to support additional features. One way would have been to define
3338 pcre_exec2() (for example) with extra arguments, but this would not have been
3339 extensible, and would also have required all calls to the original function to
3340 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3341 is used for passing in "extra" data from pcre_study().
3342
3343 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3344 contains the following fields:
3345
3346 flags a bitmap indicating which of the following fields are set
3347 study_data opaque data from pcre_study()
3348 match_limit a way of specifying a limit on match() calls for a specific
3349 call to pcre_exec()
3350 callout_data data for callouts (see 49 below)
3351
3352 The flag bits are also defined in pcre.h, and are
3353
3354 PCRE_EXTRA_STUDY_DATA
3355 PCRE_EXTRA_MATCH_LIMIT
3356 PCRE_EXTRA_CALLOUT_DATA
3357
3358 The pcre_study() function now returns one of these new pcre_extra blocks, with
3359 the actual study data pointed to by the study_data field, and the
3360 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3361 before. That is, this change is entirely upwards-compatible and requires no
3362 change to existing code.
3363
3364 If you want to pass in additional data to pcre_exec(), you can either place it
3365 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3366 block.
3367
3368 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3369 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3370 times with different match limits, until it finds the minimum value needed for
3371 pcre_exec() to complete. The value is then output. This can be instructive; for
3372 most simple matches the number is quite small, but for pathological cases it
3373 gets very large very quickly.
3374
3375 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3376 returns the size of the data block pointed to by the study_data field in a
3377 pcre_extra block, that is, the value that was passed as the argument to
3378 pcre_malloc() when PCRE was getting memory in which to place the information
3379 created by pcre_study(). The fourth argument should point to a size_t variable.
3380 pcretest has been extended so that this information is shown after a successful
3381 pcre_study() call when information about the compiled regex is being displayed.
3382
3383 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3384 because what follows is always an absolute path. (Later: it turns out that this
3385 is more than cosmetic for MinGW, because it doesn't like empty path
3386 components.)
3387
3388 49. Some changes have been made to the callout feature (see 28 above):
3389
3390 (i) A callout function now has three choices for what it returns:
3391
3392 0 => success, carry on matching
3393 > 0 => failure at this point, but backtrack if possible
3394 < 0 => serious error, return this value from pcre_exec()
3395
3396 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3397 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3398 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3399 use by callout functions. It will never be used by PCRE itself.
3400
3401 (ii) The pcre_extra structure (see 45 above) has a void * field called
3402 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3403 pcre_callout_block structure has a field of the same name. The contents of
3404 the field passed in the pcre_extra structure are passed to the callout
3405 function in the corresponding field in the callout block. This makes it
3406 easier to use the same callout-containing regex from multiple threads. For
3407 testing, the pcretest program has a new data escape
3408
3409 \C*n pass the number n (may be negative) as callout_data
3410
3411 If the callout function in pcretest receives a non-zero value as
3412 callout_data, it returns that value.
3413
3414 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3415 there were some redundant $(CFLAGS) in commands that are now specified as
3416 $(LINK), which already includes $(CFLAGS).
3417
3418 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3419 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3420 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3421 one-byte characters throughout. Note that case-insensitive matching applies
3422 only to characters whose values are less than 256. PCRE doesn't support the
3423 notion of cases for higher-valued characters.
3424
3425 (i) A character class whose characters are all within 0-255 is handled as
3426 a bit map, and the map is inverted for negative classes. Previously, a
3427 character > 255 always failed to match such a class; however it should
3428 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3429
3430 (ii) A negated character class with a single character < 255 is coded as
3431 "not this character" (OP_NOT). This wasn't working properly when the test
3432 character was multibyte, either singly or repeated.
3433
3434 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3435 mode, for example: \x{100}{2,3}.
3436
3437 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3438 singly or repeated) now correctly test multibyte characters. However,
3439 PCRE doesn't recognize any characters with values greater than 255 as
3440 digits, spaces, or word characters. Such characters always match \D, \S,
3441 and \W, and never match \d, \s, or \w.
3442
3443 (v) Classes may now contain characters and character ranges with values
3444 greater than 255. For example: [ab\x{100}-\x{400}].
3445
3446 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3447 PCRE in UTF-8 mode.
3448
3449 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3450 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3451 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3452 value.)
3453
3454 53. The single man page has become too large. I have therefore split it up into
3455 a number of separate man pages. These also give rise to individual HTML pages;
3456 these are now put in a separate directory, and there is an index.html page that
3457 lists them all. Some hyperlinking between the pages has been installed.
3458
3459 54. Added convenience functions for handling named capturing parentheses.
3460
3461 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3462 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3463 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3464 are faulted.
3465
3466 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3467 calling configure. These values are used when compiling the dftables.c program
3468 which is run to generate the source of the default character tables. They
3469 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3470 you will need to set these values.
3471
3472 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3473
3474
3475 Version 3.9 02-Jan-02
3476 ---------------------
3477
3478 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3479
3480 2. If --disable-static was given, the building process failed when trying to
3481 build pcretest and pcregrep. (For some reason it was using libtool to compile
3482 them, which is not right, as they aren't part of the library.)
3483
3484
3485 Version 3.8 18-Dec-01
3486 ---------------------
3487
3488 1. The experimental UTF-8 code was completely screwed up. It was packing the
3489 bytes in the wrong order. How dumb can you get?
3490
3491
3492 Version 3.7 29-Oct-01
3493 ---------------------
3494
3495 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3496 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3497 this didn't happen under Solaris 8, where I normally test things.
3498
3499 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3500 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3501 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3502 this fix an hour or so after the initial 3.7 release.)
3503
3504
3505 Version 3.6 23-Oct-01
3506 ---------------------
3507
3508 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3509 offsets passed as NULL with zero offset count.
3510
3511 2. The config.guess and config.sub files had not been updated when I moved to
3512 the latest autoconf.
3513
3514
3515 Version 3.5 15-Aug-01
3516 ---------------------
3517
3518 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3519 had been forgotten.
3520
3521 2. By using declared but undefined structures, we can avoid using "void"
3522 definitions in pcre.h while keeping the internal definitions of the structures
3523 private.
3524
3525 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3526 user point of view, this means that both static and shared libraries are built
3527 by default, but this can be individually controlled. More of the work of
3528 handling this static/shared cases is now inside libtool instead of PCRE's make
3529 file.
3530
3531 4. The pcretest utility is now installed along with pcregrep because it is
3532 useful for users (to test regexs) and by doing this, it automatically gets
3533 relinked by libtool. The documentation has been turned into a man page, so
3534 there are now .1, .txt, and .html versions in /doc.
3535
3536 5. Upgrades to pcregrep:
3537 (i) Added long-form option names like gnu grep.
3538 (ii) Added --help to list all options with an explanatory phrase.
3539 (iii) Added -r, --recursive to recurse into sub-directories.
3540 (iv) Added -f, --file to read patterns from a file.
3541
3542 6. pcre_exec() was referring to its "code" argument before testing that
3543 argument for NULL (and giving an error if it was NULL).
3544
3545 7. Upgraded Makefile.in to allow for compiling in a different directory from
3546 the source directory.
3547
3548 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3549 options bits, the pointer it was passed was to an int instead of to an unsigned
3550 long int. This mattered only on 64-bit systems.
3551
3552 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3553 generated) instead of pcre.in, which it its source. Also made the same change
3554 in several of the .c files.
3555
3556 10. A new release of gcc defines printf() as a macro, which broke pcretest
3557 because it had an ifdef in the middle of a string argument for printf(). Fixed
3558 by using separate calls to printf().
3559
3560 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3561 script, to force use of CR or LF instead of \n in the source. On non-Unix
3562 systems, the value can be set in config.h.
3563
3564 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3565 absolute limit. Changed the text of the error message to make this clear, and
3566 likewise updated the man page.
3567
3568 13. The limit of 99 on the number of capturing subpatterns has been removed.
3569 The new limit is 65535, which I hope will not be a "real" limit.
3570
3571
3572 Version 3.4 22-Aug-00
3573 ---------------------
3574
3575 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3576
3577 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3578
3579
3580 Version 3.3 01-Aug-00
3581 ---------------------
3582
3583 1. If an octal character was given, but the value was greater than \377, it
3584 was not getting masked to the least significant bits, as documented. This could
3585 lead to crashes in some systems.
3586
3587 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3588 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3589
3590 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3591 These just pass their arguments on to (pcre_free)(), but they are provided
3592 because some uses of PCRE bind it to non-C systems that can call its functions,
3593 but cannot call free() or pcre_free() directly.
3594
3595 4. Add "make test" as a synonym for "make check". Corrected some comments in
3596 the Makefile.
3597
3598 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3599 Makefile.
3600
3601 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3602 command called pgrep for grepping around the active processes.
3603
3604 7. Added the beginnings of support for UTF-8 character strings.
3605
3606 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3607 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3608 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3609 out for the ar command.)
3610
3611
3612 Version 3.2 12-May-00
3613 ---------------------
3614
3615 This is purely a bug fixing release.
3616
3617 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3618 of ZA. This was just one example of several cases that could provoke this bug,
3619 which was introduced by change 9 of version 2.00. The code for breaking
3620 infinite loops after an iteration that matches an empty string was't working
3621 correctly.
3622
3623 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3624 when matched against abbab (for example). After matching an empty string, it
3625 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3626 caused it to match further down the string than it should.
3627
3628 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3629 was there because it doesn't seem to be needed, and it causes trouble on some
3630 systems, as it is not a Standard C header. It has been removed.
3631
3632 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3633 were reported on the Macintosh. The changes were from
3634
3635 while ((c = *(++ptr)) != 0 && c != '\n');
3636 to
3637 while ((c = *(++ptr)) != 0 && c != '\n') ;
3638
3639 Totally extraordinary, but if that's what it takes...
3640
3641 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3642 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3643 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3644 assumes the way PCRE uses memmove() (always moving upwards).
3645
3646 6. PCRE is being used in one environment where strchr() is not available. There
3647 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3648 faster code anyway.
3649
3650
3651 Version 3.1 09-Feb-00
3652 ---------------------
3653
3654 The only change in this release is the fixing of some bugs in Makefile.in for
3655 the "install" target:
3656
3657 (1) It was failing to install pcreposix.h.
3658
3659 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3660
3661
3662 Version 3.0 01-Feb-00
3663 ---------------------
3664
3665 1. Add support for the /+ modifier to perltest (to output $` like it does in
3666 pcretest).
3667
3668 2. Add support for the /g modifier to perltest.
3669
3670 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3671 matches null strings.
3672
3673 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3674 pattern. Perl treats empty patterns specially - it reuses the most recent
3675 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3676 effect.
3677
3678 5. The POSIX interface was broken in that it was just handing over the POSIX
3679 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3680 required a bigger vector, with some working space on the end. This means that
3681 the POSIX wrapper now has to get and free some memory, and copy the results.
3682
3683 6. Added some simple autoconf support, placing the test data and the
3684 documentation in separate directories, re-organizing some of the
3685 information files, and making it build pcre-config (a GNU standard). Also added
3686 libtool support for building PCRE as a shared library, which is now the
3687 default.
3688
3689 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3690 09 are not valid octal constants. Single digits will be used for minor values
3691 less than 10.
3692
3693 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3694 existing programs that set these in the POSIX interface can use PCRE without
3695 modification.
3696
3697 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3698 return all that pcre_info() returns, plus additional data. The pcre_info()
3699 function is retained for compatibility, but is considered to be obsolete.
3700
3701 10. Added experimental recursion feature (?R) to handle one common case that
3702 Perl 5.6 will be able to do with (?p{...}).
3703
3704 11. Added support for POSIX character classes like [:alpha:], which Perl is
3705 adopting.
3706
3707
3708 Version 2.08 31-Aug-99
3709 ----------------------
3710
3711 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3712 trying to match at the startoffset position, but instead was moving forward to
3713 the next newline as if a previous match had failed.
3714
3715 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3716 and could get into a loop if a null string was matched other than at the start
3717 of the subject.
3718
3719 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3720 be distinguished at compile time, and for completeness also added PCRE_DATE.
3721
3722 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3723 in GnuWin32 environments.
3724
3725
3726 Version 2.07 29-Jul-99
3727 ----------------------
3728
3729 1. The documentation is now supplied in plain text form and HTML as well as in
3730 the form of man page sources.
3731
3732 2. C++ compilers don't like assigning (void *) values to other pointer types.
3733 In particular this affects malloc(). Although there is no problem in Standard
3734 C, I've put in casts to keep C++ compilers happy.
3735
3736 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3737 should be (const char *).
3738
3739 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3740 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3741 However, I haven't made this a standard facility. The documentation doesn't
3742 mention it, and the Makefile doesn't support it.
3743
3744 5. The Makefile now contains an "install" target, with editable destinations at
3745 the top of the file. The pcretest program is not installed.
3746
3747 6. pgrep -V now gives the PCRE version number and date.
3748
3749 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3750 causing the entire string to be ignored, instead of just the last character.
3751
3752 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3753 non-matching string, it can take a very, very long time, even for strings of
3754 quite modest length, because of the nested recursion. PCRE now does better in
3755 some of these cases. It does this by remembering the last required literal
3756 character in the pattern, and pre-searching the subject to ensure it is present
3757 before running the real match. In other words, it applies a heuristic to detect
3758 some types of certain failure quickly, and in the above example, if presented
3759 with a string that has no trailing " it gives "no match" very quickly.
3760
3761 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3762 other alternatives are tried instead.
3763
3764
3765 Version 2.06 09-Jun-99
3766 ----------------------
3767
3768 1. Change pcretest's output for amount of store used to show just the code
3769 space, because the remainder (the data block) varies in size between 32-bit and
3770 64-bit systems.
3771
3772 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3773 start matching at. This allows lookbehinds to work when searching for multiple
3774 occurrences in a string.
3775
3776 3. Added additional options to pcretest for testing multiple occurrences:
3777
3778 /+ outputs the rest of the string that follows a match
3779 /g loops for multiple occurrences, using the new startoffset argument
3780 /G loops for multiple occurrences by passing an incremented pointer
3781
3782 4. PCRE wasn't doing the "first character" optimization for patterns starting
3783 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3784 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3785 the letter 'x'. On long subject strings, this gives a significant speed-up.
3786
3787
3788 Version 2.05 21-Apr-99
3789 ----------------------
3790
3791 1. Changed the type of magic_number from int to long int so that it works
3792 properly on 16-bit systems.
3793
3794 2. Fixed a bug which caused patterns starting with .* not to work correctly
3795 when the subject string contained newline characters. PCRE was assuming
3796 anchoring for such patterns in all cases, which is not correct because .* will
3797 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3798 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3799 must be retried after every newline in the subject.
3800
3801
3802 Version 2.04 18-Feb-99
3803 ----------------------
3804
3805 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3806 computation of the store needed to hold the pattern was incorrect (too large).
3807 If such patterns were nested a few deep, this could multiply and become a real
3808 problem.
3809
3810 2. Added /M option to pcretest to show the memory requirement of a specific
3811 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3812
3813 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3814 compiled in such a way that the backtracking after subsequent failure was
3815 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3816 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3817
3818
3819 Version 2.03 02-Feb-99
3820 ----------------------
3821
3822 1. Fixed typo and small mistake in man page.
3823
3824 2. Added 4th condition (GPL supersedes if conflict) and created separate
3825 LICENCE file containing the conditions.
3826
3827 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3828 Perl, that is the internal \ allows the delimiter to be included in the
3829 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3830 the final delimiter, add \ to the end of the pattern (to test the error).
3831
3832 4. Added the convenience functions for extracting substrings after a successful
3833 match. Updated pcretest to make it able to test these functions.
3834
3835
3836 Version 2.02 14-Jan-99
3837 ----------------------
3838
3839 1. Initialized the working variables associated with each extraction so that
3840 their saving and restoring doesn't refer to uninitialized store.
3841
3842 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3843 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3844 fix the problem.
3845
3846 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3847 calls, and wasn't printing the correct value for compiling calls. Increased the
3848 default value of LOOPREPEAT, and the number of significant figures in the
3849 times.
3850
3851 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3852
3853 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3854 a building problem on Windows NT with a FAT file system.
3855
3856
3857 Version 2.01 21-Oct-98
3858 ----------------------
3859
3860 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3861 to character tables built by pcre_maketables() in the current locale. If NULL
3862 is passed, the default tables are used.
3863
3864
3865 Version 2.00 24-Sep-98
3866 ----------------------
3867
3868 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3869 it any more.
3870
3871 2. Allow quantification of (?>) groups, and make it work correctly.
3872
3873 3. The first character computation wasn't working for (?>) groups.
3874
3875 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3876 end of the subject) and add 5.005's \z, which really does match only at the
3877 very end of the subject.
3878
3879 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3880
3881 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3882 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3883 localized options. All options to pcre_study() were also removed.
3884
3885 7. Add other new features from 5.005:
3886
3887 $(?<= positive lookbehind
3888 $(?<! negative lookbehind
3889 (?imsx-imsx) added the unsetting capability
3890 such a setting is global if at outer level; local otherwise
3891 (?imsx-imsx:) non-capturing groups with option setting
3892 (?(cond)re|re) conditional pattern matching
3893
3894 A backreference to itself in a repeated group matches the previous
3895 captured string.
3896
3897 8. General tidying up of studying (both automatic and via "study")
3898 consequential on the addition of new assertions.
3899
3900 9. As in 5.005, unlimited repeated groups that could match an empty substring
3901 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3902 runtime if any iteration does actually match an empty substring.
3903
3904 10. Include the RunTest script in the distribution.
3905
3906 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3907 discrepancies, some of which were old and were also with respect to 5.004. They
3908 have now been fixed.
3909
3910
3911 Version 1.09 28-Apr-98
3912 ----------------------
3913
3914 1. A negated single character class followed by a quantifier with a minimum
3915 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3916 program crashes, or just wrong answers. This did not apply to negated classes
3917 containing more than one character, or to minima other than one.
3918
3919
3920 Version 1.08 27-Mar-98
3921 ----------------------
3922
3923 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3924
3925 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3926 latter must appear before anything that relies on it in the pattern.
3927
3928
3929 Version 1.07 16-Feb-98
3930 ----------------------
3931
3932 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3933 repeat of a potentially empty string).
3934
3935
3936 Version 1.06 23-Jan-98
3937 ----------------------
3938
3939 1. Added Markus Oberhumer's little patches for C++.
3940
3941 2. Literal strings longer than 255 characters were broken.
3942
3943
3944 Version 1.05 23-Dec-97
3945 ----------------------
3946
3947 1. Negated character classes containing more than one character were failing if
3948 PCRE_CASELESS was set at run time.
3949
3950
3951 Version 1.04 19-Dec-97
3952 ----------------------
3953
3954 1. Corrected the man page, where some "const" qualifiers had been omitted.
3955
3956 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3957 input syntax.
3958
3959 3. Fixed memory leak which occurred when a regex with back references was
3960 matched with an offsets vector that wasn't big enough. The temporary memory
3961 that is used in this case wasn't being freed if the match failed.
3962
3963 4. Tidied pcretest to ensure it frees memory that it gets.
3964
3965 5. Temporary memory was being obtained in the case where the passed offsets
3966 vector was exactly big enough.
3967
3968 6. Corrected definition of offsetof() from change 5 below.
3969
3970 7. I had screwed up change 6 below and broken the rules for the use of
3971 setjmp(). Now fixed.
3972
3973
3974 Version 1.03 18-Dec-97
3975 ----------------------
3976
3977 1. A erroneous regex with a missing opening parenthesis was correctly
3978 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3979 on some systems.
3980
3981 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3982 it was reported that one broken compiler failed on the former because "code" is
3983 also an independent variable.
3984
3985 3. The erroneous regex a[]b caused an array overrun reference.
3986
3987 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3988 fail on data ending with that character. (It was going on too far, and checking
3989 the next character, typically a binary zero.) This was specific to the
3990 optimized code for single-character negative classes.
3991
3992 5. Added a contributed patch from the TIN world which does the following:
3993
3994 + Add an undef for memmove, in case the the system defines a macro for it.
3995
3996 + Add a definition of offsetof(), in case there isn't one. (I don't know
3997 the reason behind this - offsetof() is part of the ANSI standard - but
3998 it does no harm).
3999
4000 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4001 most of the places where whitespace preceded '#'. I have given up and
4002 allowed the remaining 2 cases to be at the margin.
4003
4004 + Rename some variables in pcre to eliminate shadowing. This seems very
4005 pedantic, but does no harm, of course.
4006
4007 6. Moved the call to setjmp() into its own function, to get rid of warnings
4008 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4009
4010 7. Constructs such as \d{8,} were compiling into the equivalent of
4011 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4012 outcome, but in this particular case used more store than had been allocated,
4013 which caused the bug to be discovered because it threw up an internal error.
4014
4015 8. The debugging code in both pcre and pcretest for outputting the compiled
4016 form of a regex was going wrong in the case of back references followed by
4017 curly-bracketed repeats.
4018
4019
4020 Version 1.02 12-Dec-97
4021 ----------------------
4022
4023 1. Typos in pcre.3 and comments in the source fixed.
4024
4025 2. Applied a contributed patch to get rid of places where it used to remove
4026 'const' from variables, and fixed some signed/unsigned and uninitialized
4027 variable warnings.
4028
4029 3. Added the "runtest" target to Makefile.
4030
4031 4. Set default compiler flag to -O2 rather than just -O.
4032
4033
4034 Version 1.01 19-Nov-97
4035 ----------------------
4036
4037 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4038 like /([ab]*)*/, that is, for classes with more than one character in them.
4039
4040 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4041 as /((?>a*))*/ (a PCRE_EXTRA facility).
4042
4043
4044 Version 1.00 18-Nov-97
4045 ----------------------
4046
4047 1. Added compile-time macros to support systems such as SunOS4 which don't have
4048 memmove() or strerror() but have other things that can be used instead.
4049
4050 2. Arranged that "make clean" removes the executables.
4051
4052
4053 Version 0.99 27-Oct-97
4054 ----------------------
4055
4056 1. Fixed bug in code for optimizing classes with only one character. It was
4057 initializing a 32-byte map regardless, which could cause it to run off the end
4058 of the memory it had got.
4059
4060 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4061
4062
4063 Version 0.98 22-Oct-97
4064 ----------------------
4065
4066 1. Fixed bug in code for handling temporary memory usage when there are more
4067 back references than supplied space in the ovector. This could cause segfaults.
4068
4069
4070 Version 0.97 21-Oct-97
4071 ----------------------
4072
4073 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4074
4075 2. Optimized negated single characters not to use a bit map.
4076
4077 3. Brought error texts together as macro definitions; clarified some of them;
4078 fixed one that was wrong - it said "range out of order" when it meant "invalid
4079 escape sequence".
4080
4081 4. Changed some char * arguments to const char *.
4082
4083 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4084
4085 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4086 pcretest.
4087
4088
4089 Version 0.96 16-Oct-97
4090 ----------------------
4091
4092 1. Added a simple "pgrep" utility to the distribution.
4093
4094 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4095 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4096 where "ddd" means "one or more decimal digits".
4097
4098 3. Fixed serious bug. If a pattern had a back reference, but the call to
4099 pcre_exec() didn't supply a large enough ovector to record the related
4100 identifying subpattern, the match always failed. PCRE now remembers the number
4101 of the largest back reference, and gets some temporary memory in which to save
4102 the offsets during matching if necessary, in order to ensure that
4103 backreferences always work.
4104
4105 4. Increased the compatibility with Perl in a number of ways:
4106
4107 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4108 to request this handling. The option can be set at compile or exec time.
4109
4110 (b) $ matches before a terminating newline by default; an option
4111 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4112 mode). The option can be set at compile or exec time.
4113
4114 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4115 the same as Perl's. If the decimal number it represents is less than 10
4116 or there aren't that many previous left capturing parentheses, an octal
4117 escape is read. Inside a character class, it's always an octal escape,
4118 even if it is a single digit.
4119
4120 (d) An escaped but undefined alphabetic character is taken as a literal,
4121 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4122 escapes.
4123
4124 (e) {0} is now permitted. (The previous item is removed from the compiled
4125 pattern).
4126
4127 5. Changed all the names of code files so that the basic parts are no longer
4128 than 10 characters, and abolished the teeny "globals.c" file.
4129
4130 6. Changed the handling of character classes; they are now done with a 32-byte
4131 bit map always.
4132
4133 7. Added the -d and /D options to pcretest to make it possible to look at the
4134 internals of compilation without having to recompile pcre.
4135
4136
4137 Version 0.95 23-Sep-97
4138 ----------------------
4139
4140 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4141 \x20 at the start of a run of normal characters. These were being treated as
4142 real characters, instead of the source characters being re-checked.
4143
4144
4145 Version 0.94 18-Sep-97
4146 ----------------------
4147
4148 1. The functions are now thread-safe, with the caveat that the global variables
4149 containing pointers to malloc() and free() or alternative functions are the
4150 same for all threads.
4151
4152 2. Get pcre_study() to generate a bitmap of initial characters for non-
4153 anchored patterns when this is possible, and use it if passed to pcre_exec().
4154
4155
4156 Version 0.93 15-Sep-97
4157 ----------------------
4158
4159 1. /(b)|(:+)/ was computing an incorrect first character.
4160
4161 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4162 but not actually doing anything yet.
4163
4164 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4165 as Perl does (e.g. [-az] or [az-]).
4166
4167 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4168 all possible positions.
4169
4170 5. Split up into different modules to avoid including unneeded functions in a
4171 compiled binary. However, compile and exec are still in one module. The "study"
4172 function is split off.
4173
4174 6. The character tables are now in a separate module whose source is generated
4175 by an auxiliary program - but can then be edited by hand if required. There are
4176 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4177 toupper() in the code.
4178
4179 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4180 make them global. Abolish the function for setting them, as the caller can now
4181 set them directly.
4182
4183
4184 Version 0.92 11-Sep-97
4185 ----------------------
4186
4187 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4188 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4189
4190 2. Caseless matching was not working in character classes if the characters in
4191 the pattern were in upper case.
4192
4193 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4194
4195 4. Make PCRE_ANCHORED public and accept as a compile option.
4196
4197 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4198 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4199 pass them.
4200
4201 6. Give an error if bad option bits passed at compile or run time.
4202
4203 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4204 pcretest to cause it to pass that flag.
4205
4206 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4207 options, and the first character, if set.
4208
4209 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4210
4211
4212 Version 0.91 10-Sep-97
4213 ----------------------
4214
4215 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4216 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4217
4218 2. PCRE was looping on encountering an indefinitely repeated back reference to
4219 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4220 Perl does - treats the match as successful.
4221
4222 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5