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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 606 - (show annotations) (download)
Mon Jun 6 17:46:22 2011 UTC (3 years, 5 months ago) by ph10
File size: 202757 byte(s)
Tidy the API for _pcre_valid_utf8() to a more suitable form for a future public 
release. Also make -s in pcretest force a study for every regex.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12