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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 602 - (show annotations)
Wed May 25 08:29:03 2011 UTC (4 years, 1 month ago) by ph10
File size: 201225 byte(s)
Error occurred while calculating annotation data.
Remove OP_OPT by handling /i and /m entirely at compile time. Fixes bug with 
patterns like /(?i:([^b]))(?1)/, where the /i option was mishandled.
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
57 Version 8.12 15-Jan-2011
58 ------------------------
59
60 1. Fixed some typos in the markup of the man pages, and wrote a script that
61 checks for such things as part of the documentation building process.
62
63 2. On a big-endian 64-bit system, pcregrep did not correctly process the
64 --match-limit and --recursion-limit options (added for 8.11). In
65 particular, this made one of the standard tests fail. (The integer value
66 went into the wrong half of a long int.)
67
68 3. If the --colour option was given to pcregrep with -v (invert match), it
69 did strange things, either producing crazy output, or crashing. It should,
70 of course, ignore a request for colour when reporting lines that do not
71 match.
72
73 4. Another pcregrep bug caused similar problems if --colour was specified with
74 -M (multiline) and the pattern match finished with a line ending.
75
76 5. In pcregrep, when a pattern that ended with a literal newline sequence was
77 matched in multiline mode, the following line was shown as part of the
78 match. This seems wrong, so I have changed it.
79
80 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
81 the check for further matches in the same line (so they could be coloured)
82 to overrun the end of the current line. If another match was found, it was
83 incorrectly shown (and then shown again when found in the next line).
84
85 7. If pcregrep was compiled under Windows, there was a reference to the
86 function pcregrep_exit() before it was defined. I am assuming this was
87 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
88 reported by a user. I've moved the definition above the reference.
89
90
91 Version 8.11 10-Dec-2010
92 ------------------------
93
94 1. (*THEN) was not working properly if there were untried alternatives prior
95 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
96 backtracked to try for "b" instead of moving to the next alternative branch
97 at the same level (in this case, to look for "c"). The Perl documentation
98 is clear that when (*THEN) is backtracked onto, it goes to the "next
99 alternative in the innermost enclosing group".
100
101 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
102 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
103 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
104 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
105 (*THEN).
106
107 3. If \s appeared in a character class, it removed the VT character from
108 the class, even if it had been included by some previous item, for example
109 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
110 of \s, but is part of the POSIX "space" class.)
111
112 4. A partial match never returns an empty string (because you can always
113 match an empty string at the end of the subject); however the checking for
114 an empty string was starting at the "start of match" point. This has been
115 changed to the "earliest inspected character" point, because the returned
116 data for a partial match starts at this character. This means that, for
117 example, /(?<=abc)def/ gives a partial match for the subject "abc"
118 (previously it gave "no match").
119
120 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
121 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
122 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
123 has an implication that the given string is incomplete (because a partial
124 match is preferred over a full match). For this reason, these items now
125 give a partial match in this situation. [Aside: previously, the one case
126 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
127 match rather than a full match, which was wrong by the old rules, but is
128 now correct.]
129
130 6. There was a bug in the handling of #-introduced comments, recognized when
131 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
132 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
133 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
134 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
135 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
136 places in pcre_compile().
137
138 7. Related to (6) above, when pcre_compile() was skipping #-introduced
139 comments when looking ahead for named forward references to subpatterns,
140 the only newline sequence it recognized was NL. It now handles newlines
141 according to the set newline convention.
142
143 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
144 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
145 cater for a lack of strerror(). These oversights have been fixed.
146
147 9. Added --match-limit and --recursion-limit to pcregrep.
148
149 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
150
151 11. When the -o option was used, pcregrep was setting a return code of 1, even
152 when matches were found, and --line-buffered was not being honoured.
153
154 12. Added an optional parentheses number to the -o and --only-matching options
155 of pcregrep.
156
157 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
158 can match an empty string. The code to do it in pcretest and pcredemo
159 needed fixing:
160
161 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
162 only one byte after an empty string match just before CRLF (this case
163 just got forgotten; "any" and "anycrlf" were OK).
164
165 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
166 mode when an empty string match preceded an ASCII character followed by
167 a non-ASCII character. (The code for advancing by one character rather
168 than one byte was nonsense.)
169
170 (c) The pcredemo.c sample program did not have any code at all to handle
171 the cases when CRLF is a valid newline sequence.
172
173 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
174 as a starting offset was within the subject string. There is now a new
175 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
176 negative or greater than the length of the string. In order to test this,
177 pcretest is extended to allow the setting of negative starting offsets.
178
179 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
180 starting offset points to the beginning of a UTF-8 character was
181 unnecessarily clumsy. I tidied it up.
182
183 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
184 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
185
186 17. Nobody had reported that the --include_dir option, which was added in
187 release 7.7 should have been called --include-dir (hyphen, not underscore)
188 for compatibility with GNU grep. I have changed it to --include-dir, but
189 left --include_dir as an undocumented synonym, and the same for
190 --exclude-dir, though that is not available in GNU grep, at least as of
191 release 2.5.4.
192
193 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
194 characters from a string of bytes) have been redefined so as not to use
195 loops, in order to improve performance in some environments. At the same
196 time, I abstracted some of the common code into auxiliary macros to save
197 repetition (this should not affect the compiled code).
198
199 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
200 compile-time error is now given if \c is not followed by an ASCII
201 character, that is, a byte less than 128. (In EBCDIC mode, the code is
202 different, and any byte value is allowed.)
203
204 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
205 START_OPTIMIZE option, which is now allowed at compile time - but just
206 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
207 to pcregrep and other applications that have no direct access to PCRE
208 options. The new /Y option in pcretest sets this option when calling
209 pcre_compile().
210
211 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
212 back references. Groups containing recursive back references were forced to
213 be atomic by that change, but in the case of named groups, the amount of
214 memory required was incorrectly computed, leading to "Failed: internal
215 error: code overflow". This has been fixed.
216
217 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
218 pcretest.c, to avoid build problems in some Borland environments.
219
220
221 Version 8.10 25-Jun-2010
222 ------------------------
223
224 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
225 THEN.
226
227 2. (*ACCEPT) was not working when inside an atomic group.
228
229 3. Inside a character class, \B is treated as a literal by default, but
230 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
231 causes the error). The code is unchanged, but I tidied the documentation.
232
233 4. Inside a character class, PCRE always treated \R and \X as literals,
234 whereas Perl faults them if its -w option is set. I have changed PCRE so
235 that it faults them when PCRE_EXTRA is set.
236
237 5. Added support for \N, which always matches any character other than
238 newline. (It is the same as "." when PCRE_DOTALL is not set.)
239
240 6. When compiling pcregrep with newer versions of gcc which may have
241 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
242 declared with attribute warn_unused_result" were given. Just casting the
243 result to (void) does not stop the warnings; a more elaborate fudge is
244 needed. I've used a macro to implement this.
245
246 7. Minor change to pcretest.c to avoid a compiler warning.
247
248 8. Added four artifical Unicode properties to help with an option to make
249 \s etc use properties (see next item). The new properties are: Xan
250 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
251
252 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
253 use Unicode properties. (*UCP) at the start of a pattern can be used to set
254 this option. Modified pcretest to add /W to test this facility. Added
255 REG_UCP to make it available via the POSIX interface.
256
257 10. Added --line-buffered to pcregrep.
258
259 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
260 studied, and the match started with a letter with a code point greater than
261 127 whose first byte was different to the first byte of the other case of
262 the letter, the other case of this starting letter was not recognized
263 (#976).
264
265 12. If a pattern that was studied started with a repeated Unicode property
266 test, for example, \p{Nd}+, there was the theoretical possibility of
267 setting up an incorrect bitmap of starting bytes, but fortunately it could
268 not have actually happened in practice until change 8 above was made (it
269 added property types that matched character-matching opcodes).
270
271 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
272 possible starting bytes for non-anchored patterns.
273
274 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
275 \R, and also a number of cases that involve Unicode properties, both
276 explicit and implicit when PCRE_UCP is set.
277
278 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
279 input, it could crash or give wrong results if characters with values
280 greater than 0xc0 were present in the subject string. (Detail: it assumed
281 UTF-8 input when processing these items.)
282
283 16. Added a lot of (int) casts to avoid compiler warnings in systems where
284 size_t is 64-bit (#991).
285
286 17. Added a check for running out of memory when PCRE is compiled with
287 --disable-stack-for-recursion (#990).
288
289 18. If the last data line in a file for pcretest does not have a newline on
290 the end, a newline was missing in the output.
291
292 19. The default pcre_chartables.c file recognizes only ASCII characters (values
293 less than 128) in its various bitmaps. However, there is a facility for
294 generating tables according to the current locale when PCRE is compiled. It
295 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
296 characters, are recognized by isspace() and therefore were getting set in
297 these tables, and indeed these tables seem to approximate to ISO 8859. This
298 caused a problem in UTF-8 mode when pcre_study() was used to create a list
299 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
300 which of course cannot start UTF-8 characters. I have changed the code so
301 that only real ASCII characters (less than 128) and the correct starting
302 bytes for UTF-8 encodings are set for characters greater than 127 when in
303 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
304 altogether.)
305
306 20. Added the /T option to pcretest so as to be able to run tests with non-
307 standard character tables, thus making it possible to include the tests
308 used for 19 above in the standard set of tests.
309
310 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
311 reference to a subpattern the other side of a comment that contains an
312 opening parenthesis caused either an internal compiling error, or a
313 reference to the wrong subpattern.
314
315
316 Version 8.02 19-Mar-2010
317 ------------------------
318
319 1. The Unicode data tables have been updated to Unicode 5.2.0.
320
321 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
322 configured.
323
324 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
325 original author of that file, following a query about its status.
326
327 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
328 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
329
330 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
331 quantifier applied to a forward-referencing subroutine call, could compile
332 incorrect code or give the error "internal error: previously-checked
333 referenced subpattern not found".
334
335 6. Both MS Visual Studio and Symbian OS have problems with initializing
336 variables to point to external functions. For these systems, therefore,
337 pcre_malloc etc. are now initialized to local functions that call the
338 relevant global functions.
339
340 7. There were two entries missing in the vectors called coptable and poptable
341 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
342 I've fixed the data, and added a kludgy way of testing at compile time that
343 the lengths are correct (equal to the number of opcodes).
344
345 8. Following on from 7, I added a similar kludge to check the length of the
346 eint vector in pcreposix.c.
347
348 9. Error texts for pcre_compile() are held as one long string to avoid too
349 much relocation at load time. To find a text, the string is searched,
350 counting zeros. There was no check for running off the end of the string,
351 which could happen if a new error number was added without updating the
352 string.
353
354 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
355
356 11. \K was not working if it appeared in an atomic group or in a group that
357 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
358 \K is "not well defined" if used in an assertion. PCRE now accepts it if
359 the assertion is positive, but not if it is negative.
360
361 12. Change 11 fortuitously reduced the size of the stack frame used in the
362 "match()" function of pcre_exec.c by one pointer. Forthcoming
363 implementation of support for (*MARK) will need an extra pointer on the
364 stack; I have reserved it now, so that the stack frame size does not
365 decrease.
366
367 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
368 item in branch that calls a recursion is a subroutine call - as in the
369 second branch in the above example - was incorrectly given the compile-
370 time error "recursive call could loop indefinitely" because pcre_compile()
371 was not correctly checking the subroutine for matching a non-empty string.
372
373 14. The checks for overrunning compiling workspace could trigger after an
374 overrun had occurred. This is a "should never occur" error, but it can be
375 triggered by pathological patterns such as hundreds of nested parentheses.
376 The checks now trigger 100 bytes before the end of the workspace.
377
378 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
379
380
381 Version 8.01 19-Jan-2010
382 ------------------------
383
384 1. If a pattern contained a conditional subpattern with only one branch (in
385 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
386 computed the wrong minimum data length (which is of course zero for such
387 subpatterns). This could cause incorrect "no match" results.
388
389 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
390 the pattern is reset in the first branch, pcre_compile() failed with
391 "internal error: code overflow at offset...". This happened only when
392 the reset was to the original external option setting. (An optimization
393 abstracts leading options settings into an external setting, which was the
394 cause of this.)
395
396 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
397 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
398 assertion pattern did not match (meaning that the assertion was true), it
399 was incorrectly treated as false if the SKIP had been reached during the
400 matching. This also applied to assertions used as conditions.
401
402 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
403 assertion subpattern, including such a pattern used as a condition,
404 unpredictable results occurred, instead of the error return
405 PCRE_ERROR_DFA_UITEM.
406
407 5. The C++ GlobalReplace function was not working like Perl for the special
408 situation when an empty string is matched. It now does the fancy magic
409 stuff that is necessary.
410
411 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
412 removed. (These were left over from very, very early versions of PCRE.)
413
414 7. Some cosmetic changes to the code to make life easier when compiling it
415 as part of something else:
416
417 (a) Change DEBUG to PCRE_DEBUG.
418
419 (b) In pcre_compile(), rename the member of the "branch_chain" structure
420 called "current" as "current_branch", to prevent a collision with the
421 Linux macro when compiled as a kernel module.
422
423 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
424 prevent a collision with the Linux macro when compiled as a kernel
425 module.
426
427 8. In pcre_compile() there are some checks for integer overflows that used to
428 cast potentially large values to (double). This has been changed to that
429 when building, a check for int64_t is made, and if it is found, it is used
430 instead, thus avoiding the use of floating point arithmetic. (There is no
431 other use of FP in PCRE.) If int64_t is not found, the fallback is to
432 double.
433
434 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
435 2005 (difference between two addresses compared to an unsigned value).
436
437 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
438 custom one, because of the following reported problem in Windows:
439
440 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
441 under Win32.
442 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
443 therefore missing the function definition.
444 - The compiler thus generates a "C" signature for the test function.
445 - The linker fails to find the "C" function.
446 - PCRE fails to configure if asked to do so against libbz2.
447
448 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
449 messages were output:
450
451 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
452 rerunning libtoolize, to keep the correct libtool macros in-tree.
453 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
454
455 I have done both of these things.
456
457 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
458 most of the time, it *can* run out if it is given a pattern that contains a
459 runaway infinite recursion. I updated the discussion in the pcrestack man
460 page.
461
462 13. Now that we have gone to the x.xx style of version numbers, the minor
463 version may start with zero. Using 08 or 09 is a bad idea because users
464 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
465 interpreted as invalid octal numbers. I've updated the previous comment in
466 configure.ac, and also added a check that gives an error if 08 or 09 are
467 used.
468
469 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
470 causing partial matching to fail when the end of the subject matched \W
471 in a UTF-8 pattern where \W was quantified with a minimum of 3.
472
473 15. There were some discrepancies between the declarations in pcre_internal.h
474 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
475 their definitions. The declarations used "const uschar *" and the
476 definitions used USPTR. Even though USPTR is normally defined as "const
477 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
478 reported that: "This difference in casting confuses some C++ compilers, for
479 example, SunCC recognizes above declarations as different functions and
480 generates broken code for hbpcre." I have changed the declarations to use
481 USPTR.
482
483 16. GNU libtool is named differently on some systems. The autogen.sh script now
484 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
485 (FreeBSD).
486
487 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
488 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
489 comment: "Figure out how to create a longlong from a string: strtoll and
490 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
491 instance, but it only takes 2 args instead of 3!"
492
493 18. A subtle bug concerned with back references has been fixed by a change of
494 specification, with a corresponding code fix. A pattern such as
495 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
496 refers, was giving matches when it shouldn't. For example, xa=xaaa would
497 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
498 same bug. Such groups have to be quantified to be useful, or contained
499 inside another quantified group. (If there's no repetition, the reference
500 can never match.) The problem arises because, having left the group and
501 moved on to the rest of the pattern, a later failure that backtracks into
502 the group uses the captured value from the final iteration of the group
503 rather than the correct earlier one. I have fixed this in PCRE by forcing
504 any group that contains a reference to itself to be an atomic group; that
505 is, there cannot be any backtracking into it once it has completed. This is
506 similar to recursive and subroutine calls.
507
508
509 Version 8.00 19-Oct-09
510 ----------------------
511
512 1. The table for translating pcre_compile() error codes into POSIX error codes
513 was out-of-date, and there was no check on the pcre_compile() error code
514 being within the table. This could lead to an OK return being given in
515 error.
516
517 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
518 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
519 in a Windows environment.
520
521 3. The pcregrep --count option prints the count for each file even when it is
522 zero, as does GNU grep. However, pcregrep was also printing all files when
523 --files-with-matches was added. Now, when both options are given, it prints
524 counts only for those files that have at least one match. (GNU grep just
525 prints the file name in this circumstance, but including the count seems
526 more useful - otherwise, why use --count?) Also ensured that the
527 combination -clh just lists non-zero counts, with no names.
528
529 4. The long form of the pcregrep -F option was incorrectly implemented as
530 --fixed_strings instead of --fixed-strings. This is an incompatible change,
531 but it seems right to fix it, and I didn't think it was worth preserving
532 the old behaviour.
533
534 5. The command line items --regex=pattern and --regexp=pattern were not
535 recognized by pcregrep, which required --regex pattern or --regexp pattern
536 (with a space rather than an '='). The man page documented the '=' forms,
537 which are compatible with GNU grep; these now work.
538
539 6. No libpcreposix.pc file was created for pkg-config; there was just
540 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
541
542 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
543 when UCP support is not needed, by modifying the Python script that
544 generates it from Unicode data files. This should not matter if the module
545 is correctly used as a library, but I received one complaint about 50K of
546 unwanted data. My guess is that the person linked everything into his
547 program rather than using a library. Anyway, it does no harm.
548
549 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
550 was a minimum greater than 1 for a wide character in a possessive
551 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
552 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
553 character. Chaos in the form of incorrect output or a compiling loop could
554 result.
555
556 9. The restrictions on what a pattern can contain when partial matching is
557 requested for pcre_exec() have been removed. All patterns can now be
558 partially matched by this function. In addition, if there are at least two
559 slots in the offset vector, the offset of the earliest inspected character
560 for the match and the offset of the end of the subject are set in them when
561 PCRE_ERROR_PARTIAL is returned.
562
563 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
564 synonymous with PCRE_PARTIAL, for backwards compatibility, and
565 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
566 and may be more useful for multi-segment matching.
567
568 11. Partial matching with pcre_exec() is now more intuitive. A partial match
569 used to be given if ever the end of the subject was reached; now it is
570 given only if matching could not proceed because another character was
571 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
572 string "Z", which now yields "no match" instead of "partial match". In the
573 case of pcre_dfa_exec(), "no match" is given if every matching path for the
574 final character ended with (*FAIL).
575
576 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
577 if the pattern had a "must contain" character that was already found in the
578 earlier partial match, unless partial matching was again requested. For
579 example, with the pattern /dog.(body)?/, the "must contain" character is
580 "g". If the first part-match was for the string "dog", restarting with
581 "sbody" failed. This bug has been fixed.
582
583 13. The string returned by pcre_dfa_exec() after a partial match has been
584 changed so that it starts at the first inspected character rather than the
585 first character of the match. This makes a difference only if the pattern
586 starts with a lookbehind assertion or \b or \B (\K is not supported by
587 pcre_dfa_exec()). It's an incompatible change, but it makes the two
588 matching functions compatible, and I think it's the right thing to do.
589
590 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
591 so that the demonstration program is easily available in environments where
592 PCRE has not been installed from source.
593
594 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
595 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
596 library.
597
598 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
599 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
600 is not the first non-POSIX option to be added. Clearly some people find
601 these options useful.
602
603 17. If a caller to the POSIX matching function regexec() passes a non-zero
604 value for nmatch with a NULL value for pmatch, the value of
605 nmatch is forced to zero.
606
607 18. RunGrepTest did not have a test for the availability of the -u option of
608 the diff command, as RunTest does. It now checks in the same way as
609 RunTest, and also checks for the -b option.
610
611 19. If an odd number of negated classes containing just a single character
612 interposed, within parentheses, between a forward reference to a named
613 subpattern and the definition of the subpattern, compilation crashed with
614 an internal error, complaining that it could not find the referenced
615 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
616 [The bug was that it was starting one character too far in when skipping
617 over the character class, thus treating the ] as data rather than
618 terminating the class. This meant it could skip too much.]
619
620 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
621 /g option in pcretest when the pattern contains \K, which makes it possible
622 to have an empty string match not at the start, even when the pattern is
623 anchored. Updated pcretest and pcredemo to use this option.
624
625 21. If the maximum number of capturing subpatterns in a recursion was greater
626 than the maximum at the outer level, the higher number was returned, but
627 with unset values at the outer level. The correct (outer level) value is
628 now given.
629
630 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
631 PCRE did not set those parentheses (unlike Perl). I have now found a way to
632 make it do so. The string so far is captured, making this feature
633 compatible with Perl.
634
635 23. The tests have been re-organized, adding tests 11 and 12, to make it
636 possible to check the Perl 5.10 features against Perl 5.10.
637
638 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
639 pattern matches a fixed length string. PCRE did not allow this; now it
640 does. Neither allows recursion.
641
642 25. I finally figured out how to implement a request to provide the minimum
643 length of subject string that was needed in order to match a given pattern.
644 (It was back references and recursion that I had previously got hung up
645 on.) This code has now been added to pcre_study(); it finds a lower bound
646 to the length of subject needed. It is not necessarily the greatest lower
647 bound, but using it to avoid searching strings that are too short does give
648 some useful speed-ups. The value is available to calling programs via
649 pcre_fullinfo().
650
651 26. While implementing 25, I discovered to my embarrassment that pcretest had
652 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
653 study optimizations had never been tested with that matching function.
654 Oops. What is worse, even when it was passed study data, there was a bug in
655 pcre_dfa_exec() that meant it never actually used it. Double oops. There
656 were also very few tests of studied patterns with pcre_dfa_exec().
657
658 27. If (?| is used to create subpatterns with duplicate numbers, they are now
659 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
660 on the other side of the coin, they are no longer allowed to have different
661 names, because these cannot be distinguished in PCRE, and this has caused
662 confusion. (This is a difference from Perl.)
663
664 28. When duplicate subpattern names are present (necessarily with different
665 numbers, as required by 27 above), and a test is made by name in a
666 conditional pattern, either for a subpattern having been matched, or for
667 recursion in such a pattern, all the associated numbered subpatterns are
668 tested, and the overall condition is true if the condition is true for any
669 one of them. This is the way Perl works, and is also more like the way
670 testing by number works.
671
672
673 Version 7.9 11-Apr-09
674 ---------------------
675
676 1. When building with support for bzlib/zlib (pcregrep) and/or readline
677 (pcretest), all targets were linked against these libraries. This included
678 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
679 libraries. This caused unwanted dependencies to be created. This problem
680 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
681 pcretest is linked with readline.
682
683 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
684 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
685 moved outside it again, because FALSE and TRUE are already defined in AIX,
686 but BOOL is not.
687
688 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
689 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
690
691 4. The pcregrep documentation said spaces were inserted as well as colons (or
692 hyphens) following file names and line numbers when outputting matching
693 lines. This is not true; no spaces are inserted. I have also clarified the
694 wording for the --colour (or --color) option.
695
696 5. In pcregrep, when --colour was used with -o, the list of matching strings
697 was not coloured; this is different to GNU grep, so I have changed it to be
698 the same.
699
700 6. When --colo(u)r was used in pcregrep, only the first matching substring in
701 each matching line was coloured. Now it goes on to look for further matches
702 of any of the test patterns, which is the same behaviour as GNU grep.
703
704 7. A pattern that could match an empty string could cause pcregrep to loop; it
705 doesn't make sense to accept an empty string match in pcregrep, so I have
706 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
707 seems to be how GNU grep behaves.
708
709 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
710 start or after a newline", because the conditional assertion was not being
711 correctly handled. The rule now is that both the assertion and what follows
712 in the first alternative must satisfy the test.
713
714 9. If auto-callout was enabled in a pattern with a conditional group whose
715 condition was an assertion, PCRE could crash during matching, both with
716 pcre_exec() and pcre_dfa_exec().
717
718 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
719 used for matching.
720
721 11. Unicode property support in character classes was not working for
722 characters (bytes) greater than 127 when not in UTF-8 mode.
723
724 12. Added the -M command line option to pcretest.
725
726 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
727
728 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
729
730 16. Added comments and documentation about mis-use of no_arg in the C++
731 wrapper.
732
733 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
734 from Martin Jerabek that uses macro names for all relevant character and
735 string constants.
736
737 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
738 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
739 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
740 these, but not everybody uses configure.
741
742 19. A conditional group that had only one branch was not being correctly
743 recognized as an item that could match an empty string. This meant that an
744 enclosing group might also not be so recognized, causing infinite looping
745 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
746 with the subject "ab", where knowledge that the repeated group can match
747 nothing is needed in order to break the loop.
748
749 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
750 exec(), but without supplying a callout function, matching went wrong.
751
752 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
753 leak if the size of the offset vector was greater than 30. When the vector
754 is smaller, the saved offsets during recursion go onto a local stack
755 vector, but for larger vectors malloc() is used. It was failing to free
756 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
757 error, in fact).
758
759 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
760 heapframe that is used only when UTF-8 support is enabled. This caused no
761 problem, but was untidy.
762
763 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
764 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
765 included within another project.
766
767 24. Steven Van Ingelgem's patches to add more options to the CMake support,
768 slightly modified by me:
769
770 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
771 not building pcregrep.
772
773 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
774 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
775
776 25. Forward references, both numeric and by name, in patterns that made use of
777 duplicate group numbers, could behave incorrectly or give incorrect errors,
778 because when scanning forward to find the reference group, PCRE was not
779 taking into account the duplicate group numbers. A pattern such as
780 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
781
782 26. Changed a few more instances of "const unsigned char *" to USPTR, making
783 the feature of a custom pointer more persuasive (as requested by a user).
784
785 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
786 pcretest.c, inside #ifndefs, because it seems they are sometimes already
787 pre-defined.
788
789 28. Added support for (*UTF8) at the start of a pattern.
790
791 29. Arrange for flags added by the "release type" setting in CMake to be shown
792 in the configuration summary.
793
794
795 Version 7.8 05-Sep-08
796 ---------------------
797
798 1. Replaced UCP searching code with optimized version as implemented for Ad
799 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
800 stage table and inline lookup instead of a function, giving speed ups of 2
801 to 5 times on some simple patterns that I tested. Permission was given to
802 distribute the MultiStage2.py script that generates the tables (it's not in
803 the tarball, but is in the Subversion repository).
804
805 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
806 scripts.
807
808 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
809 a group with a zero qualifier. The result of the study could be incorrect,
810 or the function might crash, depending on the pattern.
811
812 4. Caseless matching was not working for non-ASCII characters in back
813 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
814 It now works when Unicode Property Support is available.
815
816 5. In pcretest, an escape such as \x{de} in the data was always generating
817 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
818 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
819 truncation.
820
821 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
822
823 7. Added two (int) casts to pcregrep when printing the difference of two
824 pointers, in case they are 64-bit values.
825
826 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
827 test 2 if it fails.
828
829 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
830 and a #define of that name to empty if it is not externally set. This is to
831 allow users of MSVC to set it if necessary.
832
833 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
834 the convenience functions in the pcre_get.c source file.
835
836 11. An option change at the start of a pattern that had top-level alternatives
837 could cause overwriting and/or a crash. This command provoked a crash in
838 some environments:
839
840 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
841
842 This potential security problem was recorded as CVE-2008-2371.
843
844 12. For a pattern where the match had to start at the beginning or immediately
845 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
846 pcre_dfa_exec() could read past the end of the passed subject if there was
847 no match. To help with detecting such bugs (e.g. with valgrind), I modified
848 pcretest so that it places the subject at the end of its malloc-ed buffer.
849
850 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
851 exec() might read past the end of the data buffer in UTF-8 mode.
852
853 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
854 the data contained the byte 0x85 as part of a UTF-8 character within its
855 first line. This applied both to normal and DFA matching.
856
857 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
858 /^[^d]*?$/8 failed to match "abc".
859
860 16. Added a missing copyright notice to pcrecpp_internal.h.
861
862 17. Make it more clear in the documentation that values returned from
863 pcre_exec() in ovector are byte offsets, not character counts.
864
865 18. Tidied a few places to stop certain compilers from issuing warnings.
866
867 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
868 supplied by Stefan Weber. I made a further small update for 7.8 because
869 there is a change of source arrangements: the pcre_searchfuncs.c module is
870 replaced by pcre_ucd.c.
871
872
873 Version 7.7 07-May-08
874 ---------------------
875
876 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
877 a string to a long long, pretend we don't even have a long long." This is
878 done by checking for the strtoq, strtoll, and _strtoi64 functions.
879
880 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
881 pre-7.6 versions, which defined a global no_arg variable instead of putting
882 it in the RE class. (See also #8 below.)
883
884 3. Remove a line of dead code, identified by coverity and reported by Nuno
885 Lopes.
886
887 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
888
889 (1) The include/exclude patterns were being applied to the whole pathnames
890 of files, instead of just to the final components.
891
892 (2) If there was more than one level of directory, the subdirectories were
893 skipped unless they satisfied the include/exclude conditions. This is
894 inconsistent with GNU grep (and could even be seen as contrary to the
895 pcregrep specification - which I improved to make it absolutely clear).
896 The action now is always to scan all levels of directory, and just
897 apply the include/exclude patterns to regular files.
898
899 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
900 --exclude_dir in the tests to avoid scanning .svn directories.
901
902 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
903 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
904 doesn't support NULs in patterns.
905
906 7. Added some missing "const"s to declarations of static tables in
907 pcre_compile.c and pcre_dfa_exec.c.
908
909 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
910 caused by fix #2 above. (Subsequently also a second patch to fix the
911 first patch. And a third patch - this was a messy problem.)
912
913 9. Applied Craig's patch to remove the use of push_back().
914
915 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
916 matching function regexec().
917
918 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
919 which, however, unlike Perl's \g{...}, are subroutine calls, not back
920 references. PCRE supports relative numbers with this syntax (I don't think
921 Oniguruma does).
922
923 12. Previously, a group with a zero repeat such as (...){0} was completely
924 omitted from the compiled regex. However, this means that if the group
925 was called as a subroutine from elsewhere in the pattern, things went wrong
926 (an internal error was given). Such groups are now left in the compiled
927 pattern, with a new opcode that causes them to be skipped at execution
928 time.
929
930 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
931 to the way PCRE behaves:
932
933 (a) A lone ] character is dis-allowed (Perl treats it as data).
934
935 (b) A back reference to an unmatched subpattern matches an empty string
936 (Perl fails the current match path).
937
938 (c) A data ] in a character class must be notated as \] because if the
939 first data character in a class is ], it defines an empty class. (In
940 Perl it is not possible to have an empty class.) The empty class []
941 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
942 The negative empty class [^] matches any one character, independently
943 of the DOTALL setting.
944
945 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
946 non-existent subpattern following a character class starting with ']' and
947 containing () gave an internal compiling error instead of "reference to
948 non-existent subpattern". Fortunately, when the pattern did exist, the
949 compiled code was correct. (When scanning forwards to check for the
950 existencd of the subpattern, it was treating the data ']' as terminating
951 the class, so got the count wrong. When actually compiling, the reference
952 was subsequently set up correctly.)
953
954 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
955 it was being rejected as not supported by pcre_dfa_exec(), even though
956 other assertions are supported. I have made pcre_dfa_exec() support
957 (*FAIL).
958
959 16. The implementation of 13c above involved the invention of a new opcode,
960 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
961 cannot be changed at match time, I realized I could make a small
962 improvement to matching performance by compiling OP_ALLANY instead of
963 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
964 on the OP_ANY path.
965
966 17. Compiling pcretest on Windows with readline support failed without the
967 following two fixes: (1) Make the unistd.h include conditional on
968 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
969
970 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
971 ncurses library to be included for pcretest when ReadLine support is
972 requested, but also to allow for it to be overridden. This patch came from
973 Daniel Bergström.
974
975 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
976 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
977 any errors with the current Unicode tables. Thanks to Peter Kankowski for
978 spotting this.
979
980
981 Version 7.6 28-Jan-08
982 ---------------------
983
984 1. A character class containing a very large number of characters with
985 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
986 overflow.
987
988 2. Patch to cut out the "long long" test in pcrecpp_unittest when
989 HAVE_LONG_LONG is not defined.
990
991 3. Applied Christian Ehrlicher's patch to update the CMake build files to
992 bring them up to date and include new features. This patch includes:
993
994 - Fixed PH's badly added libz and libbz2 support.
995 - Fixed a problem with static linking.
996 - Added pcredemo. [But later removed - see 7 below.]
997 - Fixed dftables problem and added an option.
998 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
999 HAVE_LONG_LONG.
1000 - Added readline support for pcretest.
1001 - Added an listing of the option settings after cmake has run.
1002
1003 4. A user submitted a patch to Makefile that makes it easy to create
1004 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1005 Makefile.am that cause it to include this special target, without
1006 affecting anything else. Note that the same mingw target plus all
1007 the other distribution libraries and programs are now supported
1008 when configuring with CMake (see 6 below) instead of with
1009 Configure/Make.
1010
1011 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1012 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1013 exported in the Windows port". It has not yet been confirmed that the patch
1014 solves the problem, but it does no harm.
1015
1016 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1017 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1018 with CMake, and also correct the comment about stack recursion.
1019
1020 7. Remove the automatic building of pcredemo from the ./configure system and
1021 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1022 of a program that users should build themselves after PCRE is installed, so
1023 building it automatically is not really right. What is more, it gave
1024 trouble in some build environments.
1025
1026 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1027
1028
1029 Version 7.5 10-Jan-08
1030 ---------------------
1031
1032 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1033 values in parens when parsing an RE using the C++ wrapper."
1034
1035 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1036 Characters greater than 255 were excluded from the class instead of being
1037 included.
1038
1039 3. The same bug as (2) above applied to negated POSIX classes such as
1040 [:^space:].
1041
1042 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1043 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1044 I have changed it.
1045
1046 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1047 first named subpattern) and a construct such as (?&a) would reference the
1048 first named subpattern whose name started with "a" (in other words, the
1049 length check was missing). Both these problems are fixed. "Subpattern name
1050 expected" is now given for (?&) (a zero-length name), and this patch also
1051 makes it give the same error for \k'' (previously it complained that that
1052 was a reference to a non-existent subpattern).
1053
1054 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1055 this is right because (?- can be followed by option settings as well as by
1056 digits. I have, however, made the messages clearer.
1057
1058 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1059 than the number used in the conditional) now cause a compile-time error.
1060 This is actually not compatible with Perl, which accepts such patterns, but
1061 treats the conditional as always being FALSE (as PCRE used to), but it
1062 seems to me that giving a diagnostic is better.
1063
1064 8. Change "alphameric" to the more common word "alphanumeric" in comments
1065 and messages.
1066
1067 9. Fix two occurrences of "backslash" in comments that should have been
1068 "backspace".
1069
1070 10. Remove two redundant lines of code that can never be obeyed (their function
1071 was moved elsewhere).
1072
1073 11. The program that makes PCRE's Unicode character property table had a bug
1074 which caused it to generate incorrect table entries for sequences of
1075 characters that have the same character type, but are in different scripts.
1076 It amalgamated them into a single range, with the script of the first of
1077 them. In other words, some characters were in the wrong script. There were
1078 thirteen such cases, affecting characters in the following ranges:
1079
1080 U+002b0 - U+002c1
1081 U+0060c - U+0060d
1082 U+0061e - U+00612
1083 U+0064b - U+0065e
1084 U+0074d - U+0076d
1085 U+01800 - U+01805
1086 U+01d00 - U+01d77
1087 U+01d9b - U+01dbf
1088 U+0200b - U+0200f
1089 U+030fc - U+030fe
1090 U+03260 - U+0327f
1091 U+0fb46 - U+0fbb1
1092 U+10450 - U+1049d
1093
1094 12. The -o option (show only the matching part of a line) for pcregrep was not
1095 compatible with GNU grep in that, if there was more than one match in a
1096 line, it showed only the first of them. It now behaves in the same way as
1097 GNU grep.
1098
1099 13. If the -o and -v options were combined for pcregrep, it printed a blank
1100 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1101 does the same. The return code can be used to tell if there were any
1102 non-matching lines.
1103
1104 14. Added --file-offsets and --line-offsets to pcregrep.
1105
1106 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1107 infinitely looping recursion. The bug was that positive lookaheads were not
1108 being skipped when checking for a possible empty match (negative lookaheads
1109 and both kinds of lookbehind were skipped).
1110
1111 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1112 inclusion of <windows.h> to before rather than after the definition of
1113 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1114
1115 17. Specifying a possessive quantifier with a specific limit for a Unicode
1116 character property caused pcre_compile() to compile bad code, which led at
1117 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1118 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1119 caused the error; without that there was no problem.
1120
1121 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1122
1123 19. Added --enable-pcretest-libreadline.
1124
1125 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1126 RE::GlobalReplace(). As a result, the number of replacements returned was
1127 double what it should be. I removed one of the increments, but Craig sent a
1128 later patch that removed the other one (the right fix) and added unit tests
1129 that check the return values (which was not done before).
1130
1131 21. Several CMake things:
1132
1133 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1134 the names libpcre and libpcreposix, not just pcre and pcreposix.
1135
1136 (2) The above change means that pcretest and pcregrep are now correctly
1137 linked with the newly-built libraries, not previously installed ones.
1138
1139 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1140
1141 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1142 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1143 UTF-8 newline character). The key issue is that the pattern starts .*;
1144 this means that the match must be either at the beginning, or after a
1145 newline. The bug was in the code for advancing after a failed match and
1146 checking that the new position followed a newline. It was not taking
1147 account of UTF-8 characters correctly.
1148
1149 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1150 character classes. PCRE was not treating the sequence [:...:] as a
1151 character class unless the ... were all letters. Perl, however, seems to
1152 allow any characters between [: and :], though of course it rejects as
1153 unknown any "names" that contain non-letters, because all the known class
1154 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1155 for example, whereas PCRE did not - it did not recognize a POSIX character
1156 class. This seemed a bit dangerous, so the code has been changed to be
1157 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1158 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1159 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1160 Perl does, and where it didn't before.
1161
1162 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1163 Windows environments %n is disabled by default.
1164
1165
1166 Version 7.4 21-Sep-07
1167 ---------------------
1168
1169 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1170 means that a class such as [\s] counted as "explicit reference to CR or
1171 LF". That isn't really right - the whole point of the change was to try to
1172 help when there was an actual mention of one of the two characters. So now
1173 the change happens only if \r or \n (or a literal CR or LF) character is
1174 encountered.
1175
1176 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1177 of both had grown to the point where there were only 3 bits left.
1178 Fortunately, there was spare space in the data structure, and so I have
1179 moved the internal flags into a new 16-bit field to free up more option
1180 bits.
1181
1182 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1183 but did not set the internal JCHANGED flag - either of these is enough to
1184 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1185 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1186 start sets both bits.
1187
1188 4. Added options (at build time, compile time, exec time) to change \R from
1189 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1190
1191 5. doc/pcresyntax.html was missing from the distribution.
1192
1193 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1194 compatibility, even though it is no longer used.
1195
1196 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1197 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1198 windows.h file is present (where different names are used). [This was
1199 reversed later after testing - see 16 below.]
1200
1201 8. Changed all #include <config.h> to #include "config.h". There were also
1202 some further <pcre.h> cases that I changed to "pcre.h".
1203
1204 9. When pcregrep was used with the --colour option, it missed the line ending
1205 sequence off the lines that it output.
1206
1207 10. It was pointed out to me that arrays of string pointers cause lots of
1208 relocations when a shared library is dynamically loaded. A technique of
1209 using a single long string with a table of offsets can drastically reduce
1210 these. I have refactored PCRE in four places to do this. The result is
1211 dramatic:
1212
1213 Originally: 290
1214 After changing UCP table: 187
1215 After changing error message table: 43
1216 After changing table of "verbs" 36
1217 After changing table of Posix names 22
1218
1219 Thanks to the folks working on Gregex for glib for this insight.
1220
1221 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1222 unicode-properties was also set.
1223
1224 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1225
1226 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1227 checked only for CRLF.
1228
1229 14. Added casts to pcretest.c to avoid compiler warnings.
1230
1231 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1232
1233 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1234 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1235 entirely. This removes changes made in 7 above.
1236
1237 17. The CMake files have been updated, and there is now more information about
1238 building with CMake in the NON-UNIX-USE document.
1239
1240
1241 Version 7.3 28-Aug-07
1242 ---------------------
1243
1244 1. In the rejigging of the build system that eventually resulted in 7.1, the
1245 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1246 brackets there is not right, since it causes compilers to look for an
1247 installed pcre.h, not the version that is in the source that is being
1248 compiled (which of course may be different). I have changed it back to:
1249
1250 #include "pcre.h"
1251
1252 I have a vague recollection that the change was concerned with compiling in
1253 different directories, but in the new build system, that is taken care of
1254 by the VPATH setting the Makefile.
1255
1256 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1257 when the subject happened to end in the byte 0x85 (e.g. if the last
1258 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1259 characters but of course it shouldn't be taken as a newline when it is part
1260 of another character. The bug was that, for an unlimited repeat of . in
1261 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1262 characters when looking for a newline.
1263
1264 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1265
1266 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1267 in debug output.
1268
1269 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1270 long printing in the pcrecpp unittest when running under MinGW.
1271
1272 6. ESC_K was left out of the EBCDIC table.
1273
1274 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1275 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1276 limit also applies to "virtual nesting" when a pattern is recursive, and in
1277 this case 1000 isn't so big. I have been able to remove this limit at the
1278 expense of backing off one optimization in certain circumstances. Normally,
1279 when pcre_exec() would call its internal match() function recursively and
1280 immediately return the result unconditionally, it uses a "tail recursion"
1281 feature to save stack. However, when a subpattern that can match an empty
1282 string has an unlimited repetition quantifier, it no longer makes this
1283 optimization. That gives it a stack frame in which to save the data for
1284 checking that an empty string has been matched. Previously this was taken
1285 from the 1000-entry workspace that had been reserved. So now there is no
1286 explicit limit, but more stack is used.
1287
1288 8. Applied Daniel's patches to solve problems with the import/export magic
1289 syntax that is required for Windows, and which was going wrong for the
1290 pcreposix and pcrecpp parts of the library. These were overlooked when this
1291 problem was solved for the main library.
1292
1293 9. There were some crude static tests to avoid integer overflow when computing
1294 the size of patterns that contain repeated groups with explicit upper
1295 limits. As the maximum quantifier is 65535, the maximum group length was
1296 set at 30,000 so that the product of these two numbers did not overflow a
1297 32-bit integer. However, it turns out that people want to use groups that
1298 are longer than 30,000 bytes (though not repeat them that many times).
1299 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1300 made it possible to implement the integer overflow checks in a much more
1301 dynamic way, which I have now done. The artificial limitation on group
1302 length has been removed - we now have only the limit on the total length of
1303 the compiled pattern, which depends on the LINK_SIZE setting.
1304
1305 10. Fixed a bug in the documentation for get/copy named substring when
1306 duplicate names are permitted. If none of the named substrings are set, the
1307 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1308 empty string.
1309
1310 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1311 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1312 because the ] is interpreted as the first data character and the
1313 terminating ] is not found. PCRE has been made compatible with Perl in this
1314 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1315 cause memory overwriting.
1316
1317 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1318 string has been matched (to stop an infinite loop). It was not recognizing
1319 a conditional subpattern that could match an empty string if that
1320 subpattern was within another subpattern. For example, it looped when
1321 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1322 condition was not nested. This bug has been fixed.
1323
1324 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1325 past the start of the subject in the presence of bytes with the top bit
1326 set, for example "\x8aBCD".
1327
1328 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1329 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1330
1331 14. Optimized (?!) to (*FAIL).
1332
1333 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1334 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1335 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1336 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1337 does: it's just the validity check that is more restrictive.
1338
1339 16. Inserted checks for integer overflows during escape sequence (backslash)
1340 processing, and also fixed erroneous offset values for syntax errors during
1341 backslash processing.
1342
1343 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1344 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1345
1346 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1347 caused an overrun.
1348
1349 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1350 something other than just ASCII characters) inside a group that had an
1351 unlimited repeat caused a loop at compile time (while checking to see
1352 whether the group could match an empty string).
1353
1354 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1355 [\P{Any}] did so. (Error in the code for printing property names.)
1356
1357 21. An orphan \E inside a character class could cause a crash.
1358
1359 22. A repeated capturing bracket such as (A)? could cause a wild memory
1360 reference during compilation.
1361
1362 23. There are several functions in pcre_compile() that scan along a compiled
1363 expression for various reasons (e.g. to see if it's fixed length for look
1364 behind). There were bugs in these functions when a repeated \p or \P was
1365 present in the pattern. These operators have additional parameters compared
1366 with \d, etc, and these were not being taken into account when moving along
1367 the compiled data. Specifically:
1368
1369 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1370 length.
1371
1372 (b) An item such as \pL+ within a repeated group could cause crashes or
1373 loops.
1374
1375 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1376 "reference to non-existent subpattern" error.
1377
1378 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1379
1380 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1381 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1382
1383 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1384
1385 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1386 character were causing crashes (broken optimization).
1387
1388 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1389 \p or \P) caused a compile-time loop.
1390
1391 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1392 break. For example, the unstudied pattern [\r\n]A does not match the string
1393 "\r\nA" because change 7.0/46 below moves the current point on by two
1394 characters after failing to match at the start. However, the pattern \nA
1395 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1396 the same is true. There doesn't seem any very clean way out of this, but
1397 what I have chosen to do makes the common cases work: PCRE now takes note
1398 of whether there can be an explicit match for \r or \n anywhere in the
1399 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1400 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1401 pattern has explicit CR or LF references.
1402
1403 29. Added (*CR) etc for changing newline setting at start of pattern.
1404
1405
1406 Version 7.2 19-Jun-07
1407 ---------------------
1408
1409 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1410 which is apparently normally available under Windows.
1411
1412 2. Re-jig the pcregrep tests with different newline settings in an attempt
1413 to make them independent of the local environment's newline setting.
1414
1415 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1416
1417 4. Some of the "internals" tests were previously cut out when the link size
1418 was not 2, because the output contained actual offsets. The recent new
1419 "Z" feature of pcretest means that these can be cut out, making the tests
1420 usable with all link sizes.
1421
1422 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1423 stack recursion. This gives a massive performance boost under BSD, but just
1424 a small improvement under Linux. However, it saves one field in the frame
1425 in all cases.
1426
1427 6. Added more features from the forthcoming Perl 5.10:
1428
1429 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1430 recursion call. It refers to the nth most recently opened parentheses.
1431
1432 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1433 to be opened parentheses.
1434
1435 (c) Conditions that refer to capturing parentheses can be specified
1436 relatively, for example, (?(-2)... or (?(+3)...
1437
1438 (d) \K resets the start of the current match so that everything before
1439 is not part of it.
1440
1441 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1442
1443 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1444 reference syntax.
1445
1446 (g) (?| introduces a group in which the numbering of parentheses in each
1447 alternative starts with the same number.
1448
1449 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1450
1451 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1452 PCRE_INFO_JCHANGED.
1453
1454 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1455 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1456 for detecting groups that can match an empty string.
1457
1458 9. A pattern with a very large number of alternatives (more than several
1459 hundred) was running out of internal workspace during the pre-compile
1460 phase, where pcre_compile() figures out how much memory will be needed. A
1461 bit of new cunning has reduced the workspace needed for groups with
1462 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1463 workspace instead of running out of the 4096 that are available.
1464
1465 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1466
1467 11. Applied patch from Google to remove an optimization that didn't quite work.
1468 The report of the bug said:
1469
1470 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1471 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1472 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1473
1474 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1475 it matched the wrong number of bytes.
1476
1477
1478 Version 7.1 24-Apr-07
1479 ---------------------
1480
1481 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1482 that is more "standard", making use of automake and other Autotools. There
1483 is some re-arrangement of the files and adjustment of comments consequent
1484 on this.
1485
1486 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1487 for recursive directory scanning broke on some systems because the files
1488 are not scanned in any specific order and on different systems the order
1489 was different. A call to "sort" has been inserted into RunGrepTest for the
1490 approprate test as a short-term fix. In the longer term there may be an
1491 alternative.
1492
1493 3. I had an email from Eric Raymond about problems translating some of PCRE's
1494 man pages to HTML (despite the fact that I distribute HTML pages, some
1495 people do their own conversions for various reasons). The problems
1496 concerned the use of low-level troff macros .br and .in. I have therefore
1497 removed all such uses from the man pages (some were redundant, some could
1498 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1499 HTML has been updated to handle .nf/.fi and to complain if it encounters
1500 .br or .in.
1501
1502 4. Updated comments in configure.ac that get placed in config.h.in and also
1503 arranged for config.h to be included in the distribution, with the name
1504 config.h.generic, for the benefit of those who have to compile without
1505 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1506
1507 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1508 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1509 makevp.bat for the current PCRE, using the additional files
1510 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1511
1512 6. A Windows user reported a minor discrepancy with test 2, which turned out
1513 to be caused by a trailing space on an input line that had got lost in his
1514 copy. The trailing space was an accident, so I've just removed it.
1515
1516 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1517 that is needed.
1518
1519 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1520 as "const" (a) because they are and (b) because it helps the PHP
1521 maintainers who have recently made a script to detect big data structures
1522 in the php code that should be moved to the .rodata section. I remembered
1523 to update Builducptable as well, so it won't revert if ucptable.h is ever
1524 re-created.
1525
1526 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1527 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1528 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1529 support is not required. This saves 1.5-2K of code, which is important in
1530 some applications.
1531
1532 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1533 so as not to refer to the tables, even though these functions will never be
1534 called when UTF-8 support is disabled. Otherwise there are problems with a
1535 shared library.
1536
1537 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1538
1539 (a) It was defining its arguments as char * instead of void *.
1540
1541 (b) It was assuming that all moves were upwards in memory; this was true
1542 a long time ago when I wrote it, but is no longer the case.
1543
1544 The emulated memove() is provided for those environments that have neither
1545 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1546 is clearly not the case, as these two bugs were recently reported.
1547
1548 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1549 and Detrail to create the HTML documentation, the .txt form of the man
1550 pages, and it removes trailing spaces from listed files. It also creates
1551 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1552 case, it wraps all the #defines with #ifndefs. This script should be run
1553 before "make dist".
1554
1555 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1556 with Unicode property support.
1557
1558 (a) For a maximizing quantifier, if the two different cases of the
1559 character were of different lengths in their UTF-8 codings (there are
1560 some cases like this - I found 11), and the matching function had to
1561 back up over a mixture of the two cases, it incorrectly assumed they
1562 were both the same length.
1563
1564 (b) When PCRE was configured to use the heap rather than the stack for
1565 recursion during matching, it was not correctly preserving the data for
1566 the other case of a UTF-8 character when checking ahead for a match
1567 while processing a minimizing repeat. If the check also involved
1568 matching a wide character, but failed, corruption could cause an
1569 erroneous result when trying to check for a repeat of the original
1570 character.
1571
1572 13. Some tidying changes to the testing mechanism:
1573
1574 (a) The RunTest script now detects the internal link size and whether there
1575 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1576 values substituted by "configure". (The RunGrepTest script already did
1577 this for UTF-8.) The configure.ac script no longer substitutes the
1578 relevant variables.
1579
1580 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1581 with length and offset values. This means that the output is different
1582 for different internal link sizes. Test 2 is skipped for link sizes
1583 other than 2 because of this, bypassing the problem. Unfortunately,
1584 there was also a test in test 3 (the locale tests) that used /B and
1585 failed for link sizes other than 2. Rather than cut the whole test out,
1586 I have added a new /Z option to pcretest that replaces the length and
1587 offset values with spaces. This is now used to make test 3 independent
1588 of link size. (Test 2 will be tidied up later.)
1589
1590 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1591 segmentation fault instead of returning the appropriate error message.
1592
1593 15. In multiline mode when the newline sequence was set to "any", the pattern
1594 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1595 This doesn't seem right; it now treats the CRLF combination as the line
1596 ending, and so does not match in that case. It's only a pattern such as ^$
1597 that would hit this one: something like ^ABC$ would have failed after \r
1598 and then tried again after \r\n.
1599
1600 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1601 in an attempt to make files that differ only in their line terminators
1602 compare equal. This works on Linux.
1603
1604 17. Under certain error circumstances pcregrep might try to free random memory
1605 as it exited. This is now fixed, thanks to valgrind.
1606
1607 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1608 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1609 was because its rules for how to advance for /g after matching an empty
1610 string at the end of a line did not allow for this case. They now check for
1611 it specially.
1612
1613 20. pcretest is supposed to handle patterns and data of any length, by
1614 extending its buffers when necessary. It was getting this wrong when the
1615 buffer for a data line had to be extended.
1616
1617 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1618 CRLF as a newline sequence.
1619
1620 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1621 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1622 I have nevertheless tidied it up.
1623
1624 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1625
1626 24. Added a man page for pcre-config.
1627
1628
1629 Version 7.0 19-Dec-06
1630 ---------------------
1631
1632 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1633 moving to gcc 4.1.1.
1634
1635 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1636 sys/time.h, which is documented as needed for this function. It doesn't
1637 seem to matter on Linux, but it showed up on some releases of OS X.
1638
1639 3. It seems that there are systems where bytes whose values are greater than
1640 127 match isprint() in the "C" locale. The "C" locale should be the
1641 default when a C program starts up. In most systems, only ASCII printing
1642 characters match isprint(). This difference caused the output from pcretest
1643 to vary, making some of the tests fail. I have changed pcretest so that:
1644
1645 (a) When it is outputting text in the compiled version of a pattern, bytes
1646 other than 32-126 are always shown as hex escapes.
1647
1648 (b) When it is outputting text that is a matched part of a subject string,
1649 it does the same, unless a different locale has been set for the match
1650 (using the /L modifier). In this case, it uses isprint() to decide.
1651
1652 4. Fixed a major bug that caused incorrect computation of the amount of memory
1653 required for a compiled pattern when options that changed within the
1654 pattern affected the logic of the preliminary scan that determines the
1655 length. The relevant options are -x, and -i in UTF-8 mode. The result was
1656 that the computed length was too small. The symptoms of this bug were
1657 either the PCRE error "internal error: code overflow" from pcre_compile(),
1658 or a glibc crash with a message such as "pcretest: free(): invalid next
1659 size (fast)". Examples of patterns that provoked this bug (shown in
1660 pcretest format) are:
1661
1662 /(?-x: )/x
1663 /(?x)(?-x: \s*#\s*)/
1664 /((?i)[\x{c0}])/8
1665 /(?i:[\x{c0}])/8
1666
1667 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1668 is now done differently.
1669
1670 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1671 wrapper classes; (b) implement a new function in the C++ scanner that is
1672 more efficient than the old way of doing things because it avoids levels of
1673 recursion in the regex matching; (c) add a paragraph to the documentation
1674 for the FullMatch() function.
1675
1676 6. The escape sequence \n was being treated as whatever was defined as
1677 "newline". Not only was this contrary to the documentation, which states
1678 that \n is character 10 (hex 0A), but it also went horribly wrong when
1679 "newline" was defined as CRLF. This has been fixed.
1680
1681 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1682 was being set to -1 for the "end of line" case (supposedly a value that no
1683 character can have). Though this value is never used (the check for end of
1684 line is "zero bytes in current character"), it caused compiler complaints.
1685 I've changed it to 0xffffffff.
1686
1687 8. In pcre_version.c, the version string was being built by a sequence of
1688 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1689 string (as it is for production releases) called a macro with an empty
1690 argument. The C standard says the result of this is undefined. The gcc
1691 compiler treats it as an empty string (which was what was wanted) but it is
1692 reported that Visual C gives an error. The source has been hacked around to
1693 avoid this problem.
1694
1695 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1696 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1697 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1698 of them did).
1699
1700 10. Originally, pcretest opened its input and output without "b"; then I was
1701 told that "b" was needed in some environments, so it was added for release
1702 5.0 to both the input and output. (It makes no difference on Unix-like
1703 systems.) Later I was told that it is wrong for the input on Windows. I've
1704 now abstracted the modes into two macros, to make it easier to fiddle with
1705 them, and removed "b" from the input mode under Windows.
1706
1707 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1708
1709 12. Added -help and --help to pcretest as an official way of being reminded
1710 of the options.
1711
1712 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1713 and pcrecpp.cc because they annoy compilers at high warning levels.
1714
1715 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1716
1717 15. Fixed an occurrence of == in configure.ac that should have been = (shell
1718 scripts are not C programs :-) and which was not noticed because it works
1719 on Linux.
1720
1721 16. pcretest is supposed to handle any length of pattern and data line (as one
1722 line or as a continued sequence of lines) by extending its input buffer if
1723 necessary. This feature was broken for very long pattern lines, leading to
1724 a string of junk being passed to pcre_compile() if the pattern was longer
1725 than about 50K.
1726
1727 17. I have done a major re-factoring of the way pcre_compile() computes the
1728 amount of memory needed for a compiled pattern. Previously, there was code
1729 that made a preliminary scan of the pattern in order to do this. That was
1730 OK when PCRE was new, but as the facilities have expanded, it has become
1731 harder and harder to keep it in step with the real compile phase, and there
1732 have been a number of bugs (see for example, 4 above). I have now found a
1733 cunning way of running the real compile function in a "fake" mode that
1734 enables it to compute how much memory it would need, while actually only
1735 ever using a few hundred bytes of working memory and without too many
1736 tests of the mode. This should make future maintenance and development
1737 easier. A side effect of this work is that the limit of 200 on the nesting
1738 depth of parentheses has been removed (though this was never a serious
1739 limitation, I suspect). However, there is a downside: pcre_compile() now
1740 runs more slowly than before (30% or more, depending on the pattern). I
1741 hope this isn't a big issue. There is no effect on runtime performance.
1742
1743 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1744 newline (only possible for the last line of a file) and it was a
1745 pattern that set a locale (followed by /Lsomething), pcretest crashed.
1746
1747 19. Added additional timing features to pcretest. (1) The -tm option now times
1748 matching only, not compiling. (2) Both -t and -tm can be followed, as a
1749 separate command line item, by a number that specifies the number of
1750 repeats to use when timing. The default is 50000; this gives better
1751 precision, but takes uncomfortably long for very large patterns.
1752
1753 20. Extended pcre_study() to be more clever in cases where a branch of a
1754 subpattern has no definite first character. For example, (a*|b*)[cd] would
1755 previously give no result from pcre_study(). Now it recognizes that the
1756 first character must be a, b, c, or d.
1757
1758 21. There was an incorrect error "recursive call could loop indefinitely" if
1759 a subpattern (or the entire pattern) that was being tested for matching an
1760 empty string contained only one non-empty item after a nested subpattern.
1761 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1762 incorrectly, because the \d was being skipped in the check.
1763
1764 22. The pcretest program now has a new pattern option /B and a command line
1765 option -b, which is equivalent to adding /B to every pattern. This causes
1766 it to show the compiled bytecode, without the additional information that
1767 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1768 is the same as /B/I).
1769
1770 23. A new optimization is now able automatically to treat some sequences such
1771 as a*b as a*+b. More specifically, if something simple (such as a character
1772 or a simple class like \d) has an unlimited quantifier, and is followed by
1773 something that cannot possibly match the quantified thing, the quantifier
1774 is automatically "possessified".
1775
1776 24. A recursive reference to a subpattern whose number was greater than 39
1777 went wrong under certain circumstances in UTF-8 mode. This bug could also
1778 have affected the operation of pcre_study().
1779
1780 25. Realized that a little bit of performance could be had by replacing
1781 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1782
1783 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1784
1785 27. Possessive quantifiers such as a++ were previously implemented by turning
1786 them into atomic groups such as ($>a+). Now they have their own opcodes,
1787 which improves performance. This includes the automatically created ones
1788 from 23 above.
1789
1790 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1791 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1792 the first matched character to be a colon. This applied both to named and
1793 numbered groups.
1794
1795 29. The ucpinternal.h header file was missing its idempotency #ifdef.
1796
1797 30. I was sent a "project" file called libpcre.a.dev which I understand makes
1798 building PCRE on Windows easier, so I have included it in the distribution.
1799
1800 31. There is now a check in pcretest against a ridiculously large number being
1801 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1802 loop, the loop is abandoned.
1803
1804 32. Forward references to subpatterns in conditions such as (?(2)...) where
1805 subpattern 2 is defined later cause pcre_compile() to search forwards in
1806 the pattern for the relevant set of parentheses. This search went wrong
1807 when there were unescaped parentheses in a character class, parentheses
1808 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1809
1810 33. "Subroutine" calls and backreferences were previously restricted to
1811 referencing subpatterns earlier in the regex. This restriction has now
1812 been removed.
1813
1814 34. Added a number of extra features that are going to be in Perl 5.10. On the
1815 whole, these are just syntactic alternatives for features that PCRE had
1816 previously implemented using the Python syntax or my own invention. The
1817 other formats are all retained for compatibility.
1818
1819 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1820 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1821 also .NET compatible.
1822
1823 (b) A recursion or subroutine call to a named group can now be defined as
1824 (?&name) as well as (?P>name).
1825
1826 (c) A backreference to a named group can now be defined as \k<name> or
1827 \k'name' as well as (?P=name). The new forms, as well as being in Perl
1828 5.10, are also .NET compatible.
1829
1830 (d) A conditional reference to a named group can now use the syntax
1831 (?(<name>) or (?('name') as well as (?(name).
1832
1833 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1834 groups (named and numbered) that are never evaluated inline, but can be
1835 called as "subroutines" from elsewhere. In effect, the DEFINE condition
1836 is always false. There may be only one alternative in such a group.
1837
1838 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1839 as the simple (?(R). The condition is true only if the most recent
1840 recursion is that of the given number or name. It does not search out
1841 through the entire recursion stack.
1842
1843 (g) The escape \gN or \g{N} has been added, where N is a positive or
1844 negative number, specifying an absolute or relative reference.
1845
1846 35. Tidied to get rid of some further signed/unsigned compiler warnings and
1847 some "unreachable code" warnings.
1848
1849 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1850 things, this adds five new scripts.
1851
1852 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1853 There were also incompatibilities regarding the handling of \Q..\E inside
1854 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1855 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1856
1857 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1858 matches an empty string, and forcibly breaks the loop. There were bugs in
1859 this code in non-simple cases. For a pattern such as ^(a()*)* matched
1860 against aaaa the result was just "a" rather than "aaaa", for example. Two
1861 separate and independent bugs (that affected different cases) have been
1862 fixed.
1863
1864 39. Refactored the code to abolish the use of different opcodes for small
1865 capturing bracket numbers. This is a tidy that I avoided doing when I
1866 removed the limit on the number of capturing brackets for 3.5 back in 2001.
1867 The new approach is not only tidier, it makes it possible to reduce the
1868 memory needed to fix the previous bug (38).
1869
1870 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1871 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1872 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1873 mode.
1874
1875 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1876 report.
1877
1878 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1879 copy construction and assignment in the C++ wrapper.
1880
1881 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1882 couple of bugs that could have given wrong results in the "--newline=crlf"
1883 case.
1884
1885 44. Added a number of casts and did some reorganization of signed/unsigned int
1886 variables following suggestions from Dair Grant. Also renamed the variable
1887 "this" as "item" because it is a C++ keyword.
1888
1889 45. Arranged for dftables to add
1890
1891 #include "pcre_internal.h"
1892
1893 to pcre_chartables.c because without it, gcc 4.x may remove the array
1894 definition from the final binary if PCRE is built into a static library and
1895 dead code stripping is activated.
1896
1897 46. For an unanchored pattern, if a match attempt fails at the start of a
1898 newline sequence, and the newline setting is CRLF or ANY, and the next two
1899 characters are CRLF, advance by two characters instead of one.
1900
1901
1902 Version 6.7 04-Jul-06
1903 ---------------------
1904
1905 1. In order to handle tests when input lines are enormously long, pcretest has
1906 been re-factored so that it automatically extends its buffers when
1907 necessary. The code is crude, but this _is_ just a test program. The
1908 default size has been increased from 32K to 50K.
1909
1910 2. The code in pcre_study() was using the value of the re argument before
1911 testing it for NULL. (Of course, in any sensible call of the function, it
1912 won't be NULL.)
1913
1914 3. The memmove() emulation function in pcre_internal.h, which is used on
1915 systems that lack both memmove() and bcopy() - that is, hardly ever -
1916 was missing a "static" storage class specifier.
1917
1918 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1919 containing an extended class (one that cannot be represented by a bitmap
1920 because it contains high-valued characters or Unicode property items, e.g.
1921 [\pZ]). Almost always one would set UTF-8 mode when processing such a
1922 pattern, but PCRE should not loop if you do not (it no longer does).
1923 [Detail: two cases were found: (a) a repeated subpattern containing an
1924 extended class; (b) a recursive reference to a subpattern that followed a
1925 previous extended class. It wasn't skipping over the extended class
1926 correctly when UTF-8 mode was not set.]
1927
1928 5. A negated single-character class was not being recognized as fixed-length
1929 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1930 compile error "lookbehind assertion is not fixed length".
1931
1932 6. The RunPerlTest auxiliary script was showing an unexpected difference
1933 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1934 write a Perl script that can interpret lines of an input file either as
1935 byte characters or as UTF-8, which is what "perltest" was being required to
1936 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1937 can't do is switch easily at run time between having the "use utf8;" pragma
1938 or not. In the end, I fudged it by using the RunPerlTest script to insert
1939 "use utf8;" explicitly for the UTF-8 tests.
1940
1941 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1942 the end of the subject string, contrary to the documentation and to what
1943 Perl does. This was true of both matching functions. Now it matches only at
1944 the start of the subject and immediately after *internal* newlines.
1945
1946 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1947 a pointer to an int instead of a pointer to an unsigned long int. This
1948 caused problems on 64-bit systems.
1949
1950 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1951 instance of the 'standard' template library not being so standard".
1952
1953 10. There was no check on the number of named subpatterns nor the maximum
1954 length of a subpattern name. The product of these values is used to compute
1955 the size of the memory block for a compiled pattern. By supplying a very
1956 long subpattern name and a large number of named subpatterns, the size
1957 computation could be caused to overflow. This is now prevented by limiting
1958 the length of names to 32 characters, and the number of named subpatterns
1959 to 10,000.
1960
1961 11. Subpatterns that are repeated with specific counts have to be replicated in
1962 the compiled pattern. The size of memory for this was computed from the
1963 length of the subpattern and the repeat count. The latter is limited to
1964 65535, but there was no limit on the former, meaning that integer overflow
1965 could in principle occur. The compiled length of a repeated subpattern is
1966 now limited to 30,000 bytes in order to prevent this.
1967
1968 12. Added the optional facility to have named substrings with the same name.
1969
1970 13. Added the ability to use a named substring as a condition, using the
1971 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1972 are numbers (not recommended). Forward references are permitted.
1973
1974 14. Added forward references in named backreferences (if you see what I mean).
1975
1976 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1977 pattern could run off the end of the subject. For example, the pattern
1978 "(?s)(.{1,5})"8 did this with the subject "ab".
1979
1980 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1981 PCRE_CASELESS was set when matching characters that were quantified with ?
1982 or *.
1983
1984 17. A character class other than a single negated character that had a minimum
1985 but no maximum quantifier - for example [ab]{6,} - was not handled
1986 correctly by pce_dfa_exec(). It would match only one character.
1987
1988 18. A valid (though odd) pattern that looked like a POSIX character
1989 class but used an invalid character after [ (for example [[,abc,]]) caused
1990 pcre_compile() to give the error "Failed: internal error: code overflow" or
1991 in some cases to crash with a glibc free() error. This could even happen if
1992 the pattern terminated after [[ but there just happened to be a sequence of
1993 letters, a binary zero, and a closing ] in the memory that followed.
1994
1995 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1996 over the years. Originally (before any Unicode support), just the bottom 8
1997 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1998 output from "man perlunicode" includes this:
1999
2000 The regular expression compiler produces polymorphic opcodes. That
2001 is, the pattern adapts to the data and automatically switches to
2002 the Unicode character scheme when presented with Unicode data--or
2003 instead uses a traditional byte scheme when presented with byte
2004 data.
2005
2006 Sadly, a wide octal escape does not cause a switch, and in a string with
2007 no other multibyte characters, these octal escapes are treated as before.
2008 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2009 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2010 Unicode string.
2011
2012 I have not perpetrated such confusion in PCRE. Up till now, it took just
2013 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2014 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2015 translate to the appropriate multibyte character.
2016
2017 29. Applied some refactoring to reduce the number of warnings from Microsoft
2018 and Borland compilers. This has included removing the fudge introduced
2019 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2020 a warning about an unused variable.
2021
2022 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2023 characters since release 4.0, because Perl (from release 5.004) does not.
2024 [Or at least, is documented not to: some releases seem to be in conflict
2025 with the documentation.] However, when a pattern was studied with
2026 pcre_study() and all its branches started with \s, PCRE still included VT
2027 as a possible starting character. Of course, this did no harm; it just
2028 caused an unnecessary match attempt.
2029
2030 22. Removed a now-redundant internal flag bit that recorded the fact that case
2031 dependency changed within the pattern. This was once needed for "required
2032 byte" processing, but is no longer used. This recovers a now-scarce options
2033 bit. Also moved the least significant internal flag bit to the most-
2034 significant bit of the word, which was not previously used (hangover from
2035 the days when it was an int rather than a uint) to free up another bit for
2036 the future.
2037
2038 23. Added support for CRLF line endings as well as CR and LF. As well as the
2039 default being selectable at build time, it can now be changed at runtime
2040 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2041 specify that it is scanning data with non-default line endings.
2042
2043 24. Changed the definition of CXXLINK to make it agree with the definition of
2044 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2045
2046 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2047 recursions. This makes a big different to stack usage for some patterns.
2048
2049 26. If a subpattern containing a named recursion or subroutine reference such
2050 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2051 the space required for the compiled pattern went wrong and gave too small a
2052 value. Depending on the environment, this could lead to "Failed: internal
2053 error: code overflow at offset 49" or "glibc detected double free or
2054 corruption" errors.
2055
2056 27. Applied patches from Google (a) to support the new newline modes and (b) to
2057 advance over multibyte UTF-8 characters in GlobalReplace.
2058
2059 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2060 difference for some implementation of PCRE in some Windows version.
2061
2062 29. Added some extra testing facilities to pcretest:
2063
2064 \q<number> in a data line sets the "match limit" value
2065 \Q<number> in a data line sets the "match recursion limt" value
2066 -S <number> sets the stack size, where <number> is in megabytes
2067
2068 The -S option isn't available for Windows.
2069
2070
2071 Version 6.6 06-Feb-06
2072 ---------------------
2073
2074 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2075 in pcreposix.h. I have copied the definition from pcre.h.
2076
2077 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2078 because pcre.h is no longer a built file.
2079
2080 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2081 not normally included in the compiled code.
2082
2083
2084 Version 6.5 01-Feb-06
2085 ---------------------
2086
2087 1. When using the partial match feature with pcre_dfa_exec(), it was not
2088 anchoring the second and subsequent partial matches at the new starting
2089 point. This could lead to incorrect results. For example, with the pattern
2090 /1234/, partially matching against "123" and then "a4" gave a match.
2091
2092 2. Changes to pcregrep:
2093
2094 (a) All non-match returns from pcre_exec() were being treated as failures
2095 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2096 error message is output. Some extra information is given for the
2097 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2098 probably the only errors that are likely to be caused by users (by
2099 specifying a regex that has nested indefinite repeats, for instance).
2100 If there are more than 20 of these errors, pcregrep is abandoned.
2101
2102 (b) A binary zero was treated as data while matching, but terminated the
2103 output line if it was written out. This has been fixed: binary zeroes
2104 are now no different to any other data bytes.
2105
2106 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2107 used to set a locale for matching. The --locale=xxxx long option has
2108 been added (no short equivalent) to specify a locale explicitly on the
2109 pcregrep command, overriding the environment variables.
2110
2111 (d) When -B was used with -n, some line numbers in the output were one less
2112 than they should have been.
2113
2114 (e) Added the -o (--only-matching) option.
2115
2116 (f) If -A or -C was used with -c (count only), some lines of context were
2117 accidentally printed for the final match.
2118
2119 (g) Added the -H (--with-filename) option.
2120
2121 (h) The combination of options -rh failed to suppress file names for files
2122 that were found from directory arguments.
2123
2124 (i) Added the -D (--devices) and -d (--directories) options.
2125
2126 (j) Added the -F (--fixed-strings) option.
2127
2128 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2129
2130 (l) Added the --colo(u)r option.
2131
2132 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2133 is not present by default.
2134
2135 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2136 items such as (?R) or (?1), when the recursion could match a number of
2137 alternatives. If it matched one of the alternatives, but subsequently,
2138 outside the recursion, there was a failure, the code tried to back up into
2139 the recursion. However, because of the way PCRE is implemented, this is not
2140 possible, and the result was an incorrect result from the match.
2141
2142 In order to prevent this happening, the specification of recursion has
2143 been changed so that all such subpatterns are automatically treated as
2144 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2145
2146 4. I had overlooked the fact that, in some locales, there are characters for
2147 which isalpha() is true but neither isupper() nor islower() are true. In
2148 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2149 and ordfeminine) are like this. This affected the treatment of \w and \W
2150 when they appeared in character classes, but not when they appeared outside
2151 a character class. The bit map for "word" characters is now created
2152 separately from the results of isalnum() instead of just taking it from the
2153 upper, lower, and digit maps. (Plus the underscore character, of course.)
2154
2155 5. The above bug also affected the handling of POSIX character classes such as
2156 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2157 permanent tables. Instead, the bit maps for such a class were previously
2158 created as the appropriate unions of the upper, lower, and digit bitmaps.
2159 Now they are created by subtraction from the [[:word:]] class, which has
2160 its own bitmap.
2161
2162 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2163 It is created by subtracting the vertical space characters (\x09, \x0a,
2164 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2165 subtraction was done in the overall bitmap for a character class, meaning
2166 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2167 be recognized. This bug has been fixed.
2168
2169 7. Patches from the folks at Google:
2170
2171 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2172 real life, but is still worth protecting against".
2173
2174 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2175 regular expressions".
2176
2177 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2178 have it.
2179
2180 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2181 "configure" and the latter not, in order to fix a problem somebody had
2182 with compiling the Arg class on HP-UX.
2183
2184 (e) Improve the error-handling of the C++ wrapper a little bit.
2185
2186 (f) New tests for checking recursion limiting.
2187
2188 8. The pcre_memmove() function, which is used only if the environment does not
2189 have a standard memmove() function (and is therefore rarely compiled),
2190 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2191 returning a result (though PCRE never actually uses the result).
2192
2193 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2194 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2195 returned instead of calling malloc() with an overflowing number that would
2196 most likely cause subsequent chaos.
2197
2198 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2199
2200 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2201 with this option is matched, the nmatch and pmatch options of regexec() are
2202 ignored.
2203
2204 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2205 provided in case anyone wants to the the POSIX interface with UTF-8
2206 strings.
2207
2208 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2209 C++ linking (needed for some HP-UX environments).
2210
2211 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2212 (unused parameter) and in the pcre_printint() function (omitted "default"
2213 switch label when the default is to do nothing).
2214
2215 15. Added some code to make it possible, when PCRE is compiled as a C++
2216 library, to replace subject pointers for pcre_exec() with a smart pointer
2217 class, thus making it possible to process discontinuous strings.
2218
2219 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2220 much the same function. They were added by different people who were trying
2221 to make PCRE easy to compile on non-Unix systems. It has been suggested
2222 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2223 for compiling on Windows systems. I have therefore replaced it with
2224 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2225 defaults to "extern" for C or "extern C" for C++, which works fine on
2226 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2227 SCOPE with something explicit in config.h. In addition:
2228
2229 (a) pcreposix.h still had just "extern" instead of either of these macros;
2230 I have replaced it with PCRE_DATA_SCOPE.
2231
2232 (b) Functions such as _pcre_xclass(), which are internal to the library,
2233 but external in the C sense, all had PCRE_EXPORT in their definitions.
2234 This is apparently wrong for the Windows case, so I have removed it.
2235 (It makes no difference on Unix-like systems.)
2236
2237 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2238 of recursive calls to match(). This is different to MATCH_LIMIT because
2239 that limits the total number of calls to match(), not all of which increase
2240 the depth of recursion. Limiting the recursion depth limits the amount of
2241 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2242 when PCRE is compiled, and changed at run time. A patch from Google adds
2243 this functionality to the C++ interface.
2244
2245 18. Changes to the handling of Unicode character properties:
2246
2247 (a) Updated the table to Unicode 4.1.0.
2248
2249 (b) Recognize characters that are not in the table as "Cn" (undefined).
2250
2251 (c) I revised the way the table is implemented to a much improved format
2252 which includes recognition of ranges. It now supports the ranges that
2253 are defined in UnicodeData.txt, and it also amalgamates other
2254 characters into ranges. This has reduced the number of entries in the
2255 table from around 16,000 to around 3,000, thus reducing its size
2256 considerably. I realized I did not need to use a tree structure after
2257 all - a binary chop search is just as efficient. Having reduced the
2258 number of entries, I extended their size from 6 bytes to 8 bytes to
2259 allow for more data.
2260
2261 (d) Added support for Unicode script names via properties such as \p{Han}.
2262
2263 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2264 matching that character.
2265
2266 20. When matching a repeated Unicode property with a minimum greater than zero,
2267 (for example \pL{2,}), PCRE could look past the end of the subject if it
2268 reached it while seeking the minimum number of characters. This could
2269 happen only if some of the characters were more than one byte long, because
2270 there is a check for at least the minimum number of bytes.
2271
2272 21. Refactored the implementation of \p and \P so as to be more general, to
2273 allow for more different types of property in future. This has changed the
2274 compiled form incompatibly. Anybody with saved compiled patterns that use
2275 \p or \P will have to recompile them.
2276
2277 22. Added "Any" and "L&" to the supported property types.
2278
2279 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2280 but give a compile time error if the value is greater than 0xff.
2281
2282 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2283 accidentally not being installed or uninstalled.
2284
2285 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2286 made were to insert the current release number. This seemed silly, because
2287 it made things harder for people building PCRE on systems that don't run
2288 "configure". I have turned pcre.h into a distributed file, no longer built
2289 by "configure", with the version identification directly included. There is
2290 no longer a pcre.h.in file.
2291
2292 However, this change necessitated a change to the pcre-config script as
2293 well. It is built from pcre-config.in, and one of the substitutions was the
2294 release number. I have updated configure.ac so that ./configure now finds
2295 the release number by grepping pcre.h.
2296
2297 26. Added the ability to run the tests under valgrind.
2298
2299
2300 Version 6.4 05-Sep-05
2301 ---------------------
2302
2303 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2304 "--" to be printed when multiple files were scanned, even when none of the
2305 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2306 consider it to be a bug, and have restored the previous behaviour.
2307
2308 2. A couple of code tidies to get rid of compiler warnings.
2309
2310 3. The pcretest program used to cheat by referring to symbols in the library
2311 whose names begin with _pcre_. These are internal symbols that are not
2312 really supposed to be visible externally, and in some environments it is
2313 possible to suppress them. The cheating is now confined to including
2314 certain files from the library's source, which is a bit cleaner.
2315
2316 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2317 file's purpose clearer.
2318
2319 5. Reorganized pcre_ucp_findchar().
2320
2321
2322 Version 6.3 15-Aug-05
2323 ---------------------
2324
2325 1. The file libpcre.pc.in did not have general read permission in the tarball.
2326
2327 2. There were some problems when building without C++ support:
2328
2329 (a) If C++ support was not built, "make install" and "make test" still
2330 tried to test it.
2331
2332 (b) There were problems when the value of CXX was explicitly set. Some
2333 changes have been made to try to fix these, and ...
2334
2335 (c) --disable-cpp can now be used to explicitly disable C++ support.
2336
2337 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2338 backslash in a target when C++ was disabled. This confuses some
2339 versions of "make", apparently. Using an intermediate variable solves
2340 this. (Same for CPP_LOBJ.)
2341
2342 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2343 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2344 necessary on certain architectures.
2345
2346 4. Added a setting of -export-symbols-regex to the link command to remove
2347 those symbols that are exported in the C sense, but actually are local
2348 within the library, and not documented. Their names all begin with
2349 "_pcre_". This is not a perfect job, because (a) we have to except some
2350 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2351 available (and never for static libraries). I have made a note to try to
2352 find a way round (a) in the future.
2353
2354
2355 Version 6.2 01-Aug-05
2356 ---------------------
2357
2358 1. There was no test for integer overflow of quantifier values. A construction
2359 such as {1111111111111111} would give undefined results. What is worse, if
2360 a minimum quantifier for a parenthesized subpattern overflowed and became
2361 negative, the calculation of the memory size went wrong. This could have
2362 led to memory overwriting.
2363
2364 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2365
2366 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2367 operating environments where this matters.
2368
2369 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2370 PCRE options from within the C++ wrapper.
2371
2372 5. Named capturing subpatterns were not being correctly counted when a pattern
2373 was compiled. This caused two problems: (a) If there were more than 100
2374 such subpatterns, the calculation of the memory needed for the whole
2375 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2376 back references of the form \12, where the number was greater than 9, were
2377 not recognized as back references, even though there were sufficient
2378 previous subpatterns.
2379
2380 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2381 versions of gcc, e.g. 2.95.4.
2382
2383
2384 Version 6.1 21-Jun-05
2385 ---------------------
2386
2387 1. There was one reference to the variable "posix" in pcretest.c that was not
2388 surrounded by "#if !defined NOPOSIX".
2389
2390 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2391 the cross-check on the old pcre_info() function, for the benefit of the
2392 cut-down version of PCRE that is currently imported into Exim.
2393
2394 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2395 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2396 patterns that start (?i) but I don't think that matters. The use of (?i) is
2397 just an example; this all applies to the other options as well.
2398
2399 4. Since libtool seems to echo the compile commands it is issuing, the output
2400 from "make" can be reduced a bit by putting "@" in front of each libtool
2401 compile command.
2402
2403 5. Patch from the folks at Google for configure.in to be a bit more thorough
2404 in checking for a suitable C++ installation before trying to compile the
2405 C++ stuff. This should fix a reported problem when a compiler was present,
2406 but no suitable headers.
2407
2408 6. The man pages all had just "PCRE" as their title. I have changed them to
2409 be the relevant file name. I have also arranged that these names are
2410 retained in the file doc/pcre.txt, which is a concatenation in text format
2411 of all the man pages except the little individual ones for each function.
2412
2413 7. The NON-UNIX-USE file had not been updated for the different set of source
2414 files that come with release 6. I also added a few comments about the C++
2415 wrapper.
2416
2417
2418 Version 6.0 07-Jun-05
2419 ---------------------
2420
2421 1. Some minor internal re-organization to help with my DFA experiments.
2422
2423 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2424 didn't matter for the library itself when fully configured, but did matter
2425 when compiling without UCP support, or within Exim, where the ucp files are
2426 not imported.
2427
2428 3. Refactoring of the library code to split up the various functions into
2429 different source modules. The addition of the new DFA matching code (see
2430 below) to a single monolithic source would have made it really too
2431 unwieldy, quite apart from causing all the code to be include in a
2432 statically linked application, when only some functions are used. This is
2433 relevant even without the DFA addition now that patterns can be compiled in
2434 one application and matched in another.
2435
2436 The downside of splitting up is that there have to be some external
2437 functions and data tables that are used internally in different modules of
2438 the library but which are not part of the API. These have all had their
2439 names changed to start with "_pcre_" so that they are unlikely to clash
2440 with other external names.
2441
2442 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2443 a different (DFA) algorithm. Although it is slower than the original
2444 function, it does have some advantages for certain types of matching
2445 problem.
2446
2447 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2448 including restarting after a partial match.
2449
2450 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2451 defined when compiling for Windows was sent to me. I have put it into the
2452 code, though I have no means of testing or verifying it.
2453
2454 7. Added the pcre_refcount() auxiliary function.
2455
2456 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2457 match before or at the first newline in the subject string. In pcretest,
2458 the /f option on a pattern can be used to set this.
2459
2460 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2461 would behave wrongly. This has been present in PCRE since release 4.0.
2462
2463 10. A number of changes to the pcregrep command:
2464
2465 (a) Refactored how -x works; insert ^(...)$ instead of setting
2466 PCRE_ANCHORED and checking the length, in preparation for adding
2467 something similar for -w.
2468
2469 (b) Added the -w (match as a word) option.
2470
2471 (c) Refactored the way lines are read and buffered so as to have more
2472 than one at a time available.
2473
2474 (d) Implemented a pcregrep test script.
2475
2476 (e) Added the -M (multiline match) option. This allows patterns to match
2477 over several lines of the subject. The buffering ensures that at least
2478 8K, or the rest of the document (whichever is the shorter) is available
2479 for matching (and similarly the previous 8K for lookbehind assertions).
2480
2481 (f) Changed the --help output so that it now says
2482
2483 -w, --word-regex(p)
2484
2485 instead of two lines, one with "regex" and the other with "regexp"
2486 because that confused at least one person since the short forms are the
2487 same. (This required a bit of code, as the output is generated
2488 automatically from a table. It wasn't just a text change.)
2489
2490 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2491 option but starts with a hyphen. Could be a pattern or a path name
2492 starting with a hyphen, for instance.
2493
2494 (h) "-" can be given as a file name to represent stdin.
2495
2496 (i) When file names are being printed, "(standard input)" is used for
2497 the standard input, for compatibility with GNU grep. Previously
2498 "<stdin>" was used.
2499
2500 (j) The option --label=xxx can be used to supply a name to be used for
2501 stdin when file names are being printed. There is no short form.
2502
2503 (k) Re-factored the options decoding logic because we are going to add
2504 two more options that take data. Such options can now be given in four
2505 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2506
2507 (l) Added the -A, -B, and -C options for requesting that lines of context
2508 around matches be printed.
2509
2510 (m) Added the -L option to print the names of files that do not contain
2511 any matching lines, that is, the complement of -l.
2512
2513 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2514 continue to scan other files.
2515
2516 (o) The -s option was incorrectly implemented. For compatibility with other
2517 greps, it now suppresses the error message for a non-existent or non-
2518 accessible file (but not the return code). There is a new option called
2519 -q that suppresses the output of matching lines, which was what -s was
2520 previously doing.
2521
2522 (p) Added --include and --exclude options to specify files for inclusion
2523 and exclusion when recursing.
2524
2525 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2526 Hopefully, it now does.
2527
2528 12. Missing cast in pcre_study().
2529
2530 13. Added an "uninstall" target to the makefile.
2531
2532 14. Replaced "extern" in the function prototypes in Makefile.in with
2533 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2534 world, but is set differently for Windows.
2535
2536 15. Added a second compiling function called pcre_compile2(). The only
2537 difference is that it has an extra argument, which is a pointer to an
2538 integer error code. When there is a compile-time failure, this is set
2539 non-zero, in addition to the error test pointer being set to point to an
2540 error message. The new argument may be NULL if no error number is required
2541 (but then you may as well call pcre_compile(), which is now just a
2542 wrapper). This facility is provided because some applications need a
2543 numeric error indication, but it has also enabled me to tidy up the way
2544 compile-time errors are handled in the POSIX wrapper.
2545
2546 16. Added VPATH=.libs to the makefile; this should help when building with one
2547 prefix path and installing with another. (Or so I'm told by someone who
2548 knows more about this stuff than I do.)
2549
2550 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2551 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2552 match everything, including newlines. This is not POSIX-compatible, but
2553 somebody wanted the feature. From pcretest it can be activated by using
2554 both the P and the s flags.
2555
2556 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2557
2558 19. libpcre.pc was being incorrectly installed as executable.
2559
2560 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2561 it now also looks for '\r' so that it will work unmodified on Windows.
2562
2563 21. Added Google's contributed C++ wrapper to the distribution.
2564
2565 22. Added some untidy missing memory free() calls in pcretest, to keep
2566 Electric Fence happy when testing.
2567
2568
2569
2570 Version 5.0 13-Sep-04
2571 ---------------------
2572
2573 1. Internal change: literal characters are no longer packed up into items
2574 containing multiple characters in a single byte-string. Each character
2575 is now matched using a separate opcode. However, there may be more than one
2576 byte in the character in UTF-8 mode.
2577
2578 2. The pcre_callout_block structure has two new fields: pattern_position and
2579 next_item_length. These contain the offset in the pattern to the next match
2580 item, and its length, respectively.
2581
2582 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2583 insertion of callouts before each pattern item. Added the /C option to
2584 pcretest to make use of this.
2585
2586 4. On the advice of a Windows user, the lines
2587
2588 #if defined(_WIN32) || defined(WIN32)
2589 _setmode( _fileno( stdout ), 0x8000 );
2590 #endif /* defined(_WIN32) || defined(WIN32) */
2591
2592 have been added to the source of pcretest. This apparently does useful
2593 magic in relation to line terminators.
2594
2595 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2596 for the benefit of those environments where the "b" makes a difference.
2597
2598 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2599 to know about it. I have put a hack into configure.in that adds in code
2600 to set GCC=yes if CC=icc. This seems to end up at a point in the
2601 generated configure script that is early enough to affect the setting of
2602 compiler options, which is what is needed, but I have no means of testing
2603 whether it really works. (The user who reported this had patched the
2604 generated configure script, which of course I cannot do.)
2605
2606 LATER: After change 22 below (new libtool files), the configure script
2607 seems to know about icc (and also ecc). Therefore, I have commented out
2608 this hack in configure.in.
2609
2610 7. Added support for pkg-config (2 patches were sent in).
2611
2612 8. Negated POSIX character classes that used a combination of internal tables
2613 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2614 [[:^ascii]]. Typically, they would match almost any characters. The other
2615 POSIX classes were not broken in this way.
2616
2617 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2618 to find the match, as PCRE was deluded into thinking that the match had to
2619 start at the start point or following a newline. The same bug applied to
2620 patterns with negative forward assertions or any backward assertions
2621 preceding ".*" at the start, unless the pattern required a fixed first
2622 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2623
2624 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2625 starting at the last subject character, bytes beyond the end of the subject
2626 string were read.
2627
2628 11. Renamed the variable "class" as "classbits" to make life easier for C++
2629 users. (Previously there was a macro definition, but it apparently wasn't
2630 enough.)
2631
2632 12. Added the new field "tables" to the extra data so that tables can be passed
2633 in at exec time, or the internal tables can be re-selected. This allows
2634 a compiled regex to be saved and re-used at a later time by a different
2635 program that might have everything at different addresses.
2636
2637 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2638 -R library as well as a -L library.
2639
2640 14. The debugging options of pcretest (-d on the command line or D on a
2641 pattern) showed incorrect output for anything following an extended class
2642 that contained multibyte characters and which was followed by a quantifier.
2643
2644 15. Added optional support for general category Unicode character properties
2645 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2646 support. It adds about 90K to the size of the library. The meanings of the
2647 inbuilt class escapes such as \d and \s have NOT been changed.
2648
2649 16. Updated pcredemo.c to include calls to free() to release the memory for the
2650 compiled pattern.
2651
2652 17. The generated file chartables.c was being created in the source directory
2653 instead of in the building directory. This caused the build to fail if the
2654 source directory was different from the building directory, and was
2655 read-only.
2656
2657 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2658 file. No doubt somebody will tell me if they don't make sense... Also added
2659 Dan Mooney's comments about building on OpenVMS.
2660
2661 19. Added support for partial matching via the PCRE_PARTIAL option for
2662 pcre_exec() and the \P data escape in pcretest.
2663
2664 20. Extended pcretest with 3 new pattern features:
2665
2666 (i) A pattern option of the form ">rest-of-line" causes pcretest to
2667 write the compiled pattern to the file whose name is "rest-of-line".
2668 This is a straight binary dump of the data, with the saved pointer to
2669 the character tables forced to be NULL. The study data, if any, is
2670 written too. After writing, pcretest reads a new pattern.
2671
2672 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2673 compiled pattern from the given file. There must not be any
2674 occurrences of "<" in the file name (pretty unlikely); if there are,
2675 pcretest will instead treat the initial "<" as a pattern delimiter.
2676 After reading in the pattern, pcretest goes on to read data lines as
2677 usual.
2678
2679 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2680 and 16-bit fields in a compiled pattern, to simulate a pattern that
2681 was compiled on a host of opposite endianness.
2682
2683 21. The pcre-exec() function can now cope with patterns that were compiled on
2684 hosts of opposite endianness, with this restriction:
2685
2686 As for any compiled expression that is saved and used later, the tables
2687 pointer field cannot be preserved; the extra_data field in the arguments
2688 to pcre_exec() should be used to pass in a tables address if a value
2689 other than the default internal tables were used at compile time.
2690
2691 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2692 now diagnosed as an error. Previously, most of the time, a negative number
2693 would have been treated as zero, but if in addition "ovector" was passed as
2694 NULL, a crash could occur.
2695
2696 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2697 new versions from the libtool 1.5 distribution (the last one is a copy of
2698 a file called libtool.m4). This seems to have fixed the need to patch
2699 "configure" to support Darwin 1.3 (which I used to do). However, I still
2700 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2701 workstation).
2702
2703 24. Changed the PCRE licence to be the more standard "BSD" licence.
2704
2705
2706 Version 4.5 01-Dec-03
2707 ---------------------
2708
2709 1. There has been some re-arrangement of the code for the match() function so
2710 that it can be compiled in a version that does not call itself recursively.
2711 Instead, it keeps those local variables that need separate instances for
2712 each "recursion" in a frame on the heap, and gets/frees frames whenever it
2713 needs to "recurse". Keeping track of where control must go is done by means
2714 of setjmp/longjmp. The whole thing is implemented by a set of macros that
2715 hide most of the details from the main code, and operates only if
2716 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2717 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2718 operating.
2719
2720 To make it easier for callers to provide specially tailored get/free
2721 functions for this usage, two new functions, pcre_stack_malloc, and
2722 pcre_stack_free, are used. They are always called in strict stacking order,
2723 and the size of block requested is always the same.
2724
2725 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2726 PCRE has been compiled to use the stack or the heap for recursion. The
2727 -C option of pcretest uses this to show which version is compiled.
2728
2729 A new data escape \S, is added to pcretest; it causes the amounts of store
2730 obtained and freed by both kinds of malloc/free at match time to be added
2731 to the output.
2732
2733 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2734 what's available on my current Linux desktop machine.
2735
2736 3. When matching a UTF-8 string, the test for a valid string at the start has
2737 been extended. If start_offset is not zero, PCRE now checks that it points
2738 to a byte that is the start of a UTF-8 character. If not, it returns
2739 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2740 this is necessary because there may be backward assertions in the pattern.
2741 When matching the same subject several times, it may save resources to use
2742 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2743
2744 4. The code for checking the validity of UTF-8 strings has been tightened so
2745 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2746 containing "overlong sequences".
2747
2748 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2749 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2750 should have been "&", but it just so happened that all the cases this let
2751 through by mistake were picked up later in the function.
2752
2753 6. I had used a variable called "isblank" - this is a C99 function, causing
2754 some compilers to warn. To avoid this, I renamed it (as "blankclass").
2755
2756 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2757 prompting; (b) run "./pcretest /dev/null" at the start of the test script
2758 so the version is shown; (c) stop "make test" echoing "./RunTest".
2759
2760 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2761
2762 9. The prototype for memmove() for systems that don't have it was using
2763 size_t, but the inclusion of the header that defines size_t was later. I've
2764 moved the #includes for the C headers earlier to avoid this.
2765
2766 10. Added some adjustments to the code to make it easier to compiler on certain
2767 special systems:
2768
2769 (a) Some "const" qualifiers were missing.
2770 (b) Added the macro EXPORT before all exported functions; by default this
2771 is defined to be empty.
2772 (c) Changed the dftables auxiliary program (that builds chartables.c) so
2773 that it reads its output file name as an argument instead of writing
2774 to the standard output and assuming this can be redirected.
2775
2776 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2777 class containing characters with values greater than 255, PCRE compilation
2778 went into a loop.
2779
2780 12. A recursive reference to a subpattern that was within another subpattern
2781 that had a minimum quantifier of zero caused PCRE to crash. For example,
2782 (x(y(?2))z)? provoked this bug with a subject that got as far as the
2783 recursion. If the recursively-called subpattern itself had a zero repeat,
2784 that was OK.
2785
2786 13. In pcretest, the buffer for reading a data line was set at 30K, but the
2787 buffer into which it was copied (for escape processing) was still set at
2788 1024, so long lines caused crashes.
2789
2790 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2791 "internal error: code overflow...". This applied to any character class
2792 that was followed by a possessive quantifier.
2793
2794 15. Modified the Makefile to add libpcre.la as a prerequisite for
2795 libpcreposix.la because I was told this is needed for a parallel build to
2796 work.
2797
2798 16. If a pattern that contained .* following optional items at the start was
2799 studied, the wrong optimizing data was generated, leading to matching
2800 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2801 matching string must start with a or b or c. The correct conclusion for
2802 this pattern is that a match can start with any character.
2803
2804
2805 Version 4.4 13-Aug-03
2806 ---------------------
2807
2808 1. In UTF-8 mode, a character class containing characters with values between
2809 127 and 255 was not handled correctly if the compiled pattern was studied.
2810 In fixing this, I have also improved the studying algorithm for such
2811 classes (slightly).
2812
2813 2. Three internal functions had redundant arguments passed to them. Removal
2814 might give a very teeny performance improvement.
2815
2816 3. Documentation bug: the value of the capture_top field in a callout is *one
2817 more than* the number of the hightest numbered captured substring.
2818
2819 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2820 in incorrectly linking with a previously installed version. They now link
2821 explicitly with libpcre.la.
2822
2823 5. configure.in no longer needs to recognize Cygwin specially.
2824
2825 6. A problem in pcre.in for Windows platforms is fixed.
2826
2827 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2828 pcretest, it used to include the size of the study block as part of its
2829 output. Unfortunately, the structure contains a field that has a different
2830 size on different hardware architectures. This meant that the tests that
2831 showed this size failed. As the block is currently always of a fixed size,
2832 this information isn't actually particularly useful in pcretest output, so
2833 I have just removed it.
2834
2835 8. Three pre-processor statements accidentally did not start in column 1.
2836 Sadly, there are *still* compilers around that complain, even though
2837 standard C has not required this for well over a decade. Sigh.
2838
2839 9. In pcretest, the code for checking callouts passed small integers in the
2840 callout_data field, which is a void * field. However, some picky compilers
2841 complained about the casts involved for this on 64-bit systems. Now
2842 pcretest passes the address of the small integer instead, which should get
2843 rid of the warnings.
2844
2845 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2846 both compile and run time, and gives an error if an invalid UTF-8 sequence
2847 is found. There is a option for disabling this check in cases where the
2848 string is known to be correct and/or the maximum performance is wanted.
2849
2850 11. In response to a bug report, I changed one line in Makefile.in from
2851
2852 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2853 to
2854 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2855
2856 to look similar to other lines, but I have no way of telling whether this
2857 is the right thing to do, as I do not use Windows. No doubt I'll get told
2858 if it's wrong...
2859
2860
2861 Version 4.3 21-May-03
2862 ---------------------
2863
2864 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2865 Makefile.
2866
2867 2. Some refactoring to improve the quality of the code:
2868
2869 (i) The utf8_table... variables are now declared "const".
2870
2871 (ii) The code for \cx, which used the "case flipping" table to upper case
2872 lower case letters, now just substracts 32. This is ASCII-specific,
2873 but the whole concept of \cx is ASCII-specific, so it seems
2874 reasonable.
2875
2876 (iii) PCRE was using its character types table to recognize decimal and
2877 hexadecimal digits in the pattern. This is silly, because it handles
2878 only 0-9, a-f, and A-F, but the character types table is locale-
2879 specific, which means strange things might happen. A private
2880 table is now used for this - though it costs 256 bytes, a table is
2881 much faster than multiple explicit tests. Of course, the standard
2882 character types table is still used for matching digits in subject
2883 strings against \d.
2884
2885 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2886 ending in _t are). So I've renamed it as ESC_tee.
2887
2888 3. The first argument for regexec() in the POSIX wrapper should have been
2889 defined as "const".
2890
2891 4. Changed pcretest to use malloc() for its buffers so that they can be
2892 Electric Fenced for debugging.
2893
2894 5. There were several places in the code where, in UTF-8 mode, PCRE would try
2895 to read one or more bytes before the start of the subject string. Often this
2896 had no effect on PCRE's behaviour, but in some circumstances it could
2897 provoke a segmentation fault.
2898
2899 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2900 to try to read one or more bytes before the start of the subject string.
2901
2902 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2903 UTF-8 support could misbehave in various ways if the subject string
2904 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2905 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2906 back over UTF-8 characters.)
2907
2908
2909 Version 4.2 14-Apr-03
2910 ---------------------
2911
2912 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2913
2914 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2915 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2916 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2917 [WIN_PREFIX]: new variable, "cyg" for Cygwin
2918 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2919 and BUILD_EXEEXT
2920 Note: automatic setting of the BUILD variables is not yet working
2921 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2922 compile-time but not at link-time
2923 [LINK]: use for linking executables only
2924 make different versions for Windows and non-Windows
2925 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2926 libraries
2927 [LINK_FOR_BUILD]: new variable
2928 [OBJEXT]: use throughout
2929 [EXEEXT]: use throughout
2930 <winshared>: new target
2931 <wininstall>: new target
2932 <dftables.o>: use native compiler
2933 <dftables>: use native linker
2934 <install>: handle Windows platform correctly
2935 <clean>: ditto
2936 <check>: ditto
2937 copy DLL to top builddir before testing
2938
2939 As part of these changes, -no-undefined was removed again. This was reported
2940 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2941 in any case.
2942
2943 3. Some tidies to get rid of compiler warnings:
2944
2945 . In the match_data structure, match_limit was an unsigned long int, whereas
2946 match_call_count was an int. I've made them both unsigned long ints.
2947
2948 . In pcretest the fact that a const uschar * doesn't automatically cast to
2949 a void * provoked a warning.
2950
2951 . Turning on some more compiler warnings threw up some "shadow" variables
2952 and a few more missing casts.
2953
2954 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2955 option, a class that contained a single character with a value between 128
2956 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2957
2958 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2959 option, a class that contained several characters, but with at least one
2960 whose value was between 128 and 255 caused PCRE to crash.
2961
2962
2963 Version 4.1 12-Mar-03
2964 ---------------------
2965
2966 1. Compiling with gcc -pedantic found a couple of places where casts were
2967 needed, and a string in dftables.c that was longer than standard compilers are
2968 required to support.
2969
2970 2. Compiling with Sun's compiler found a few more places where the code could
2971 be tidied up in order to avoid warnings.
2972
2973 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2974 first of these names is deprecated in the latest Autoconf in favour of the name
2975 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2976 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2977 analogy I have changed it to CFLAGS_FOR_BUILD.
2978
2979 4. Added -no-undefined to the linking command in the Makefile, because this is
2980 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2981 linking step for the pcreposix library.
2982
2983 5. PCRE was failing to diagnose the case of two named groups with the same
2984 name.
2985
2986 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2987 literal character that is needed in the subject for a match, and scans along to
2988 ensure that it is present before embarking on the full matching process. This
2989 saves time in cases of nested unlimited repeats that are never going to match.
2990 Problem: the scan can take a lot of time if the subject is very long (e.g.
2991 megabytes), thus penalizing straightforward matches. It is now done only if the
2992 amount of subject to be scanned is less than 1000 bytes.
2993
2994 7. A lesser problem with the same optimization is that it was recording the
2995 first character of an anchored pattern as "needed", thus provoking a search
2996 right along the subject, even when the first match of the pattern was going to
2997 fail. The "needed" character is now not set for anchored patterns, unless it
2998 follows something in the pattern that is of non-fixed length. Thus, it still
2999 fulfils its original purpose of finding quick non-matches in cases of nested
3000 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3001
3002
3003 Version 4.0 17-Feb-03
3004 ---------------------
3005
3006 1. If a comment in an extended regex that started immediately after a meta-item
3007 extended to the end of string, PCRE compiled incorrect data. This could lead to
3008 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3009
3010 2. Moved to autoconf 2.53 and libtool 1.4.2.
3011
3012 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3013 the special perltest8 script is no longer needed - all the tests can be run
3014 from a single perltest script.
3015
3016 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3017 by \s. It has now been removed in PCRE. This means it isn't recognized as
3018 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3019 class [:space:] *does* include VT, thereby creating a mess.
3020
3021 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3022 space and tab.
3023
3024 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3025 its new features into the main test script, reducing the number of scripts.
3026
3027 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3028 were backward compatible, and made the (?i) apply to the whole pattern, as if
3029 /i were given. Now it behaves more logically, and applies the option setting
3030 only to what follows. PCRE has been changed to follow suit. However, if it
3031 finds options settings right at the start of the pattern, it extracts them into
3032 the global options, as before. Thus, they show up in the info data.
3033
3034 8. Added support for the \Q...\E escape sequence. Characters in between are
3035 treated as literals. This is slightly different from Perl in that $ and @ are
3036 also handled as literals inside the quotes. In Perl, they will cause variable
3037 interpolation. Note the following examples:
3038
3039 Pattern PCRE matches Perl matches
3040
3041 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3042 \Qabc\$xyz\E abc\$xyz abc\$xyz
3043 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3044
3045 For compatibility with Perl, \Q...\E sequences are recognized inside character
3046 classes as well as outside them.
3047
3048 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3049 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3050 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3051 signed/unsigned warnings.
3052
3053 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3054 option for pcretest, so I've replaced it by a simple function that does just
3055 that job.
3056
3057 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3058 "pcregrep -".
3059
3060 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3061 Java package. This provides some syntactic sugar for simple cases of what my
3062 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3063 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3064 item, you can use this simplified notation. Note that only makes sense with
3065 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3066 greediness, whatever the setting of the PCRE_UNGREEDY option.
3067
3068 13. A change of greediness default within a pattern was not taking effect at
3069 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3070 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3071 was abstracted outside.
3072
3073 14. PCRE now supports the \G assertion. It is true when the current matching
3074 position is at the start point of the match. This differs from \A when the
3075 starting offset is non-zero. Used with the /g option of pcretest (or similar
3076 code), it works in the same way as it does for Perl's /g option. If all
3077 alternatives of a regex begin with \G, the expression is anchored to the start
3078 match position, and the "anchored" flag is set in the compiled expression.
3079
3080 15. Some bugs concerning the handling of certain option changes within patterns
3081 have been fixed. These applied to options other than (?ims). For example,
3082 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3083 been the other way round. Some of this was related to change 7 above.
3084
3085 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3086 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3087 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3088 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3089
3090 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3091 mode. Unlike ".", it always matches newline, whatever the setting of
3092 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3093 assertions. Perl allows it, but it doesn't (in general) work because it can't
3094 calculate the length of the lookbehind. At least, that's the case for Perl
3095 5.8.0 - I've been told they are going to document that it doesn't work in
3096 future.
3097
3098 18. Added an error diagnosis for escapes that PCRE does not support: these are
3099 \L, \l, \N, \P, \p, \U, \u, and \X.
3100
3101 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3102 reading past the end of the pattern in cases such as /[abcd/.
3103
3104 20. PCRE was getting more memory than necessary for patterns with classes that
3105 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3106
3107 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3108 compiling PCRE for use with Virtual Pascal.
3109
3110 22. Small fix to the Makefile to make it work properly if the build is done
3111 outside the source tree.
3112
3113 23. Added a new extension: a condition to go with recursion. If a conditional
3114 subpattern starts with (?(R) the "true" branch is used if recursion has
3115 happened, whereas the "false" branch is used only at the top level.
3116
3117 24. When there was a very long string of literal characters (over 255 bytes
3118 without UTF support, over 250 bytes with UTF support), the computation of how
3119 much memory was required could be incorrect, leading to segfaults or other
3120 strange effects.
3121
3122 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3123 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3124 there was a subsequent back reference to those brackets. This meant that, for
3125 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3126 possible to check for precisely this case. All we can do is abandon the
3127 optimization if .* occurs inside capturing brackets when there are any back
3128 references whatsoever. (See below for a better fix that came later.)
3129
3130 26. The handling of the optimization for finding the first character of a
3131 non-anchored pattern, and for finding a character that is required later in the
3132 match were failing in some cases. This didn't break the matching; it just
3133 failed to optimize when it could. The way this is done has been re-implemented.
3134
3135 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3136
3137 28. Added a new feature that provides some of the functionality that Perl
3138 provides with (?{...}). The facility is termed a "callout". The way it is done
3139 in PCRE is for the caller to provide an optional function, by setting
3140 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3141 global variable. By default it is unset, which disables all calling out. To get
3142 the function called, the regex must include (?C) at appropriate points. This
3143 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3144 This provides a means of identifying different callout points. When PCRE
3145 reaches such a point in the regex, if pcre_callout has been set, the external
3146 function is called. It is provided with data in a structure called
3147 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3148 matching continues; if it returns a non-zero value, the match at the current
3149 point fails. However, backtracking will occur if possible. [This was changed
3150 later and other features added - see item 49 below.]
3151
3152 29. pcretest is upgraded to test the callout functionality. It provides a
3153 callout function that displays information. By default, it shows the start of
3154 the match and the current position in the text. There are some new data escapes
3155 to vary what happens:
3156
3157 \C+ in addition, show current contents of captured substrings
3158 \C- do not supply a callout function
3159 \C!n return 1 when callout number n is reached
3160 \C!n!m return 1 when callout number n is reached for the mth time
3161
3162 30. If pcregrep was called with the -l option and just a single file name, it
3163 output "<stdin>" if a match was found, instead of the file name.
3164
3165 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3166 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3167 pcre_exec(). This saves a malloc/free per call. The default value of
3168 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3169 when configuring.
3170
3171 32. The default maximum size of a compiled pattern is 64K. There have been a
3172 few cases of people hitting this limit. The code now uses macros to handle the
3173 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3174 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3175 configuring. Tests 2 and 5 work only with 2-byte links because they output
3176 debugging information about compiled patterns.
3177
3178 33. Internal code re-arrangements:
3179
3180 (a) Moved the debugging function for printing out a compiled regex into
3181 its own source file (printint.c) and used #include to pull it into
3182 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3183 separate copies.
3184
3185 (b) Defined the list of op-code names for debugging as a macro in
3186 internal.h so that it is next to the definition of the opcodes.
3187
3188 (c) Defined a table of op-code lengths for simpler skipping along compiled
3189 code. This is again a macro in internal.h so that it is next to the
3190 definition of the opcodes.
3191
3192 34. Added support for recursive calls to individual subpatterns, along the
3193 lines of Robin Houston's patch (but implemented somewhat differently).
3194
3195 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3196 allow it to read and process whole directories in Win32. This code was
3197 contributed by Lionel Fourquaux; it has not been tested by me.
3198
3199 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3200 used to name a group. Names consist of alphanumerics and underscores, and must
3201 be unique. Back references use the syntax (?P=name) and recursive calls use
3202 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3203 numbers. The function pcre_fullinfo() can be used after compilation to extract
3204 a name/number map. There are three relevant calls:
3205
3206 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3207 PCRE_INFO_NAMECOUNT yields the number of entries
3208 PCRE_INFO_NAMETABLE yields a pointer to the map.
3209
3210 The map is a vector of fixed-size entries. The size of each entry depends on
3211 the length of the longest name used. The first two bytes of each entry are the
3212 group number, most significant byte first. There follows the corresponding
3213 name, zero terminated. The names are in alphabetical order.
3214
3215 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3216 case instead of 255. Making it the same both with and without UTF-8 support
3217 means that the same test output works with both.
3218
3219 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3220 calling malloc() with a zero argument.
3221
3222 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3223 optimization. I've improved things by keeping a bitmap of backreferences with
3224 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3225 fact referenced, the optimization can be applied. It is unlikely that a
3226 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3227 the match to follow \n) will appear inside brackets with a number greater than
3228 31, but if it does, any back reference > 31 suppresses the optimization.
3229
3230 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3231 of disabling numbered capturing parentheses. Any opening parenthesis that is
3232 not followed by ? behaves as if it were followed by ?: but named parentheses
3233 can still be used for capturing (and they will acquire numbers in the usual
3234 way).
3235
3236 41. Redesigned the return codes from the match() function into yes/no/error so
3237 that errors can be passed back from deep inside the nested calls. A malloc
3238 failure while inside a recursive subpattern call now causes the
3239 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3240
3241 42. It is now possible to set a limit on the number of times the match()
3242 function is called in a call to pcre_exec(). This facility makes it possible to
3243 limit the amount of recursion and backtracking, though not in a directly
3244 obvious way, because the match() function is used in a number of different
3245 circumstances. The count starts from zero for each position in the subject
3246 string (for non-anchored patterns). The default limit is, for compatibility, a
3247 large number, namely 10 000 000. You can change this in two ways:
3248
3249 (a) When configuring PCRE before making, you can use --with-match-limit=n
3250 to set a default value for the compiled library.
3251
3252 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3253 a different value is set. See 45 below.
3254
3255 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3256
3257 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3258 of things that can be changed at compile time. The first argument specifies
3259 what is wanted and the second points to where the information is to be placed.
3260 The current list of available information is:
3261
3262 PCRE_CONFIG_UTF8
3263
3264 The output is an integer that is set to one if UTF-8 support is available;
3265 otherwise it is set to zero.
3266
3267 PCRE_CONFIG_NEWLINE
3268
3269 The output is an integer that it set to the value of the code that is used for
3270 newline. It is either LF (10) or CR (13).
3271
3272 PCRE_CONFIG_LINK_SIZE
3273
3274 The output is an integer that contains the number of bytes used for internal
3275 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3276
3277 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3278
3279 The output is an integer that contains the threshold above which the POSIX
3280 interface uses malloc() for output vectors. See item 31 above.
3281
3282 PCRE_CONFIG_MATCH_LIMIT
3283
3284 The output is an unsigned integer that contains the default limit of the number
3285 of match() calls in a pcre_exec() execution. See 42 above.
3286
3287 44. pcretest has been upgraded by the addition of the -C option. This causes it
3288 to extract all the available output from the new pcre_config() function, and to
3289 output it. The program then exits immediately.
3290
3291 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3292 order to support additional features. One way would have been to define
3293 pcre_exec2() (for example) with extra arguments, but this would not have been
3294 extensible, and would also have required all calls to the original function to
3295 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3296 is used for passing in "extra" data from pcre_study().
3297
3298 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3299 contains the following fields:
3300
3301 flags a bitmap indicating which of the following fields are set
3302 study_data opaque data from pcre_study()
3303 match_limit a way of specifying a limit on match() calls for a specific
3304 call to pcre_exec()
3305 callout_data data for callouts (see 49 below)
3306
3307 The flag bits are also defined in pcre.h, and are
3308
3309 PCRE_EXTRA_STUDY_DATA
3310 PCRE_EXTRA_MATCH_LIMIT
3311 PCRE_EXTRA_CALLOUT_DATA
3312
3313 The pcre_study() function now returns one of these new pcre_extra blocks, with
3314 the actual study data pointed to by the study_data field, and the
3315 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3316 before. That is, this change is entirely upwards-compatible and requires no
3317 change to existing code.
3318
3319 If you want to pass in additional data to pcre_exec(), you can either place it
3320 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3321 block.
3322
3323 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3324 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3325 times with different match limits, until it finds the minimum value needed for
3326 pcre_exec() to complete. The value is then output. This can be instructive; for
3327 most simple matches the number is quite small, but for pathological cases it
3328 gets very large very quickly.
3329
3330 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3331 returns the size of the data block pointed to by the study_data field in a
3332 pcre_extra block, that is, the value that was passed as the argument to
3333 pcre_malloc() when PCRE was getting memory in which to place the information
3334 created by pcre_study(). The fourth argument should point to a size_t variable.
3335 pcretest has been extended so that this information is shown after a successful
3336 pcre_study() call when information about the compiled regex is being displayed.
3337
3338 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3339 because what follows is always an absolute path. (Later: it turns out that this
3340 is more than cosmetic for MinGW, because it doesn't like empty path
3341 components.)
3342
3343 49. Some changes have been made to the callout feature (see 28 above):
3344
3345 (i) A callout function now has three choices for what it returns:
3346
3347 0 => success, carry on matching
3348 > 0 => failure at this point, but backtrack if possible
3349 < 0 => serious error, return this value from pcre_exec()
3350
3351 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3352 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3353 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3354 use by callout functions. It will never be used by PCRE itself.
3355
3356 (ii) The pcre_extra structure (see 45 above) has a void * field called
3357 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3358 pcre_callout_block structure has a field of the same name. The contents of
3359 the field passed in the pcre_extra structure are passed to the callout
3360 function in the corresponding field in the callout block. This makes it
3361 easier to use the same callout-containing regex from multiple threads. For
3362 testing, the pcretest program has a new data escape
3363
3364 \C*n pass the number n (may be negative) as callout_data
3365
3366 If the callout function in pcretest receives a non-zero value as
3367 callout_data, it returns that value.
3368
3369 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3370 there were some redundant $(CFLAGS) in commands that are now specified as
3371 $(LINK), which already includes $(CFLAGS).
3372
3373 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3374 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3375 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3376 one-byte characters throughout. Note that case-insensitive matching applies
3377 only to characters whose values are less than 256. PCRE doesn't support the
3378 notion of cases for higher-valued characters.
3379
3380 (i) A character class whose characters are all within 0-255 is handled as
3381 a bit map, and the map is inverted for negative classes. Previously, a
3382 character > 255 always failed to match such a class; however it should
3383 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3384
3385 (ii) A negated character class with a single character < 255 is coded as
3386 "not this character" (OP_NOT). This wasn't working properly when the test
3387 character was multibyte, either singly or repeated.
3388
3389 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3390 mode, for example: \x{100}{2,3}.
3391
3392 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3393 singly or repeated) now correctly test multibyte characters. However,
3394 PCRE doesn't recognize any characters with values greater than 255 as
3395 digits, spaces, or word characters. Such characters always match \D, \S,
3396 and \W, and never match \d, \s, or \w.
3397
3398 (v) Classes may now contain characters and character ranges with values
3399 greater than 255. For example: [ab\x{100}-\x{400}].
3400
3401 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3402 PCRE in UTF-8 mode.
3403
3404 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3405 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3406 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3407 value.)
3408
3409 53. The single man page has become too large. I have therefore split it up into
3410 a number of separate man pages. These also give rise to individual HTML pages;
3411 these are now put in a separate directory, and there is an index.html page that
3412 lists them all. Some hyperlinking between the pages has been installed.
3413
3414 54. Added convenience functions for handling named capturing parentheses.
3415
3416 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3417 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3418 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3419 are faulted.
3420
3421 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3422 calling configure. These values are used when compiling the dftables.c program
3423 which is run to generate the source of the default character tables. They
3424 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3425 you will need to set these values.
3426
3427 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3428
3429
3430 Version 3.9 02-Jan-02
3431 ---------------------
3432
3433 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3434
3435 2. If --disable-static was given, the building process failed when trying to
3436 build pcretest and pcregrep. (For some reason it was using libtool to compile
3437 them, which is not right, as they aren't part of the library.)
3438
3439
3440 Version 3.8 18-Dec-01
3441 ---------------------
3442
3443 1. The experimental UTF-8 code was completely screwed up. It was packing the
3444 bytes in the wrong order. How dumb can you get?
3445
3446
3447 Version 3.7 29-Oct-01
3448 ---------------------
3449
3450 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3451 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3452 this didn't happen under Solaris 8, where I normally test things.
3453
3454 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3455 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3456 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3457 this fix an hour or so after the initial 3.7 release.)
3458
3459
3460 Version 3.6 23-Oct-01
3461 ---------------------
3462
3463 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3464 offsets passed as NULL with zero offset count.
3465
3466 2. The config.guess and config.sub files had not been updated when I moved to
3467 the latest autoconf.
3468
3469
3470 Version 3.5 15-Aug-01
3471 ---------------------
3472
3473 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3474 had been forgotten.
3475
3476 2. By using declared but undefined structures, we can avoid using "void"
3477 definitions in pcre.h while keeping the internal definitions of the structures
3478 private.
3479
3480 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3481 user point of view, this means that both static and shared libraries are built
3482 by default, but this can be individually controlled. More of the work of
3483 handling this static/shared cases is now inside libtool instead of PCRE's make
3484 file.
3485
3486 4. The pcretest utility is now installed along with pcregrep because it is
3487 useful for users (to test regexs) and by doing this, it automatically gets
3488 relinked by libtool. The documentation has been turned into a man page, so
3489 there are now .1, .txt, and .html versions in /doc.
3490
3491 5. Upgrades to pcregrep:
3492 (i) Added long-form option names like gnu grep.
3493 (ii) Added --help to list all options with an explanatory phrase.
3494 (iii) Added -r, --recursive to recurse into sub-directories.
3495 (iv) Added -f, --file to read patterns from a file.
3496
3497 6. pcre_exec() was referring to its "code" argument before testing that
3498 argument for NULL (and giving an error if it was NULL).
3499
3500 7. Upgraded Makefile.in to allow for compiling in a different directory from
3501 the source directory.
3502
3503 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3504 options bits, the pointer it was passed was to an int instead of to an unsigned
3505 long int. This mattered only on 64-bit systems.
3506
3507 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3508 generated) instead of pcre.in, which it its source. Also made the same change
3509 in several of the .c files.
3510
3511 10. A new release of gcc defines printf() as a macro, which broke pcretest
3512 because it had an ifdef in the middle of a string argument for printf(). Fixed
3513 by using separate calls to printf().
3514
3515 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3516 script, to force use of CR or LF instead of \n in the source. On non-Unix
3517 systems, the value can be set in config.h.
3518
3519 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3520 absolute limit. Changed the text of the error message to make this clear, and
3521 likewise updated the man page.
3522
3523 13. The limit of 99 on the number of capturing subpatterns has been removed.
3524 The new limit is 65535, which I hope will not be a "real" limit.
3525
3526
3527 Version 3.4 22-Aug-00
3528 ---------------------
3529
3530 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3531
3532 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3533
3534
3535 Version 3.3 01-Aug-00
3536 ---------------------
3537
3538 1. If an octal character was given, but the value was greater than \377, it
3539 was not getting masked to the least significant bits, as documented. This could
3540 lead to crashes in some systems.
3541
3542 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3543 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3544
3545 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3546 These just pass their arguments on to (pcre_free)(), but they are provided
3547 because some uses of PCRE bind it to non-C systems that can call its functions,
3548 but cannot call free() or pcre_free() directly.
3549
3550 4. Add "make test" as a synonym for "make check". Corrected some comments in
3551 the Makefile.
3552
3553 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3554 Makefile.
3555
3556 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3557 command called pgrep for grepping around the active processes.
3558
3559 7. Added the beginnings of support for UTF-8 character strings.
3560
3561 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3562 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3563 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3564 out for the ar command.)
3565
3566
3567 Version 3.2 12-May-00
3568 ---------------------
3569
3570 This is purely a bug fixing release.
3571
3572 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3573 of ZA. This was just one example of several cases that could provoke this bug,
3574 which was introduced by change 9 of version 2.00. The code for breaking
3575 infinite loops after an iteration that matches an empty string was't working
3576 correctly.
3577
3578 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3579 when matched against abbab (for example). After matching an empty string, it
3580 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3581 caused it to match further down the string than it should.
3582
3583 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3584 was there because it doesn't seem to be needed, and it causes trouble on some
3585 systems, as it is not a Standard C header. It has been removed.
3586
3587 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3588 were reported on the Macintosh. The changes were from
3589
3590 while ((c = *(++ptr)) != 0 && c != '\n');
3591 to
3592 while ((c = *(++ptr)) != 0 && c != '\n') ;
3593
3594 Totally extraordinary, but if that's what it takes...
3595
3596 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3597 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3598 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3599 assumes the way PCRE uses memmove() (always moving upwards).
3600
3601 6. PCRE is being used in one environment where strchr() is not available. There
3602 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3603 faster code anyway.
3604
3605
3606 Version 3.1 09-Feb-00
3607 ---------------------
3608
3609 The only change in this release is the fixing of some bugs in Makefile.in for
3610 the "install" target:
3611
3612 (1) It was failing to install pcreposix.h.
3613
3614 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3615
3616
3617 Version 3.0 01-Feb-00
3618 ---------------------
3619
3620 1. Add support for the /+ modifier to perltest (to output $` like it does in
3621 pcretest).
3622
3623 2. Add support for the /g modifier to perltest.
3624
3625 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3626 matches null strings.
3627
3628 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3629 pattern. Perl treats empty patterns specially - it reuses the most recent
3630 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3631 effect.
3632
3633 5. The POSIX interface was broken in that it was just handing over the POSIX
3634 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3635 required a bigger vector, with some working space on the end. This means that
3636 the POSIX wrapper now has to get and free some memory, and copy the results.
3637
3638 6. Added some simple autoconf support, placing the test data and the
3639 documentation in separate directories, re-organizing some of the
3640 information files, and making it build pcre-config (a GNU standard). Also added
3641 libtool support for building PCRE as a shared library, which is now the
3642 default.
3643
3644 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3645 09 are not valid octal constants. Single digits will be used for minor values
3646 less than 10.
3647
3648 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3649 existing programs that set these in the POSIX interface can use PCRE without
3650 modification.
3651
3652 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3653 return all that pcre_info() returns, plus additional data. The pcre_info()
3654 function is retained for compatibility, but is considered to be obsolete.
3655
3656 10. Added experimental recursion feature (?R) to handle one common case that
3657 Perl 5.6 will be able to do with (?p{...}).
3658
3659 11. Added support for POSIX character classes like [:alpha:], which Perl is
3660 adopting.
3661
3662
3663 Version 2.08 31-Aug-99
3664 ----------------------
3665
3666 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3667 trying to match at the startoffset position, but instead was moving forward to
3668 the next newline as if a previous match had failed.
3669
3670 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3671 and could get into a loop if a null string was matched other than at the start
3672 of the subject.
3673
3674 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3675 be distinguished at compile time, and for completeness also added PCRE_DATE.
3676
3677 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3678 in GnuWin32 environments.
3679
3680
3681 Version 2.07 29-Jul-99
3682 ----------------------
3683
3684 1. The documentation is now supplied in plain text form and HTML as well as in
3685 the form of man page sources.
3686
3687 2. C++ compilers don't like assigning (void *) values to other pointer types.
3688 In particular this affects malloc(). Although there is no problem in Standard
3689 C, I've put in casts to keep C++ compilers happy.
3690
3691 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3692 should be (const char *).
3693
3694 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3695 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3696 However, I haven't made this a standard facility. The documentation doesn't
3697 mention it, and the Makefile doesn't support it.
3698
3699 5. The Makefile now contains an "install" target, with editable destinations at
3700 the top of the file. The pcretest program is not installed.
3701
3702 6. pgrep -V now gives the PCRE version number and date.
3703
3704 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3705 causing the entire string to be ignored, instead of just the last character.
3706
3707 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3708 non-matching string, it can take a very, very long time, even for strings of
3709 quite modest length, because of the nested recursion. PCRE now does better in
3710 some of these cases. It does this by remembering the last required literal
3711 character in the pattern, and pre-searching the subject to ensure it is present
3712 before running the real match. In other words, it applies a heuristic to detect
3713 some types of certain failure quickly, and in the above example, if presented
3714 with a string that has no trailing " it gives "no match" very quickly.
3715
3716 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3717 other alternatives are tried instead.
3718
3719
3720 Version 2.06 09-Jun-99
3721 ----------------------
3722
3723 1. Change pcretest's output for amount of store used to show just the code
3724 space, because the remainder (the data block) varies in size between 32-bit and
3725 64-bit systems.
3726
3727 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3728 start matching at. This allows lookbehinds to work when searching for multiple
3729 occurrences in a string.
3730
3731 3. Added additional options to pcretest for testing multiple occurrences:
3732
3733 /+ outputs the rest of the string that follows a match
3734 /g loops for multiple occurrences, using the new startoffset argument
3735 /G loops for multiple occurrences by passing an incremented pointer
3736
3737 4. PCRE wasn't doing the "first character" optimization for patterns starting
3738 with \b or \B, though it was doing it for other lookbehind assertions. That is,
3739 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3740 the letter 'x'. On long subject strings, this gives a significant speed-up.
3741
3742
3743 Version 2.05 21-Apr-99
3744 ----------------------
3745
3746 1. Changed the type of magic_number from int to long int so that it works
3747 properly on 16-bit systems.
3748
3749 2. Fixed a bug which caused patterns starting with .* not to work correctly
3750 when the subject string contained newline characters. PCRE was assuming
3751 anchoring for such patterns in all cases, which is not correct because .* will
3752 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3753 DOTALL is set at top level; otherwise it knows that patterns starting with .*
3754 must be retried after every newline in the subject.
3755
3756
3757 Version 2.04 18-Feb-99
3758 ----------------------
3759
3760 1. For parenthesized subpatterns with repeats whose minimum was zero, the
3761 computation of the store needed to hold the pattern was incorrect (too large).
3762 If such patterns were nested a few deep, this could multiply and become a real
3763 problem.
3764
3765 2. Added /M option to pcretest to show the memory requirement of a specific
3766 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3767
3768 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3769 compiled in such a way that the backtracking after subsequent failure was
3770 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3771 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3772
3773
3774 Version 2.03 02-Feb-99
3775 ----------------------
3776
3777 1. Fixed typo and small mistake in man page.
3778
3779 2. Added 4th condition (GPL supersedes if conflict) and created separate
3780 LICENCE file containing the conditions.
3781
3782 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3783 Perl, that is the internal \ allows the delimiter to be included in the
3784 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3785 the final delimiter, add \ to the end of the pattern (to test the error).
3786
3787 4. Added the convenience functions for extracting substrings after a successful
3788 match. Updated pcretest to make it able to test these functions.
3789
3790
3791 Version 2.02 14-Jan-99
3792 ----------------------
3793
3794 1. Initialized the working variables associated with each extraction so that
3795 their saving and restoring doesn't refer to uninitialized store.
3796
3797 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3798 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3799 fix the problem.
3800
3801 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3802 calls, and wasn't printing the correct value for compiling calls. Increased the
3803 default value of LOOPREPEAT, and the number of significant figures in the
3804 times.
3805
3806 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3807
3808 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3809 a building problem on Windows NT with a FAT file system.
3810
3811
3812 Version 2.01 21-Oct-98
3813 ----------------------
3814
3815 1. Changed the API for pcre_compile() to allow for the provision of a pointer
3816 to character tables built by pcre_maketables() in the current locale. If NULL
3817 is passed, the default tables are used.
3818
3819
3820 Version 2.00 24-Sep-98
3821 ----------------------
3822
3823 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3824 it any more.
3825
3826 2. Allow quantification of (?>) groups, and make it work correctly.
3827
3828 3. The first character computation wasn't working for (?>) groups.
3829
3830 4. Correct the implementation of \Z (it is permitted to match on the \n at the
3831 end of the subject) and add 5.005's \z, which really does match only at the
3832 very end of the subject.
3833
3834 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3835
3836 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3837 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3838 localized options. All options to pcre_study() were also removed.
3839
3840 7. Add other new features from 5.005:
3841
3842 $(?<= positive lookbehind
3843 $(?<! negative lookbehind
3844 (?imsx-imsx) added the unsetting capability
3845 such a setting is global if at outer level; local otherwise
3846 (?imsx-imsx:) non-capturing groups with option setting
3847 (?(cond)re|re) conditional pattern matching
3848
3849 A backreference to itself in a repeated group matches the previous
3850 captured string.
3851
3852 8. General tidying up of studying (both automatic and via "study")
3853 consequential on the addition of new assertions.
3854
3855 9. As in 5.005, unlimited repeated groups that could match an empty substring
3856 are no longer faulted at compile time. Instead, the loop is forcibly broken at
3857 runtime if any iteration does actually match an empty substring.
3858
3859 10. Include the RunTest script in the distribution.
3860
3861 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3862 discrepancies, some of which were old and were also with respect to 5.004. They
3863 have now been fixed.
3864
3865
3866 Version 1.09 28-Apr-98
3867 ----------------------
3868
3869 1. A negated single character class followed by a quantifier with a minimum
3870 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3871 program crashes, or just wrong answers. This did not apply to negated classes
3872 containing more than one character, or to minima other than one.
3873
3874
3875 Version 1.08 27-Mar-98
3876 ----------------------
3877
3878 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3879
3880 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3881 latter must appear before anything that relies on it in the pattern.
3882
3883
3884 Version 1.07 16-Feb-98
3885 ----------------------
3886
3887 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3888 repeat of a potentially empty string).
3889
3890
3891 Version 1.06 23-Jan-98
3892 ----------------------
3893
3894 1. Added Markus Oberhumer's little patches for C++.
3895
3896 2. Literal strings longer than 255 characters were broken.
3897
3898
3899 Version 1.05 23-Dec-97
3900 ----------------------
3901
3902 1. Negated character classes containing more than one character were failing if
3903 PCRE_CASELESS was set at run time.
3904
3905
3906 Version 1.04 19-Dec-97
3907 ----------------------
3908
3909 1. Corrected the man page, where some "const" qualifiers had been omitted.
3910
3911 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3912 input syntax.
3913
3914 3. Fixed memory leak which occurred when a regex with back references was
3915 matched with an offsets vector that wasn't big enough. The temporary memory
3916 that is used in this case wasn't being freed if the match failed.
3917
3918 4. Tidied pcretest to ensure it frees memory that it gets.
3919
3920 5. Temporary memory was being obtained in the case where the passed offsets
3921 vector was exactly big enough.
3922
3923 6. Corrected definition of offsetof() from change 5 below.
3924
3925 7. I had screwed up change 6 below and broken the rules for the use of
3926 setjmp(). Now fixed.
3927
3928
3929 Version 1.03 18-Dec-97
3930 ----------------------
3931
3932 1. A erroneous regex with a missing opening parenthesis was correctly
3933 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3934 on some systems.
3935
3936 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3937 it was reported that one broken compiler failed on the former because "code" is
3938 also an independent variable.
3939
3940 3. The erroneous regex a[]b caused an array overrun reference.
3941
3942 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3943 fail on data ending with that character. (It was going on too far, and checking
3944 the next character, typically a binary zero.) This was specific to the
3945 optimized code for single-character negative classes.
3946
3947 5. Added a contributed patch from the TIN world which does the following:
3948
3949 + Add an undef for memmove, in case the the system defines a macro for it.
3950
3951 + Add a definition of offsetof(), in case there isn't one. (I don't know
3952 the reason behind this - offsetof() is part of the ANSI standard - but
3953 it does no harm).
3954
3955 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3956 most of the places where whitespace preceded '#'. I have given up and
3957 allowed the remaining 2 cases to be at the margin.
3958
3959 + Rename some variables in pcre to eliminate shadowing. This seems very
3960 pedantic, but does no harm, of course.
3961
3962 6. Moved the call to setjmp() into its own function, to get rid of warnings
3963 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3964
3965 7. Constructs such as \d{8,} were compiling into the equivalent of
3966 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3967 outcome, but in this particular case used more store than had been allocated,
3968 which caused the bug to be discovered because it threw up an internal error.
3969
3970 8. The debugging code in both pcre and pcretest for outputting the compiled
3971 form of a regex was going wrong in the case of back references followed by
3972 curly-bracketed repeats.
3973
3974
3975 Version 1.02 12-Dec-97
3976 ----------------------
3977
3978 1. Typos in pcre.3 and comments in the source fixed.
3979
3980 2. Applied a contributed patch to get rid of places where it used to remove
3981 'const' from variables, and fixed some signed/unsigned and uninitialized
3982 variable warnings.
3983
3984 3. Added the "runtest" target to Makefile.
3985
3986 4. Set default compiler flag to -O2 rather than just -O.
3987
3988
3989 Version 1.01 19-Nov-97
3990 ----------------------
3991
3992 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3993 like /([ab]*)*/, that is, for classes with more than one character in them.
3994
3995 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3996 as /((?>a*))*/ (a PCRE_EXTRA facility).
3997
3998
3999 Version 1.00 18-Nov-97
4000 ----------------------
4001
4002 1. Added compile-time macros to support systems such as SunOS4 which don't have
4003 memmove() or strerror() but have other things that can be used instead.
4004
4005 2. Arranged that "make clean" removes the executables.
4006
4007
4008 Version 0.99 27-Oct-97
4009 ----------------------
4010
4011 1. Fixed bug in code for optimizing classes with only one character. It was
4012 initializing a 32-byte map regardless, which could cause it to run off the end
4013 of the memory it had got.
4014
4015 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4016
4017
4018 Version 0.98 22-Oct-97
4019 ----------------------
4020
4021 1. Fixed bug in code for handling temporary memory usage when there are more
4022 back references than supplied space in the ovector. This could cause segfaults.
4023
4024
4025 Version 0.97 21-Oct-97
4026 ----------------------
4027
4028 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4029
4030 2. Optimized negated single characters not to use a bit map.
4031
4032 3. Brought error texts together as macro definitions; clarified some of them;
4033 fixed one that was wrong - it said "range out of order" when it meant "invalid
4034 escape sequence".
4035
4036 4. Changed some char * arguments to const char *.
4037
4038 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4039
4040 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4041 pcretest.
4042
4043
4044 Version 0.96 16-Oct-97
4045 ----------------------
4046
4047 1. Added a simple "pgrep" utility to the distribution.
4048
4049 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4050 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4051 where "ddd" means "one or more decimal digits".
4052
4053 3. Fixed serious bug. If a pattern had a back reference, but the call to
4054 pcre_exec() didn't supply a large enough ovector to record the related
4055 identifying subpattern, the match always failed. PCRE now remembers the number
4056 of the largest back reference, and gets some temporary memory in which to save
4057 the offsets during matching if necessary, in order to ensure that
4058 backreferences always work.
4059
4060 4. Increased the compatibility with Perl in a number of ways:
4061
4062 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4063 to request this handling. The option can be set at compile or exec time.
4064
4065 (b) $ matches before a terminating newline by default; an option
4066 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4067 mode). The option can be set at compile or exec time.
4068
4069 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4070 the same as Perl's. If the decimal number it represents is less than 10
4071 or there aren't that many previous left capturing parentheses, an octal
4072 escape is read. Inside a character class, it's always an octal escape,
4073 even if it is a single digit.
4074
4075 (d) An escaped but undefined alphabetic character is taken as a literal,
4076 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4077 escapes.
4078
4079 (e) {0} is now permitted. (The previous item is removed from the compiled
4080 pattern).
4081
4082 5. Changed all the names of code files so that the basic parts are no longer
4083 than 10 characters, and abolished the teeny "globals.c" file.
4084
4085 6. Changed the handling of character classes; they are now done with a 32-byte
4086 bit map always.
4087
4088 7. Added the -d and /D options to pcretest to make it possible to look at the
4089 internals of compilation without having to recompile pcre.
4090
4091
4092 Version 0.95 23-Sep-97
4093 ----------------------
4094
4095 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4096 \x20 at the start of a run of normal characters. These were being treated as
4097 real characters, instead of the source characters being re-checked.
4098
4099
4100 Version 0.94 18-Sep-97
4101 ----------------------
4102
4103 1. The functions are now thread-safe, with the caveat that the global variables
4104 containing pointers to malloc() and free() or alternative functions are the
4105 same for all threads.
4106
4107 2. Get pcre_study() to generate a bitmap of initial characters for non-
4108 anchored patterns when this is possible, and use it if passed to pcre_exec().
4109
4110
4111 Version 0.93 15-Sep-97
4112 ----------------------
4113
4114 1. /(b)|(:+)/ was computing an incorrect first character.
4115
4116 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4117 but not actually doing anything yet.
4118
4119 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4120 as Perl does (e.g. [-az] or [az-]).
4121
4122 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4123 all possible positions.
4124
4125 5. Split up into different modules to avoid including unneeded functions in a
4126 compiled binary. However, compile and exec are still in one module. The "study"
4127 function is split off.
4128
4129 6. The character tables are now in a separate module whose source is generated
4130 by an auxiliary program - but can then be edited by hand if required. There are
4131 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4132 toupper() in the code.
4133
4134 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4135 make them global. Abolish the function for setting them, as the caller can now
4136 set them directly.
4137
4138
4139 Version 0.92 11-Sep-97
4140 ----------------------
4141
4142 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4143 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4144
4145 2. Caseless matching was not working in character classes if the characters in
4146 the pattern were in upper case.
4147
4148 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4149
4150 4. Make PCRE_ANCHORED public and accept as a compile option.
4151
4152 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4153 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4154 pass them.
4155
4156 6. Give an error if bad option bits passed at compile or run time.
4157
4158 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4159 pcretest to cause it to pass that flag.
4160
4161 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4162 options, and the first character, if set.
4163
4164 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4165
4166
4167 Version 0.91 10-Sep-97
4168 ----------------------
4169
4170 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4171 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4172
4173 2. PCRE was looping on encountering an indefinitely repeated back reference to
4174 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4175 Perl does - treats the match as successful.
4176
4177 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5