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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 503 - (show annotations) (download)
Sun Mar 7 17:35:52 2010 UTC (4 years, 6 months ago) by ph10
File size: 184429 byte(s)
Fix incorrect compile time error for certain types of recursive patterns.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12