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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 491 - (hide annotations) (download)
Mon Mar 1 17:45:08 2010 UTC (4 years, 1 month ago) by ph10
File size: 181657 byte(s)
Update Unicode tables to Unicode version 5.2.0.

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

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