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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 494 - (hide annotations) (download)
Tue Mar 2 11:53:12 2010 UTC (4 years, 4 months ago) by ph10
File size: 182068 byte(s)
If stdint.h is not available, include inttypes.h if available.

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