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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 516 - (hide annotations) (download)
Tue May 4 15:51:35 2010 UTC (3 years, 11 months ago) by ph10
File size: 185927 byte(s)
Avoid compiler warning in pcretest.

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