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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 523 - (hide annotations) (download)
Sun May 23 18:48:54 2010 UTC (4 years, 5 months ago) by ph10
File size: 187464 byte(s)
Fix pcre_study() bug for \p{Nd}* type items at pattern start.

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