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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 511 - (hide annotations) (download)
Mon Mar 29 09:25:38 2010 UTC (4 years, 3 months ago) by ph10
File size: 184931 byte(s)
Make (*ACCEPT) work inside an atomic group.

1 nigel 3 ChangeLog for PCRE
2     ------------------
3    
4 ph10 510 Version 8.03 26-Mar-2010
5     ------------------------
6    
7     1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
8     THEN.
9 ph10 511
10     2. (*ACCEPT) was not working when inside an atomic group.
11 ph10 510
12    
13 ph10 508 Version 8.02 19-Mar-2010
14 ph10 491 ------------------------
15    
16     1. The Unicode data tables have been updated to Unicode 5.2.0.
17    
18 ph10 492 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
19     configured.
20 ph10 507
21 ph10 493 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
22 ph10 507 original author of that file, following a query about its status.
23    
24     4. On systems that do not have stdint.h (e.g. Solaris), check for and include
25 ph10 494 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
26 ph10 507
27 ph10 496 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
28     quantifier applied to a forward-referencing subroutine call, could compile
29     incorrect code or give the error "internal error: previously-checked
30     referenced subpattern not found".
31 ph10 507
32     6. Both MS Visual Studio and Symbian OS have problems with initializing
33 ph10 497 variables to point to external functions. For these systems, therefore,
34     pcre_malloc etc. are now initialized to local functions that call the
35     relevant global functions.
36 ph10 507
37 ph10 498 7. There were two entries missing in the vectors called coptable and poptable
38     in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
39 ph10 507 I've fixed the data, and added a kludgy way of testing at compile time that
40     the lengths are correct (equal to the number of opcodes).
41    
42     8. Following on from 7, I added a similar kludge to check the length of the
43     eint vector in pcreposix.c.
44    
45     9. Error texts for pcre_compile() are held as one long string to avoid too
46     much relocation at load time. To find a text, the string is searched,
47 ph10 499 counting zeros. There was no check for running off the end of the string,
48     which could happen if a new error number was added without updating the
49 ph10 507 string.
50    
51     10. \K gave a compile-time error if it appeared in a lookbehind assersion.
52    
53 ph10 500 11. \K was not working if it appeared in an atomic group or in a group that
54 ph10 507 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
55     \K is "not well defined" if used in an assertion. PCRE now accepts it if
56     the assertion is positive, but not if it is negative.
57    
58 ph10 501 12. Change 11 fortuitously reduced the size of the stack frame used in the
59 ph10 507 "match()" function of pcre_exec.c by one pointer. Forthcoming
60     implementation of support for (*MARK) will need an extra pointer on the
61 ph10 501 stack; I have reserved it now, so that the stack frame size does not
62     decrease.
63 ph10 507
64     13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
65     item in branch that calls a recursion is a subroutine call - as in the
66 ph10 503 second branch in the above example - was incorrectly given the compile-
67     time error "recursive call could loop indefinitely" because pcre_compile()
68 ph10 507 was not correctly checking the subroutine for matching a non-empty string.
69    
70 ph10 505 14. The checks for overrunning compiling workspace could trigger after an
71 ph10 507 overrun had occurred. This is a "should never occur" error, but it can be
72 ph10 505 triggered by pathological patterns such as hundreds of nested parentheses.
73 ph10 507 The checks now trigger 100 bytes before the end of the workspace.
74 ph10 491
75 ph10 507 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
76 ph10 492
77 ph10 507
78 ph10 489 Version 8.01 19-Jan-2010
79     ------------------------
80 ph10 471
81     1. If a pattern contained a conditional subpattern with only one branch (in
82 ph10 489 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
83 ph10 473 computed the wrong minimum data length (which is of course zero for such
84 ph10 489 subpatterns). This could cause incorrect "no match" results.
85 ph10 473
86     2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
87 ph10 472 the pattern is reset in the first branch, pcre_compile() failed with
88     "internal error: code overflow at offset...". This happened only when
89 ph10 473 the reset was to the original external option setting. (An optimization
90     abstracts leading options settings into an external setting, which was the
91 ph10 472 cause of this.)
92 ph10 471
93 ph10 473 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
94     of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
95     assertion pattern did not match (meaning that the assertion was true), it
96     was incorrectly treated as false if the SKIP had been reached during the
97     matching. This also applied to assertions used as conditions.
98 ph10 471
99 ph10 473 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
100     assertion subpattern, including such a pattern used as a condition,
101     unpredictable results occurred, instead of the error return
102     PCRE_ERROR_DFA_UITEM.
103 ph10 478
104 ph10 474 5. The C++ GlobalReplace function was not working like Perl for the special
105     situation when an empty string is matched. It now does the fancy magic
106 ph10 478 stuff that is necessary.
107    
108     6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
109 ph10 475 removed. (These were left over from very, very early versions of PCRE.)
110 ph10 478
111 ph10 475 7. Some cosmetic changes to the code to make life easier when compiling it
112     as part of something else:
113 ph10 478
114     (a) Change DEBUG to PCRE_DEBUG.
115    
116     (b) In pcre_compile(), rename the member of the "branch_chain" structure
117     called "current" as "current_branch", to prevent a collision with the
118 ph10 475 Linux macro when compiled as a kernel module.
119 ph10 478
120     (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
121     prevent a collision with the Linux macro when compiled as a kernel
122 ph10 475 module.
123 ph10 478
124 ph10 475 8. In pcre_compile() there are some checks for integer overflows that used to
125     cast potentially large values to (double). This has been changed to that
126 ph10 478 when building, a check for int64_t is made, and if it is found, it is used
127     instead, thus avoiding the use of floating point arithmetic. (There is no
128     other use of FP in PCRE.) If int64_t is not found, the fallback is to
129     double.
130    
131     9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
132 ph10 476 2005 (difference between two addresses compared to an unsigned value).
133 ph10 473
134 ph10 478 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
135     custom one, because of the following reported problem in Windows:
136 ph10 473
137 ph10 478 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
138 ph10 489 under Win32.
139 ph10 478 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
140 ph10 489 therefore missing the function definition.
141 ph10 478 - The compiler thus generates a "C" signature for the test function.
142     - The linker fails to find the "C" function.
143     - PCRE fails to configure if asked to do so against libbz2.
144 ph10 486
145 ph10 479 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
146     messages were output:
147 ph10 478
148 ph10 479 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
149     rerunning libtoolize, to keep the correct libtool macros in-tree.
150     Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
151 ph10 486
152 ph10 479 I have done both of these things.
153 ph10 486
154 ph10 480 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
155 ph10 486 most of the time, it *can* run out if it is given a pattern that contains a
156     runaway infinite recursion. I updated the discussion in the pcrestack man
157     page.
158    
159     13. Now that we have gone to the x.xx style of version numbers, the minor
160     version may start with zero. Using 08 or 09 is a bad idea because users
161     might check the value of PCRE_MINOR in their code, and 08 or 09 may be
162     interpreted as invalid octal numbers. I've updated the previous comment in
163     configure.ac, and also added a check that gives an error if 08 or 09 are
164 ph10 481 used.
165 ph10 486
166 ph10 482 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
167 ph10 483 causing partial matching to fail when the end of the subject matched \W
168     in a UTF-8 pattern where \W was quantified with a minimum of 3.
169 ph10 486
170 ph10 483 15. There were some discrepancies between the declarations in pcre_internal.h
171 ph10 486 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
172 ph10 483 their definitions. The declarations used "const uschar *" and the
173     definitions used USPTR. Even though USPTR is normally defined as "const
174     unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
175     reported that: "This difference in casting confuses some C++ compilers, for
176     example, SunCC recognizes above declarations as different functions and
177 ph10 486 generates broken code for hbpcre." I have changed the declarations to use
178 ph10 483 USPTR.
179 ph10 486
180     16. GNU libtool is named differently on some systems. The autogen.sh script now
181 ph10 484 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
182     (FreeBSD).
183 ph10 478
184 ph10 486 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
185     (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
186     comment: "Figure out how to create a longlong from a string: strtoll and
187     equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
188     instance, but it only takes 2 args instead of 3!"
189 ph10 479
190 ph10 488 18. A subtle bug concerned with back references has been fixed by a change of
191     specification, with a corresponding code fix. A pattern such as
192     ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
193     refers, was giving matches when it shouldn't. For example, xa=xaaa would
194     match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
195     same bug. Such groups have to be quantified to be useful, or contained
196     inside another quantified group. (If there's no repetition, the reference
197     can never match.) The problem arises because, having left the group and
198     moved on to the rest of the pattern, a later failure that backtracks into
199     the group uses the captured value from the final iteration of the group
200     rather than the correct earlier one. I have fixed this in PCRE by forcing
201     any group that contains a reference to itself to be an atomic group; that
202     is, there cannot be any backtracking into it once it has completed. This is
203     similar to recursive and subroutine calls.
204 ph10 486
205 ph10 488
206 ph10 469 Version 8.00 19-Oct-09
207 ph10 418 ----------------------
208    
209     1. The table for translating pcre_compile() error codes into POSIX error codes
210 ph10 461 was out-of-date, and there was no check on the pcre_compile() error code
211     being within the table. This could lead to an OK return being given in
212 ph10 418 error.
213 ph10 461
214     2. Changed the call to open a subject file in pcregrep from fopen(pathname,
215     "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
216     in a Windows environment.
217    
218 ph10 420 3. The pcregrep --count option prints the count for each file even when it is
219     zero, as does GNU grep. However, pcregrep was also printing all files when
220     --files-with-matches was added. Now, when both options are given, it prints
221     counts only for those files that have at least one match. (GNU grep just
222 ph10 461 prints the file name in this circumstance, but including the count seems
223     more useful - otherwise, why use --count?) Also ensured that the
224 ph10 420 combination -clh just lists non-zero counts, with no names.
225 ph10 461
226     4. The long form of the pcregrep -F option was incorrectly implemented as
227     --fixed_strings instead of --fixed-strings. This is an incompatible change,
228     but it seems right to fix it, and I didn't think it was worth preserving
229     the old behaviour.
230    
231     5. The command line items --regex=pattern and --regexp=pattern were not
232 ph10 422 recognized by pcregrep, which required --regex pattern or --regexp pattern
233 ph10 461 (with a space rather than an '='). The man page documented the '=' forms,
234 ph10 422 which are compatible with GNU grep; these now work.
235 ph10 461
236     6. No libpcreposix.pc file was created for pkg-config; there was just
237 ph10 423 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
238 ph10 461
239 ph10 425 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
240 ph10 461 when UCP support is not needed, by modifying the Python script that
241 ph10 425 generates it from Unicode data files. This should not matter if the module
242     is correctly used as a library, but I received one complaint about 50K of
243     unwanted data. My guess is that the person linked everything into his
244     program rather than using a library. Anyway, it does no harm.
245 ph10 461
246 ph10 426 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
247 ph10 461 was a minimum greater than 1 for a wide character in a possessive
248 ph10 431 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
249     which had an unlimited repeat of a nested, fixed maximum repeat of a wide
250     character. Chaos in the form of incorrect output or a compiling loop could
251     result.
252 ph10 461
253 ph10 426 9. The restrictions on what a pattern can contain when partial matching is
254 ph10 461 requested for pcre_exec() have been removed. All patterns can now be
255 ph10 426 partially matched by this function. In addition, if there are at least two
256 ph10 435 slots in the offset vector, the offset of the earliest inspected character
257     for the match and the offset of the end of the subject are set in them when
258 ph10 461 PCRE_ERROR_PARTIAL is returned.
259    
260 ph10 428 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
261     synonymous with PCRE_PARTIAL, for backwards compatibility, and
262     PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
263 ph10 462 and may be more useful for multi-segment matching.
264 ph10 461
265     11. Partial matching with pcre_exec() is now more intuitive. A partial match
266     used to be given if ever the end of the subject was reached; now it is
267     given only if matching could not proceed because another character was
268     needed. This makes a difference in some odd cases such as Z(*FAIL) with the
269     string "Z", which now yields "no match" instead of "partial match". In the
270     case of pcre_dfa_exec(), "no match" is given if every matching path for the
271     final character ended with (*FAIL).
272    
273 ph10 428 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
274 ph10 461 if the pattern had a "must contain" character that was already found in the
275 ph10 428 earlier partial match, unless partial matching was again requested. For
276     example, with the pattern /dog.(body)?/, the "must contain" character is
277     "g". If the first part-match was for the string "dog", restarting with
278 ph10 435 "sbody" failed. This bug has been fixed.
279 ph10 461
280     13. The string returned by pcre_dfa_exec() after a partial match has been
281     changed so that it starts at the first inspected character rather than the
282     first character of the match. This makes a difference only if the pattern
283     starts with a lookbehind assertion or \b or \B (\K is not supported by
284     pcre_dfa_exec()). It's an incompatible change, but it makes the two
285 ph10 435 matching functions compatible, and I think it's the right thing to do.
286 ph10 461
287 ph10 435 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
288 ph10 461 so that the demonstration program is easily available in environments where
289     PCRE has not been installed from source.
290    
291 ph10 435 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
292 ph10 430 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
293     library.
294 ph10 461
295 ph10 435 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
296 ph10 432 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
297 ph10 461 is not the first non-POSIX option to be added. Clearly some people find
298 ph10 433 these options useful.
299 ph10 461
300     17. If a caller to the POSIX matching function regexec() passes a non-zero
301 ph10 438 value for nmatch with a NULL value for pmatch, the value of
302 ph10 461 nmatch is forced to zero.
303    
304 ph10 437 18. RunGrepTest did not have a test for the availability of the -u option of
305 ph10 461 the diff command, as RunTest does. It now checks in the same way as
306 ph10 437 RunTest, and also checks for the -b option.
307 ph10 461
308 ph10 438 19. If an odd number of negated classes containing just a single character
309     interposed, within parentheses, between a forward reference to a named
310 ph10 461 subpattern and the definition of the subpattern, compilation crashed with
311     an internal error, complaining that it could not find the referenced
312 ph10 438 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
313 ph10 461 [The bug was that it was starting one character too far in when skipping
314     over the character class, thus treating the ] as data rather than
315     terminating the class. This meant it could skip too much.]
316    
317 ph10 442 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
318 ph10 461 /g option in pcretest when the pattern contains \K, which makes it possible
319 ph10 442 to have an empty string match not at the start, even when the pattern is
320 ph10 461 anchored. Updated pcretest and pcredemo to use this option.
321    
322 ph10 446 21. If the maximum number of capturing subpatterns in a recursion was greater
323 ph10 461 than the maximum at the outer level, the higher number was returned, but
324     with unset values at the outer level. The correct (outer level) value is
325 ph10 446 now given.
326 ph10 461
327 ph10 447 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
328     PCRE did not set those parentheses (unlike Perl). I have now found a way to
329     make it do so. The string so far is captured, making this feature
330     compatible with Perl.
331 ph10 461
332     23. The tests have been re-organized, adding tests 11 and 12, to make it
333 ph10 448 possible to check the Perl 5.10 features against Perl 5.10.
334 ph10 461
335 ph10 454 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
336 ph10 461 pattern matches a fixed length string. PCRE did not allow this; now it
337     does. Neither allows recursion.
338    
339     25. I finally figured out how to implement a request to provide the minimum
340     length of subject string that was needed in order to match a given pattern.
341     (It was back references and recursion that I had previously got hung up
342     on.) This code has now been added to pcre_study(); it finds a lower bound
343 ph10 455 to the length of subject needed. It is not necessarily the greatest lower
344     bound, but using it to avoid searching strings that are too short does give
345     some useful speed-ups. The value is available to calling programs via
346     pcre_fullinfo().
347 ph10 461
348 ph10 455 26. While implementing 25, I discovered to my embarrassment that pcretest had
349     not been passing the result of pcre_study() to pcre_dfa_exec(), so the
350     study optimizations had never been tested with that matching function.
351     Oops. What is worse, even when it was passed study data, there was a bug in
352     pcre_dfa_exec() that meant it never actually used it. Double oops. There
353     were also very few tests of studied patterns with pcre_dfa_exec().
354 ph10 461
355 ph10 457 27. If (?| is used to create subpatterns with duplicate numbers, they are now
356     allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
357     on the other side of the coin, they are no longer allowed to have different
358     names, because these cannot be distinguished in PCRE, and this has caused
359     confusion. (This is a difference from Perl.)
360 ph10 461
361     28. When duplicate subpattern names are present (necessarily with different
362     numbers, as required by 27 above), and a test is made by name in a
363     conditional pattern, either for a subpattern having been matched, or for
364     recursion in such a pattern, all the associated numbered subpatterns are
365 ph10 459 tested, and the overall condition is true if the condition is true for any
366     one of them. This is the way Perl works, and is also more like the way
367     testing by number works.
368 ph10 418
369 ph10 461
370 ph10 412 Version 7.9 11-Apr-09
371 ph10 376 ---------------------
372    
373 ph10 392 1. When building with support for bzlib/zlib (pcregrep) and/or readline
374 ph10 376 (pcretest), all targets were linked against these libraries. This included
375     libpcre, libpcreposix, and libpcrecpp, even though they do not use these
376     libraries. This caused unwanted dependencies to be created. This problem
377 ph10 392 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
378 ph10 376 pcretest is linked with readline.
379 ph10 392
380 ph10 376 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
381     "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
382     moved outside it again, because FALSE and TRUE are already defined in AIX,
383     but BOOL is not.
384 ph10 392
385     3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
386 ph10 409 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
387 ph10 392
388 ph10 376 4. The pcregrep documentation said spaces were inserted as well as colons (or
389     hyphens) following file names and line numbers when outputting matching
390 ph10 392 lines. This is not true; no spaces are inserted. I have also clarified the
391 ph10 376 wording for the --colour (or --color) option.
392 ph10 392
393 ph10 378 5. In pcregrep, when --colour was used with -o, the list of matching strings
394     was not coloured; this is different to GNU grep, so I have changed it to be
395     the same.
396 ph10 392
397 ph10 378 6. When --colo(u)r was used in pcregrep, only the first matching substring in
398 ph10 392 each matching line was coloured. Now it goes on to look for further matches
399     of any of the test patterns, which is the same behaviour as GNU grep.
400    
401     7. A pattern that could match an empty string could cause pcregrep to loop; it
402     doesn't make sense to accept an empty string match in pcregrep, so I have
403 ph10 379 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
404     seems to be how GNU grep behaves.
405 ph10 392
406 ph10 380 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
407 ph10 392 start or after a newline", because the conditional assertion was not being
408     correctly handled. The rule now is that both the assertion and what follows
409     in the first alternative must satisfy the test.
410    
411 ph10 405 9. If auto-callout was enabled in a pattern with a conditional group whose
412 ph10 399 condition was an assertion, PCRE could crash during matching, both with
413     pcre_exec() and pcre_dfa_exec().
414 ph10 392
415     10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
416     used for matching.
417    
418     11. Unicode property support in character classes was not working for
419 ph10 385 characters (bytes) greater than 127 when not in UTF-8 mode.
420 ph10 388
421 ph10 392 12. Added the -M command line option to pcretest.
422    
423 ph10 388 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
424 ph10 389
425     15. Added the PCRE_NO_START_OPTIMIZE match-time option.
426 ph10 390
427 ph10 392 16. Added comments and documentation about mis-use of no_arg in the C++
428 ph10 390 wrapper.
429 ph10 392
430 ph10 391 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
431 ph10 392 from Martin Jerabek that uses macro names for all relevant character and
432 ph10 391 string constants.
433 ph10 405
434 ph10 393 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
435     SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
436 ph10 405 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
437     these, but not everybody uses configure.
438    
439     19. A conditional group that had only one branch was not being correctly
440     recognized as an item that could match an empty string. This meant that an
441     enclosing group might also not be so recognized, causing infinite looping
442     (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
443     with the subject "ab", where knowledge that the repeated group can match
444 ph10 396 nothing is needed in order to break the loop.
445 ph10 405
446 ph10 397 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
447 ph10 405 exec(), but without supplying a callout function, matching went wrong.
448    
449     21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
450     leak if the size of the offset vector was greater than 30. When the vector
451     is smaller, the saved offsets during recursion go onto a local stack
452     vector, but for larger vectors malloc() is used. It was failing to free
453     when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
454 ph10 400 error, in fact).
455 ph10 405
456 ph10 402 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
457 ph10 405 heapframe that is used only when UTF-8 support is enabled. This caused no
458     problem, but was untidy.
459 ph10 376
460 ph10 405 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
461     CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
462     included within another project.
463 ph10 411
464 ph10 407 24. Steven Van Ingelgem's patches to add more options to the CMake support,
465     slightly modified by me:
466 ph10 411
467     (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
468     not building pcregrep.
469    
470 ph10 407 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
471 ph10 411 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
472    
473 ph10 408 25. Forward references, both numeric and by name, in patterns that made use of
474     duplicate group numbers, could behave incorrectly or give incorrect errors,
475     because when scanning forward to find the reference group, PCRE was not
476     taking into account the duplicate group numbers. A pattern such as
477     ^X(?3)(a)(?|(b)|(q))(Y) is an example.
478 ph10 411
479     26. Changed a few more instances of "const unsigned char *" to USPTR, making
480     the feature of a custom pointer more persuasive (as requested by a user).
481 ph10 416
482 ph10 411 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
483 ph10 416 pcretest.c, inside #ifndefs, because it seems they are sometimes already
484     pre-defined.
485 ph10 392
486 ph10 416 28. Added support for (*UTF8) at the start of a pattern.
487    
488 ph10 413 29. Arrange for flags added by the "release type" setting in CMake to be shown
489     in the configuration summary.
490 ph10 405
491 ph10 411
492 ph10 374 Version 7.8 05-Sep-08
493 ph10 349 ---------------------
494    
495     1. Replaced UCP searching code with optimized version as implemented for Ad
496     Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
497 ph10 351 stage table and inline lookup instead of a function, giving speed ups of 2
498     to 5 times on some simple patterns that I tested. Permission was given to
499     distribute the MultiStage2.py script that generates the tables (it's not in
500     the tarball, but is in the Subversion repository).
501 ph10 349
502 ph10 351 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
503     scripts.
504 ph10 358
505 ph10 353 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
506     a group with a zero qualifier. The result of the study could be incorrect,
507 ph10 358 or the function might crash, depending on the pattern.
508    
509     4. Caseless matching was not working for non-ASCII characters in back
510 ph10 354 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
511 ph10 358 It now works when Unicode Property Support is available.
512    
513 ph10 355 5. In pcretest, an escape such as \x{de} in the data was always generating
514     a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
515     non-UTF-8 mode. If the value is greater than 255, it gives a warning about
516 ph10 358 truncation.
517 ph10 349
518 ph10 358 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
519    
520 ph10 357 7. Added two (int) casts to pcregrep when printing the difference of two
521     pointers, in case they are 64-bit values.
522 ph10 371
523     8. Added comments about Mac OS X stack usage to the pcrestack man page and to
524     test 2 if it fails.
525    
526 ph10 359 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
527 ph10 371 and a #define of that name to empty if it is not externally set. This is to
528     allow users of MSVC to set it if necessary.
529    
530     10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
531 ph10 359 the convenience functions in the pcre_get.c source file.
532 ph10 371
533 ph10 360 11. An option change at the start of a pattern that had top-level alternatives
534     could cause overwriting and/or a crash. This command provoked a crash in
535 ph10 371 some environments:
536    
537     printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
538    
539 ph10 360 This potential security problem was recorded as CVE-2008-2371.
540 ph10 371
541 ph10 361 12. For a pattern where the match had to start at the beginning or immediately
542     after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
543 ph10 371 pcre_dfa_exec() could read past the end of the passed subject if there was
544 ph10 361 no match. To help with detecting such bugs (e.g. with valgrind), I modified
545     pcretest so that it places the subject at the end of its malloc-ed buffer.
546 ph10 371
547 ph10 364 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
548 ph10 371 exec() might read past the end of the data buffer in UTF-8 mode.
549    
550 ph10 364 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
551 ph10 371 the data contained the byte 0x85 as part of a UTF-8 character within its
552     first line. This applied both to normal and DFA matching.
553    
554 ph10 366 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
555 ph10 371 /^[^d]*?$/8 failed to match "abc".
556 ph10 351
557 ph10 371 16. Added a missing copyright notice to pcrecpp_internal.h.
558    
559     17. Make it more clear in the documentation that values returned from
560 ph10 368 pcre_exec() in ovector are byte offsets, not character counts.
561 ph10 357
562 ph10 371 18. Tidied a few places to stop certain compilers from issuing warnings.
563 ph10 368
564 ph10 373 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
565     supplied by Stefan Weber. I made a further small update for 7.8 because
566 ph10 374 there is a change of source arrangements: the pcre_searchfuncs.c module is
567     replaced by pcre_ucd.c.
568 ph10 371
569 ph10 373
570 ph10 347 Version 7.7 07-May-08
571 ph10 321 ---------------------
572    
573     1. Applied Craig's patch to sort out a long long problem: "If we can't convert
574 ph10 345 a string to a long long, pretend we don't even have a long long." This is
575 ph10 321 done by checking for the strtoq, strtoll, and _strtoi64 functions.
576 ph10 345
577 ph10 322 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
578     pre-7.6 versions, which defined a global no_arg variable instead of putting
579 ph10 332 it in the RE class. (See also #8 below.)
580 ph10 345
581     3. Remove a line of dead code, identified by coverity and reported by Nuno
582     Lopes.
583    
584 ph10 324 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
585 ph10 321
586 ph10 324 (1) The include/exclude patterns were being applied to the whole pathnames
587 ph10 345 of files, instead of just to the final components.
588    
589 ph10 324 (2) If there was more than one level of directory, the subdirectories were
590     skipped unless they satisfied the include/exclude conditions. This is
591     inconsistent with GNU grep (and could even be seen as contrary to the
592     pcregrep specification - which I improved to make it absolutely clear).
593     The action now is always to scan all levels of directory, and just
594     apply the include/exclude patterns to regular files.
595 ph10 345
596 ph10 325 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
597 ph10 345 --exclude_dir in the tests to avoid scanning .svn directories.
598    
599 ph10 326 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
600 ph10 345 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
601 ph10 326 doesn't support NULs in patterns.
602 ph10 345
603     7. Added some missing "const"s to declarations of static tables in
604     pcre_compile.c and pcre_dfa_exec.c.
605    
606 ph10 328 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
607 ph10 329 caused by fix #2 above. (Subsequently also a second patch to fix the
608 ph10 345 first patch. And a third patch - this was a messy problem.)
609 ph10 321
610 ph10 345 9. Applied Craig's patch to remove the use of push_back().
611    
612 ph10 332 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
613     matching function regexec().
614 ph10 345
615     11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
616     which, however, unlike Perl's \g{...}, are subroutine calls, not back
617 ph10 333 references. PCRE supports relative numbers with this syntax (I don't think
618     Oniguruma does).
619 ph10 345
620     12. Previously, a group with a zero repeat such as (...){0} was completely
621 ph10 335 omitted from the compiled regex. However, this means that if the group
622     was called as a subroutine from elsewhere in the pattern, things went wrong
623 ph10 345 (an internal error was given). Such groups are now left in the compiled
624     pattern, with a new opcode that causes them to be skipped at execution
625 ph10 335 time.
626 ph10 345
627 ph10 341 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
628     to the way PCRE behaves:
629 ph10 345
630 ph10 336 (a) A lone ] character is dis-allowed (Perl treats it as data).
631 ph10 345
632     (b) A back reference to an unmatched subpattern matches an empty string
633 ph10 336 (Perl fails the current match path).
634 ph10 345
635 ph10 341 (c) A data ] in a character class must be notated as \] because if the
636 ph10 345 first data character in a class is ], it defines an empty class. (In
637 ph10 341 Perl it is not possible to have an empty class.) The empty class []
638     never matches; it forces failure and is equivalent to (*FAIL) or (?!).
639     The negative empty class [^] matches any one character, independently
640     of the DOTALL setting.
641 ph10 345
642     14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
643 ph10 340 non-existent subpattern following a character class starting with ']' and
644     containing () gave an internal compiling error instead of "reference to
645     non-existent subpattern". Fortunately, when the pattern did exist, the
646 ph10 345 compiled code was correct. (When scanning forwards to check for the
647 ph10 340 existencd of the subpattern, it was treating the data ']' as terminating
648 ph10 345 the class, so got the count wrong. When actually compiling, the reference
649 ph10 340 was subsequently set up correctly.)
650 ph10 345
651 ph10 341 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
652 ph10 345 it was being rejected as not supported by pcre_dfa_exec(), even though
653     other assertions are supported. I have made pcre_dfa_exec() support
654     (*FAIL).
655    
656 ph10 342 16. The implementation of 13c above involved the invention of a new opcode,
657 ph10 345 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
658     cannot be changed at match time, I realized I could make a small
659     improvement to matching performance by compiling OP_ALLANY instead of
660     OP_ANY for "." when DOTALL was set, and then removing the runtime tests
661     on the OP_ANY path.
662    
663     17. Compiling pcretest on Windows with readline support failed without the
664 ph10 343 following two fixes: (1) Make the unistd.h include conditional on
665 ph10 345 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
666    
667 ph10 344 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
668 ph10 345 ncurses library to be included for pcretest when ReadLine support is
669     requested, but also to allow for it to be overridden. This patch came from
670     Daniel Bergström.
671 ph10 347
672 ph10 346 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
673 ph10 347 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
674     any errors with the current Unicode tables. Thanks to Peter Kankowski for
675     spotting this.
676 ph10 324
677 ph10 332
678 ph10 319 Version 7.6 28-Jan-08
679 ph10 300 ---------------------
680    
681 ph10 302 1. A character class containing a very large number of characters with
682     codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
683     overflow.
684 ph10 309
685     2. Patch to cut out the "long long" test in pcrecpp_unittest when
686     HAVE_LONG_LONG is not defined.
687    
688 ph10 303 3. Applied Christian Ehrlicher's patch to update the CMake build files to
689 ph10 304 bring them up to date and include new features. This patch includes:
690 ph10 309
691 ph10 304 - Fixed PH's badly added libz and libbz2 support.
692     - Fixed a problem with static linking.
693 ph10 312 - Added pcredemo. [But later removed - see 7 below.]
694 ph10 304 - Fixed dftables problem and added an option.
695     - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
696     HAVE_LONG_LONG.
697     - Added readline support for pcretest.
698 ph10 309 - Added an listing of the option settings after cmake has run.
699    
700 ph10 314 4. A user submitted a patch to Makefile that makes it easy to create
701     "pcre.dll" under mingw when using Configure/Make. I added stuff to
702     Makefile.am that cause it to include this special target, without
703     affecting anything else. Note that the same mingw target plus all
704     the other distribution libraries and programs are now supported
705     when configuring with CMake (see 6 below) instead of with
706     Configure/Make.
707 ph10 309
708 ph10 308 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
709     This is an attempt to solve the reported problem "pcrecpp::no_arg is not
710 ph10 309 exported in the Windows port". It has not yet been confirmed that the patch
711     solves the problem, but it does no harm.
712 ph10 313
713 ph10 311 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
714 ph10 319 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
715 ph10 314 with CMake, and also correct the comment about stack recursion.
716 ph10 313
717 ph10 312 7. Remove the automatic building of pcredemo from the ./configure system and
718 ph10 313 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
719     of a program that users should build themselves after PCRE is installed, so
720     building it automatically is not really right. What is more, it gave
721 ph10 312 trouble in some build environments.
722 ph10 300
723 ph10 319 8. Further tidies to CMakeLists.txt from Sheri and Christian.
724 ph10 308
725 ph10 319
726 ph10 298 Version 7.5 10-Jan-08
727 ph10 263 ---------------------
728    
729     1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
730     values in parens when parsing an RE using the C++ wrapper."
731 ph10 286
732 ph10 264 2. Negative specials like \S did not work in character classes in UTF-8 mode.
733     Characters greater than 255 were excluded from the class instead of being
734     included.
735 ph10 286
736     3. The same bug as (2) above applied to negated POSIX classes such as
737 ph10 265 [:^space:].
738 ph10 286
739 ph10 267 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
740     defined or documented. It seems to have been a typo for PCRE_STATIC, so
741 ph10 286 I have changed it.
742    
743 ph10 268 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
744 ph10 286 first named subpattern) and a construct such as (?&a) would reference the
745     first named subpattern whose name started with "a" (in other words, the
746 ph10 272 length check was missing). Both these problems are fixed. "Subpattern name
747     expected" is now given for (?&) (a zero-length name), and this patch also
748     makes it give the same error for \k'' (previously it complained that that
749     was a reference to a non-existent subpattern).
750 ph10 286
751 ph10 269 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
752 ph10 286 this is right because (?- can be followed by option settings as well as by
753 ph10 269 digits. I have, however, made the messages clearer.
754 ph10 286
755 ph10 270 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
756     than the number used in the conditional) now cause a compile-time error.
757     This is actually not compatible with Perl, which accepts such patterns, but
758     treats the conditional as always being FALSE (as PCRE used to), but it
759     seems to me that giving a diagnostic is better.
760 ph10 286
761 ph10 274 8. Change "alphameric" to the more common word "alphanumeric" in comments
762 ph10 275 and messages.
763 ph10 286
764     9. Fix two occurrences of "backslash" in comments that should have been
765     "backspace".
766    
767     10. Remove two redundant lines of code that can never be obeyed (their function
768     was moved elsewhere).
769    
770     11. The program that makes PCRE's Unicode character property table had a bug
771     which caused it to generate incorrect table entries for sequences of
772 ph10 277 characters that have the same character type, but are in different scripts.
773 ph10 286 It amalgamated them into a single range, with the script of the first of
774 ph10 277 them. In other words, some characters were in the wrong script. There were
775     thirteen such cases, affecting characters in the following ranges:
776 ph10 286
777 ph10 277 U+002b0 - U+002c1
778     U+0060c - U+0060d
779 ph10 286 U+0061e - U+00612
780 ph10 277 U+0064b - U+0065e
781     U+0074d - U+0076d
782     U+01800 - U+01805
783     U+01d00 - U+01d77
784     U+01d9b - U+01dbf
785     U+0200b - U+0200f
786     U+030fc - U+030fe
787     U+03260 - U+0327f
788     U+0fb46 - U+0fbb1
789     U+10450 - U+1049d
790 ph10 286
791 ph10 279 12. The -o option (show only the matching part of a line) for pcregrep was not
792     compatible with GNU grep in that, if there was more than one match in a
793     line, it showed only the first of them. It now behaves in the same way as
794     GNU grep.
795 ph10 286
796     13. If the -o and -v options were combined for pcregrep, it printed a blank
797     line for every non-matching line. GNU grep prints nothing, and pcregrep now
798     does the same. The return code can be used to tell if there were any
799     non-matching lines.
800    
801 ph10 289 14. Added --file-offsets and --line-offsets to pcregrep.
802    
803     15. The pattern (?=something)(?R) was not being diagnosed as a potentially
804 ph10 282 infinitely looping recursion. The bug was that positive lookaheads were not
805 ph10 286 being skipped when checking for a possible empty match (negative lookaheads
806     and both kinds of lookbehind were skipped).
807    
808 ph10 289 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
809 ph10 284 inclusion of <windows.h> to before rather than after the definition of
810     INVALID_FILE_ATTRIBUTES (patch from David Byron).
811 ph10 263
812 ph10 289 17. Specifying a possessive quantifier with a specific limit for a Unicode
813 ph10 286 character property caused pcre_compile() to compile bad code, which led at
814     runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
815     are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
816     caused the error; without that there was no problem.
817 ph10 263
818 ph10 289 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
819 ph10 286
820 ph10 289 19. Added --enable-pcretest-libreadline.
821 ph10 286
822 ph10 289 20. In pcrecpp.cc, the variable 'count' was incremented twice in
823 ph10 288 RE::GlobalReplace(). As a result, the number of replacements returned was
824 ph10 298 double what it should be. I removed one of the increments, but Craig sent a
825     later patch that removed the other one (the right fix) and added unit tests
826     that check the return values (which was not done before).
827 ph10 292
828 ph10 291 21. Several CMake things:
829 ph10 287
830 ph10 291 (1) Arranged that, when cmake is used on Unix, the libraries end up with
831 ph10 292 the names libpcre and libpcreposix, not just pcre and pcreposix.
832 ph10 288
833 ph10 292 (2) The above change means that pcretest and pcregrep are now correctly
834     linked with the newly-built libraries, not previously installed ones.
835 ph10 291
836 ph10 292 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
837 ph10 298
838     22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
839     crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
840     UTF-8 newline character). The key issue is that the pattern starts .*;
841     this means that the match must be either at the beginning, or after a
842     newline. The bug was in the code for advancing after a failed match and
843     checking that the new position followed a newline. It was not taking
844 ph10 294 account of UTF-8 characters correctly.
845 ph10 298
846     23. PCRE was behaving differently from Perl in the way it recognized POSIX
847     character classes. PCRE was not treating the sequence [:...:] as a
848     character class unless the ... were all letters. Perl, however, seems to
849     allow any characters between [: and :], though of course it rejects as
850     unknown any "names" that contain non-letters, because all the known class
851     names consist only of letters. Thus, Perl gives an error for [[:1234:]],
852     for example, whereas PCRE did not - it did not recognize a POSIX character
853     class. This seemed a bit dangerous, so the code has been changed to be
854     closer to Perl. The behaviour is not identical to Perl, because PCRE will
855     diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
856     treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
857     Perl does, and where it didn't before.
858    
859 ph10 296 24. Rewrite so as to remove the single use of %n from pcregrep because in some
860     Windows environments %n is disabled by default.
861 ph10 292
862    
863 ph10 260 Version 7.4 21-Sep-07
864 ph10 230 ---------------------
865    
866     1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
867 ph10 231 means that a class such as [\s] counted as "explicit reference to CR or
868     LF". That isn't really right - the whole point of the change was to try to
869     help when there was an actual mention of one of the two characters. So now
870     the change happens only if \r or \n (or a literal CR or LF) character is
871 ph10 230 encountered.
872 ph10 231
873     2. The 32-bit options word was also used for 6 internal flags, but the numbers
874     of both had grown to the point where there were only 3 bits left.
875     Fortunately, there was spare space in the data structure, and so I have
876     moved the internal flags into a new 16-bit field to free up more option
877 ph10 230 bits.
878 ph10 231
879     3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
880     but did not set the internal JCHANGED flag - either of these is enough to
881     control the way the "get" function works - but the PCRE_INFO_JCHANGED
882     facility is supposed to tell if (?J) was ever used, so now (?J) at the
883 ph10 230 start sets both bits.
884    
885 ph10 231 4. Added options (at build time, compile time, exec time) to change \R from
886     matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
887 ph10 230
888 ph10 243 5. doc/pcresyntax.html was missing from the distribution.
889    
890     6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
891 ph10 233 compatibility, even though it is no longer used.
892 ph10 243
893     7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
894 ph10 254 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
895 ph10 259 windows.h file is present (where different names are used). [This was
896 ph10 257 reversed later after testing - see 16 below.]
897 ph10 243
898     8. Changed all #include <config.h> to #include "config.h". There were also
899 ph10 236 some further <pcre.h> cases that I changed to "pcre.h".
900 ph10 243
901 ph10 239 9. When pcregrep was used with the --colour option, it missed the line ending
902 ph10 243 sequence off the lines that it output.
903    
904     10. It was pointed out to me that arrays of string pointers cause lots of
905     relocations when a shared library is dynamically loaded. A technique of
906     using a single long string with a table of offsets can drastically reduce
907     these. I have refactored PCRE in four places to do this. The result is
908 ph10 240 dramatic:
909 ph10 243
910 ph10 240 Originally: 290
911     After changing UCP table: 187
912 ph10 243 After changing error message table: 43
913 ph10 240 After changing table of "verbs" 36
914     After changing table of Posix names 22
915 ph10 243
916 ph10 240 Thanks to the folks working on Gregex for glib for this insight.
917 ph10 247
918 ph10 244 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
919 ph10 247 unicode-properties was also set.
920    
921 ph10 248 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
922 ph10 231
923 ph10 253 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
924     checked only for CRLF.
925 ph10 233
926 ph10 259 14. Added casts to pcretest.c to avoid compiler warnings.
927    
928 ph10 256 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
929 ph10 253
930 ph10 257 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
931 ph10 259 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
932     entirely. This removes changes made in 7 above.
933 ph10 256
934 ph10 261 17. The CMake files have been updated, and there is now more information about
935     building with CMake in the NON-UNIX-USE document.
936 ph10 257
937 ph10 261
938 ph10 228 Version 7.3 28-Aug-07
939 ph10 157 ---------------------
940    
941 ph10 189 1. In the rejigging of the build system that eventually resulted in 7.1, the
942     line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
943     brackets there is not right, since it causes compilers to look for an
944     installed pcre.h, not the version that is in the source that is being
945     compiled (which of course may be different). I have changed it back to:
946 ph10 197
947 ph10 189 #include "pcre.h"
948 ph10 197
949     I have a vague recollection that the change was concerned with compiling in
950     different directories, but in the new build system, that is taken care of
951     by the VPATH setting the Makefile.
952    
953 ph10 190 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
954     when the subject happened to end in the byte 0x85 (e.g. if the last
955     character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
956     characters but of course it shouldn't be taken as a newline when it is part
957     of another character. The bug was that, for an unlimited repeat of . in
958     not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
959     characters when looking for a newline.
960 ph10 197
961     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
962    
963     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
964     in debug output.
965    
966 ph10 201 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
967     long printing in the pcrecpp unittest when running under MinGW.
968 ph10 189
969 ph10 197 6. ESC_K was left out of the EBCDIC table.
970 ph10 189
971 ph10 197 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
972     parentheses; I made it 1000, which seemed large enough. Unfortunately, the
973     limit also applies to "virtual nesting" when a pattern is recursive, and in
974     this case 1000 isn't so big. I have been able to remove this limit at the
975     expense of backing off one optimization in certain circumstances. Normally,
976     when pcre_exec() would call its internal match() function recursively and
977     immediately return the result unconditionally, it uses a "tail recursion"
978     feature to save stack. However, when a subpattern that can match an empty
979     string has an unlimited repetition quantifier, it no longer makes this
980     optimization. That gives it a stack frame in which to save the data for
981     checking that an empty string has been matched. Previously this was taken
982     from the 1000-entry workspace that had been reserved. So now there is no
983     explicit limit, but more stack is used.
984 ph10 189
985 ph10 200 8. Applied Daniel's patches to solve problems with the import/export magic
986     syntax that is required for Windows, and which was going wrong for the
987     pcreposix and pcrecpp parts of the library. These were overlooked when this
988     problem was solved for the main library.
989 ph10 197
990 ph10 202 9. There were some crude static tests to avoid integer overflow when computing
991     the size of patterns that contain repeated groups with explicit upper
992     limits. As the maximum quantifier is 65535, the maximum group length was
993     set at 30,000 so that the product of these two numbers did not overflow a
994     32-bit integer. However, it turns out that people want to use groups that
995     are longer than 30,000 bytes (though not repeat them that many times).
996     Change 7.0/17 (the refactoring of the way the pattern size is computed) has
997     made it possible to implement the integer overflow checks in a much more
998     dynamic way, which I have now done. The artificial limitation on group
999     length has been removed - we now have only the limit on the total length of
1000     the compiled pattern, which depends on the LINK_SIZE setting.
1001 ph10 208
1002     10. Fixed a bug in the documentation for get/copy named substring when
1003     duplicate names are permitted. If none of the named substrings are set, the
1004     functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1005     empty string.
1006    
1007     11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1008     instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1009     because the ] is interpreted as the first data character and the
1010     terminating ] is not found. PCRE has been made compatible with Perl in this
1011     regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1012     cause memory overwriting.
1013    
1014 ph10 206 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1015     string has been matched (to stop an infinite loop). It was not recognizing
1016 ph10 208 a conditional subpattern that could match an empty string if that
1017 ph10 206 subpattern was within another subpattern. For example, it looped when
1018 ph10 208 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1019 ph10 206 condition was not nested. This bug has been fixed.
1020 ph10 208
1021 ph10 207 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1022     past the start of the subject in the presence of bytes with the top bit
1023     set, for example "\x8aBCD".
1024 ph10 211
1025 ph10 210 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1026     (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1027 ph10 200
1028 ph10 211 14. Optimized (?!) to (*FAIL).
1029 ph10 202
1030 ph10 212 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1031     This restricts code points to be within the range 0 to 0x10FFFF, excluding
1032     the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1033     full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1034     does: it's just the validity check that is more restrictive.
1035 ph10 220
1036     16. Inserted checks for integer overflows during escape sequence (backslash)
1037     processing, and also fixed erroneous offset values for syntax errors during
1038     backslash processing.
1039    
1040 ph10 214 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1041 ph10 220 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1042    
1043 ph10 215 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1044     caused an overrun.
1045 ph10 220
1046     19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1047     something other than just ASCII characters) inside a group that had an
1048     unlimited repeat caused a loop at compile time (while checking to see
1049     whether the group could match an empty string).
1050    
1051 ph10 217 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1052     [\P{Any}] did so. (Error in the code for printing property names.)
1053 ph10 210
1054 ph10 220 21. An orphan \E inside a character class could cause a crash.
1055    
1056     22. A repeated capturing bracket such as (A)? could cause a wild memory
1057 ph10 218 reference during compilation.
1058 ph10 220
1059     23. There are several functions in pcre_compile() that scan along a compiled
1060     expression for various reasons (e.g. to see if it's fixed length for look
1061 ph10 218 behind). There were bugs in these functions when a repeated \p or \P was
1062 ph10 220 present in the pattern. These operators have additional parameters compared
1063     with \d, etc, and these were not being taken into account when moving along
1064 ph10 218 the compiled data. Specifically:
1065 ph10 220
1066     (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1067     length.
1068    
1069     (b) An item such as \pL+ within a repeated group could cause crashes or
1070 ph10 218 loops.
1071 ph10 220
1072 ph10 218 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1073 ph10 220 "reference to non-existent subpattern" error.
1074    
1075 ph10 221 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1076    
1077 ph10 219 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1078 ph10 220 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1079 ph10 211
1080 ph10 222 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1081 ph10 218
1082 ph10 227 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1083 ph10 223 character were causing crashes (broken optimization).
1084 ph10 227
1085 ph10 224 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1086 ph10 227 \p or \P) caused a compile-time loop.
1087    
1088 ph10 226 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1089     break. For example, the unstudied pattern [\r\n]A does not match the string
1090     "\r\nA" because change 7.0/46 below moves the current point on by two
1091     characters after failing to match at the start. However, the pattern \nA
1092 ph10 227 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1093     the same is true. There doesn't seem any very clean way out of this, but
1094     what I have chosen to do makes the common cases work: PCRE now takes note
1095     of whether there can be an explicit match for \r or \n anywhere in the
1096     pattern, and if so, 7.0/46 no longer applies. As part of this change,
1097     there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1098     pattern has explicit CR or LF references.
1099 ph10 222
1100 ph10 227 29. Added (*CR) etc for changing newline setting at start of pattern.
1101 ph10 223
1102 ph10 227
1103 ph10 189 Version 7.2 19-Jun-07
1104     ---------------------
1105    
1106 ph10 157 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1107     which is apparently normally available under Windows.
1108 ph10 159
1109     2. Re-jig the pcregrep tests with different newline settings in an attempt
1110     to make them independent of the local environment's newline setting.
1111 ph10 157
1112 ph10 160 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1113 ph10 165
1114 ph10 161 4. Some of the "internals" tests were previously cut out when the link size
1115     was not 2, because the output contained actual offsets. The recent new
1116 ph10 165 "Z" feature of pcretest means that these can be cut out, making the tests
1117     usable with all link sizes.
1118    
1119 ph10 164 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1120     stack recursion. This gives a massive performance boost under BSD, but just
1121 ph10 165 a small improvement under Linux. However, it saves one field in the frame
1122 ph10 164 in all cases.
1123 ph10 172
1124 ph10 166 6. Added more features from the forthcoming Perl 5.10:
1125 ph10 172
1126 ph10 166 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1127     recursion call. It refers to the nth most recently opened parentheses.
1128 ph10 172
1129 ph10 166 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1130 ph10 172 to be opened parentheses.
1131    
1132     (c) Conditions that refer to capturing parentheses can be specified
1133 ph10 167 relatively, for example, (?(-2)... or (?(+3)...
1134 ph10 172
1135 ph10 168 (d) \K resets the start of the current match so that everything before
1136 ph10 172 is not part of it.
1137    
1138 ph10 171 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1139 ph10 172
1140 ph10 171 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1141 ph10 172 reference syntax.
1142 ph10 182
1143 ph10 175 (g) (?| introduces a group in which the numbering of parentheses in each
1144 ph10 182 alternative starts with the same number.
1145 ph10 172
1146 ph10 182 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1147    
1148 ph10 172 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1149     PCRE_INFO_JCHANGED.
1150    
1151     8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1152     terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1153 ph10 170 for detecting groups that can match an empty string.
1154 ph10 159
1155 ph10 172 9. A pattern with a very large number of alternatives (more than several
1156     hundred) was running out of internal workspace during the pre-compile
1157     phase, where pcre_compile() figures out how much memory will be needed. A
1158     bit of new cunning has reduced the workspace needed for groups with
1159     alternatives. The 1000-alternative test pattern now uses 12 bytes of
1160     workspace instead of running out of the 4096 that are available.
1161 ph10 182
1162 ph10 176 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1163 ph10 172
1164 ph10 179 11. Applied patch from Google to remove an optimization that didn't quite work.
1165     The report of the bug said:
1166 ph10 182
1167 ph10 179 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1168     pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1169 ph10 182 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1170 ph10 185
1171 ph10 184 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1172 ph10 185 it matched the wrong number of bytes.
1173 ph10 172
1174 ph10 179
1175 ph10 155 Version 7.1 24-Apr-07
1176 ph10 98 ---------------------
1177    
1178 ph10 111 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1179 ph10 122 that is more "standard", making use of automake and other Autotools. There
1180 ph10 99 is some re-arrangement of the files and adjustment of comments consequent
1181     on this.
1182 ph10 111
1183     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1184     for recursive directory scanning broke on some systems because the files
1185     are not scanned in any specific order and on different systems the order
1186     was different. A call to "sort" has been inserted into RunGrepTest for the
1187     approprate test as a short-term fix. In the longer term there may be an
1188 ph10 100 alternative.
1189 ph10 111
1190 ph10 100 3. I had an email from Eric Raymond about problems translating some of PCRE's
1191 ph10 111 man pages to HTML (despite the fact that I distribute HTML pages, some
1192     people do their own conversions for various reasons). The problems
1193     concerned the use of low-level troff macros .br and .in. I have therefore
1194     removed all such uses from the man pages (some were redundant, some could
1195 ph10 113 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1196     HTML has been updated to handle .nf/.fi and to complain if it encounters
1197     .br or .in.
1198 ph10 111
1199 ph10 100 4. Updated comments in configure.ac that get placed in config.h.in and also
1200 ph10 123 arranged for config.h to be included in the distribution, with the name
1201 ph10 111 config.h.generic, for the benefit of those who have to compile without
1202     Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1203    
1204     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1205     Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1206 ph10 127 makevp.bat for the current PCRE, using the additional files
1207 ph10 135 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1208 ph10 111
1209     6. A Windows user reported a minor discrepancy with test 2, which turned out
1210     to be caused by a trailing space on an input line that had got lost in his
1211 ph10 102 copy. The trailing space was an accident, so I've just removed it.
1212 ph10 111
1213 ph10 104 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1214 ph10 111 that is needed.
1215    
1216 ph10 105 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1217     as "const" (a) because they are and (b) because it helps the PHP
1218     maintainers who have recently made a script to detect big data structures
1219 ph10 111 in the php code that should be moved to the .rodata section. I remembered
1220     to update Builducptable as well, so it won't revert if ucptable.h is ever
1221 ph10 105 re-created.
1222 ph10 111
1223     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1224 ph10 107 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1225 ph10 111 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1226     support is not required. This saves 1.5-2K of code, which is important in
1227 ph10 107 some applications.
1228 ph10 111
1229 ph10 110 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1230 ph10 111 so as not to refer to the tables, even though these functions will never be
1231     called when UTF-8 support is disabled. Otherwise there are problems with a
1232     shared library.
1233    
1234 ph10 118 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1235    
1236     (a) It was defining its arguments as char * instead of void *.
1237    
1238 ph10 123 (b) It was assuming that all moves were upwards in memory; this was true
1239     a long time ago when I wrote it, but is no longer the case.
1240    
1241 ph10 118 The emulated memove() is provided for those environments that have neither
1242 ph10 123 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1243 ph10 118 is clearly not the case, as these two bugs were recently reported.
1244 ph10 123
1245 ph10 111 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1246 ph10 123 and Detrail to create the HTML documentation, the .txt form of the man
1247     pages, and it removes trailing spaces from listed files. It also creates
1248     pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1249     case, it wraps all the #defines with #ifndefs. This script should be run
1250 ph10 111 before "make dist".
1251 ph10 123
1252 ph10 115 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1253     with Unicode property support.
1254 ph10 123
1255     (a) For a maximizing quantifier, if the two different cases of the
1256     character were of different lengths in their UTF-8 codings (there are
1257     some cases like this - I found 11), and the matching function had to
1258 ph10 115 back up over a mixture of the two cases, it incorrectly assumed they
1259     were both the same length.
1260 ph10 123
1261     (b) When PCRE was configured to use the heap rather than the stack for
1262     recursion during matching, it was not correctly preserving the data for
1263     the other case of a UTF-8 character when checking ahead for a match
1264     while processing a minimizing repeat. If the check also involved
1265 ph10 115 matching a wide character, but failed, corruption could cause an
1266     erroneous result when trying to check for a repeat of the original
1267     character.
1268 ph10 123
1269 ph10 116 13. Some tidying changes to the testing mechanism:
1270 ph10 98
1271 ph10 116 (a) The RunTest script now detects the internal link size and whether there
1272     is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1273 ph10 123 values substituted by "configure". (The RunGrepTest script already did
1274     this for UTF-8.) The configure.ac script no longer substitutes the
1275     relevant variables.
1276    
1277 ph10 116 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1278     with length and offset values. This means that the output is different
1279     for different internal link sizes. Test 2 is skipped for link sizes
1280     other than 2 because of this, bypassing the problem. Unfortunately,
1281     there was also a test in test 3 (the locale tests) that used /B and
1282 ph10 123 failed for link sizes other than 2. Rather than cut the whole test out,
1283     I have added a new /Z option to pcretest that replaces the length and
1284     offset values with spaces. This is now used to make test 3 independent
1285 ph10 122 of link size. (Test 2 will be tidied up later.)
1286 ph10 123
1287     14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1288 ph10 122 segmentation fault instead of returning the appropriate error message.
1289 ph10 134
1290 ph10 131 15. In multiline mode when the newline sequence was set to "any", the pattern
1291 ph10 134 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1292     This doesn't seem right; it now treats the CRLF combination as the line
1293     ending, and so does not match in that case. It's only a pattern such as ^$
1294     that would hit this one: something like ^ABC$ would have failed after \r
1295     and then tried again after \r\n.
1296    
1297 ph10 131 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1298 ph10 134 in an attempt to make files that differ only in their line terminators
1299     compare equal. This works on Linux.
1300 ph10 142
1301 ph10 141 17. Under certain error circumstances pcregrep might try to free random memory
1302     as it exited. This is now fixed, thanks to valgrind.
1303 ph10 142
1304 ph10 141 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1305 ph10 142 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1306     was because its rules for how to advance for /g after matching an empty
1307 ph10 143 string at the end of a line did not allow for this case. They now check for
1308     it specially.
1309 ph10 150
1310     20. pcretest is supposed to handle patterns and data of any length, by
1311     extending its buffers when necessary. It was getting this wrong when the
1312 ph10 147 buffer for a data line had to be extended.
1313 ph10 150
1314 ph10 149 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1315 ph10 150 CRLF as a newline sequence.
1316 ph10 152
1317 ph10 151 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1318 ph10 152 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1319     I have nevertheless tidied it up.
1320 ph10 111
1321 ph10 152 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1322 ph10 123
1323 ph10 153 24. Added a man page for pcre-config.
1324 ph10 152
1325 ph10 153
1326 nigel 93 Version 7.0 19-Dec-06
1327     ---------------------
1328    
1329     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1330     moving to gcc 4.1.1.
1331    
1332     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1333     sys/time.h, which is documented as needed for this function. It doesn't
1334     seem to matter on Linux, but it showed up on some releases of OS X.
1335    
1336     3. It seems that there are systems where bytes whose values are greater than
1337     127 match isprint() in the "C" locale. The "C" locale should be the
1338     default when a C program starts up. In most systems, only ASCII printing
1339     characters match isprint(). This difference caused the output from pcretest
1340     to vary, making some of the tests fail. I have changed pcretest so that:
1341    
1342     (a) When it is outputting text in the compiled version of a pattern, bytes
1343     other than 32-126 are always shown as hex escapes.
1344    
1345     (b) When it is outputting text that is a matched part of a subject string,
1346     it does the same, unless a different locale has been set for the match
1347     (using the /L modifier). In this case, it uses isprint() to decide.
1348    
1349     4. Fixed a major bug that caused incorrect computation of the amount of memory
1350     required for a compiled pattern when options that changed within the
1351     pattern affected the logic of the preliminary scan that determines the
1352     length. The relevant options are -x, and -i in UTF-8 mode. The result was
1353     that the computed length was too small. The symptoms of this bug were
1354     either the PCRE error "internal error: code overflow" from pcre_compile(),
1355     or a glibc crash with a message such as "pcretest: free(): invalid next
1356     size (fast)". Examples of patterns that provoked this bug (shown in
1357     pcretest format) are:
1358    
1359     /(?-x: )/x
1360     /(?x)(?-x: \s*#\s*)/
1361     /((?i)[\x{c0}])/8
1362     /(?i:[\x{c0}])/8
1363    
1364     HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1365     is now done differently.
1366    
1367     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1368     wrapper classes; (b) implement a new function in the C++ scanner that is
1369     more efficient than the old way of doing things because it avoids levels of
1370     recursion in the regex matching; (c) add a paragraph to the documentation
1371     for the FullMatch() function.
1372    
1373     6. The escape sequence \n was being treated as whatever was defined as
1374     "newline". Not only was this contrary to the documentation, which states
1375     that \n is character 10 (hex 0A), but it also went horribly wrong when
1376     "newline" was defined as CRLF. This has been fixed.
1377    
1378     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1379     was being set to -1 for the "end of line" case (supposedly a value that no
1380     character can have). Though this value is never used (the check for end of
1381     line is "zero bytes in current character"), it caused compiler complaints.
1382     I've changed it to 0xffffffff.
1383    
1384     8. In pcre_version.c, the version string was being built by a sequence of
1385     C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1386     string (as it is for production releases) called a macro with an empty
1387     argument. The C standard says the result of this is undefined. The gcc
1388     compiler treats it as an empty string (which was what was wanted) but it is
1389     reported that Visual C gives an error. The source has been hacked around to
1390     avoid this problem.
1391    
1392     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1393     builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1394     instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1395     of them did).
1396    
1397     10. Originally, pcretest opened its input and output without "b"; then I was
1398     told that "b" was needed in some environments, so it was added for release
1399     5.0 to both the input and output. (It makes no difference on Unix-like
1400     systems.) Later I was told that it is wrong for the input on Windows. I've
1401     now abstracted the modes into two macros, to make it easier to fiddle with
1402     them, and removed "b" from the input mode under Windows.
1403    
1404     11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1405    
1406     12. Added -help and --help to pcretest as an official way of being reminded
1407     of the options.
1408    
1409     13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1410     and pcrecpp.cc because they annoy compilers at high warning levels.
1411    
1412     14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1413    
1414     15. Fixed an occurrence of == in configure.ac that should have been = (shell
1415     scripts are not C programs :-) and which was not noticed because it works
1416     on Linux.
1417    
1418     16. pcretest is supposed to handle any length of pattern and data line (as one
1419     line or as a continued sequence of lines) by extending its input buffer if
1420     necessary. This feature was broken for very long pattern lines, leading to
1421     a string of junk being passed to pcre_compile() if the pattern was longer
1422     than about 50K.
1423    
1424     17. I have done a major re-factoring of the way pcre_compile() computes the
1425     amount of memory needed for a compiled pattern. Previously, there was code
1426     that made a preliminary scan of the pattern in order to do this. That was
1427     OK when PCRE was new, but as the facilities have expanded, it has become
1428     harder and harder to keep it in step with the real compile phase, and there
1429     have been a number of bugs (see for example, 4 above). I have now found a
1430     cunning way of running the real compile function in a "fake" mode that
1431     enables it to compute how much memory it would need, while actually only
1432     ever using a few hundred bytes of working memory and without too many
1433     tests of the mode. This should make future maintenance and development
1434     easier. A side effect of this work is that the limit of 200 on the nesting
1435     depth of parentheses has been removed (though this was never a serious
1436     limitation, I suspect). However, there is a downside: pcre_compile() now
1437     runs more slowly than before (30% or more, depending on the pattern). I
1438     hope this isn't a big issue. There is no effect on runtime performance.
1439    
1440     18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1441     newline (only possible for the last line of a file) and it was a
1442     pattern that set a locale (followed by /Lsomething), pcretest crashed.
1443    
1444     19. Added additional timing features to pcretest. (1) The -tm option now times
1445     matching only, not compiling. (2) Both -t and -tm can be followed, as a
1446     separate command line item, by a number that specifies the number of
1447     repeats to use when timing. The default is 50000; this gives better
1448     precision, but takes uncomfortably long for very large patterns.
1449    
1450     20. Extended pcre_study() to be more clever in cases where a branch of a
1451     subpattern has no definite first character. For example, (a*|b*)[cd] would
1452     previously give no result from pcre_study(). Now it recognizes that the
1453     first character must be a, b, c, or d.
1454    
1455     21. There was an incorrect error "recursive call could loop indefinitely" if
1456     a subpattern (or the entire pattern) that was being tested for matching an
1457     empty string contained only one non-empty item after a nested subpattern.
1458     For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1459     incorrectly, because the \d was being skipped in the check.
1460    
1461     22. The pcretest program now has a new pattern option /B and a command line
1462     option -b, which is equivalent to adding /B to every pattern. This causes
1463     it to show the compiled bytecode, without the additional information that
1464     -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1465     is the same as /B/I).
1466    
1467     23. A new optimization is now able automatically to treat some sequences such
1468     as a*b as a*+b. More specifically, if something simple (such as a character
1469     or a simple class like \d) has an unlimited quantifier, and is followed by
1470     something that cannot possibly match the quantified thing, the quantifier
1471     is automatically "possessified".
1472    
1473     24. A recursive reference to a subpattern whose number was greater than 39
1474     went wrong under certain circumstances in UTF-8 mode. This bug could also
1475     have affected the operation of pcre_study().
1476    
1477     25. Realized that a little bit of performance could be had by replacing
1478     (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1479    
1480     26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1481    
1482     27. Possessive quantifiers such as a++ were previously implemented by turning
1483     them into atomic groups such as ($>a+). Now they have their own opcodes,
1484     which improves performance. This includes the automatically created ones
1485     from 23 above.
1486    
1487     28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1488     lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1489     the first matched character to be a colon. This applied both to named and
1490     numbered groups.
1491    
1492     29. The ucpinternal.h header file was missing its idempotency #ifdef.
1493    
1494     30. I was sent a "project" file called libpcre.a.dev which I understand makes
1495     building PCRE on Windows easier, so I have included it in the distribution.
1496    
1497     31. There is now a check in pcretest against a ridiculously large number being
1498     returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1499     loop, the loop is abandoned.
1500    
1501     32. Forward references to subpatterns in conditions such as (?(2)...) where
1502     subpattern 2 is defined later cause pcre_compile() to search forwards in
1503     the pattern for the relevant set of parentheses. This search went wrong
1504     when there were unescaped parentheses in a character class, parentheses
1505     escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1506    
1507     33. "Subroutine" calls and backreferences were previously restricted to
1508     referencing subpatterns earlier in the regex. This restriction has now
1509     been removed.
1510    
1511     34. Added a number of extra features that are going to be in Perl 5.10. On the
1512     whole, these are just syntactic alternatives for features that PCRE had
1513     previously implemented using the Python syntax or my own invention. The
1514     other formats are all retained for compatibility.
1515    
1516     (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1517     as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1518     also .NET compatible.
1519    
1520     (b) A recursion or subroutine call to a named group can now be defined as
1521     (?&name) as well as (?P>name).
1522    
1523     (c) A backreference to a named group can now be defined as \k<name> or
1524     \k'name' as well as (?P=name). The new forms, as well as being in Perl
1525     5.10, are also .NET compatible.
1526    
1527     (d) A conditional reference to a named group can now use the syntax
1528     (?(<name>) or (?('name') as well as (?(name).
1529    
1530     (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1531     groups (named and numbered) that are never evaluated inline, but can be
1532     called as "subroutines" from elsewhere. In effect, the DEFINE condition
1533     is always false. There may be only one alternative in such a group.
1534    
1535     (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1536     as the simple (?(R). The condition is true only if the most recent
1537     recursion is that of the given number or name. It does not search out
1538     through the entire recursion stack.
1539    
1540     (g) The escape \gN or \g{N} has been added, where N is a positive or
1541     negative number, specifying an absolute or relative reference.
1542    
1543     35. Tidied to get rid of some further signed/unsigned compiler warnings and
1544     some "unreachable code" warnings.
1545    
1546     36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1547     things, this adds five new scripts.
1548    
1549     37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1550     There were also incompatibilities regarding the handling of \Q..\E inside
1551     character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1552     hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1553    
1554     38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1555     matches an empty string, and forcibly breaks the loop. There were bugs in
1556     this code in non-simple cases. For a pattern such as ^(a()*)* matched
1557     against aaaa the result was just "a" rather than "aaaa", for example. Two
1558     separate and independent bugs (that affected different cases) have been
1559     fixed.
1560    
1561     39. Refactored the code to abolish the use of different opcodes for small
1562     capturing bracket numbers. This is a tidy that I avoided doing when I
1563     removed the limit on the number of capturing brackets for 3.5 back in 2001.
1564     The new approach is not only tidier, it makes it possible to reduce the
1565     memory needed to fix the previous bug (38).
1566    
1567     40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1568     sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1569     processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1570     mode.
1571    
1572     41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1573     report.
1574    
1575     42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1576     copy construction and assignment in the C++ wrapper.
1577    
1578     43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1579     couple of bugs that could have given wrong results in the "--newline=crlf"
1580     case.
1581    
1582     44. Added a number of casts and did some reorganization of signed/unsigned int
1583     variables following suggestions from Dair Grant. Also renamed the variable
1584     "this" as "item" because it is a C++ keyword.
1585    
1586     45. Arranged for dftables to add
1587    
1588     #include "pcre_internal.h"
1589    
1590     to pcre_chartables.c because without it, gcc 4.x may remove the array
1591     definition from the final binary if PCRE is built into a static library and
1592     dead code stripping is activated.
1593    
1594     46. For an unanchored pattern, if a match attempt fails at the start of a
1595     newline sequence, and the newline setting is CRLF or ANY, and the next two
1596     characters are CRLF, advance by two characters instead of one.
1597    
1598    
1599 nigel 91 Version 6.7 04-Jul-06
1600     ---------------------
1601    
1602     1. In order to handle tests when input lines are enormously long, pcretest has
1603     been re-factored so that it automatically extends its buffers when
1604     necessary. The code is crude, but this _is_ just a test program. The
1605     default size has been increased from 32K to 50K.
1606    
1607     2. The code in pcre_study() was using the value of the re argument before
1608     testing it for NULL. (Of course, in any sensible call of the function, it
1609     won't be NULL.)
1610    
1611     3. The memmove() emulation function in pcre_internal.h, which is used on
1612     systems that lack both memmove() and bcopy() - that is, hardly ever -
1613     was missing a "static" storage class specifier.
1614    
1615     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1616     containing an extended class (one that cannot be represented by a bitmap
1617     because it contains high-valued characters or Unicode property items, e.g.
1618     [\pZ]). Almost always one would set UTF-8 mode when processing such a
1619     pattern, but PCRE should not loop if you do not (it no longer does).
1620     [Detail: two cases were found: (a) a repeated subpattern containing an
1621     extended class; (b) a recursive reference to a subpattern that followed a
1622     previous extended class. It wasn't skipping over the extended class
1623     correctly when UTF-8 mode was not set.]
1624    
1625     5. A negated single-character class was not being recognized as fixed-length
1626     in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1627     compile error "lookbehind assertion is not fixed length".
1628    
1629     6. The RunPerlTest auxiliary script was showing an unexpected difference
1630     between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1631     write a Perl script that can interpret lines of an input file either as
1632     byte characters or as UTF-8, which is what "perltest" was being required to
1633     do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1634     can't do is switch easily at run time between having the "use utf8;" pragma
1635     or not. In the end, I fudged it by using the RunPerlTest script to insert
1636     "use utf8;" explicitly for the UTF-8 tests.
1637    
1638     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1639     the end of the subject string, contrary to the documentation and to what
1640     Perl does. This was true of both matching functions. Now it matches only at
1641     the start of the subject and immediately after *internal* newlines.
1642    
1643     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1644     a pointer to an int instead of a pointer to an unsigned long int. This
1645     caused problems on 64-bit systems.
1646    
1647     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1648     instance of the 'standard' template library not being so standard".
1649    
1650     10. There was no check on the number of named subpatterns nor the maximum
1651     length of a subpattern name. The product of these values is used to compute
1652     the size of the memory block for a compiled pattern. By supplying a very
1653     long subpattern name and a large number of named subpatterns, the size
1654     computation could be caused to overflow. This is now prevented by limiting
1655     the length of names to 32 characters, and the number of named subpatterns
1656     to 10,000.
1657    
1658     11. Subpatterns that are repeated with specific counts have to be replicated in
1659     the compiled pattern. The size of memory for this was computed from the
1660     length of the subpattern and the repeat count. The latter is limited to
1661     65535, but there was no limit on the former, meaning that integer overflow
1662     could in principle occur. The compiled length of a repeated subpattern is
1663     now limited to 30,000 bytes in order to prevent this.
1664    
1665     12. Added the optional facility to have named substrings with the same name.
1666    
1667     13. Added the ability to use a named substring as a condition, using the
1668     Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1669     are numbers (not recommended). Forward references are permitted.
1670    
1671     14. Added forward references in named backreferences (if you see what I mean).
1672    
1673     15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1674     pattern could run off the end of the subject. For example, the pattern
1675     "(?s)(.{1,5})"8 did this with the subject "ab".
1676    
1677     16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1678     PCRE_CASELESS was set when matching characters that were quantified with ?
1679     or *.
1680    
1681     17. A character class other than a single negated character that had a minimum
1682     but no maximum quantifier - for example [ab]{6,} - was not handled
1683     correctly by pce_dfa_exec(). It would match only one character.
1684    
1685     18. A valid (though odd) pattern that looked like a POSIX character
1686     class but used an invalid character after [ (for example [[,abc,]]) caused
1687     pcre_compile() to give the error "Failed: internal error: code overflow" or
1688     in some cases to crash with a glibc free() error. This could even happen if
1689     the pattern terminated after [[ but there just happened to be a sequence of
1690     letters, a binary zero, and a closing ] in the memory that followed.
1691    
1692     19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1693     over the years. Originally (before any Unicode support), just the bottom 8
1694     bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1695     output from "man perlunicode" includes this:
1696    
1697     The regular expression compiler produces polymorphic opcodes. That
1698     is, the pattern adapts to the data and automatically switches to
1699     the Unicode character scheme when presented with Unicode data--or
1700     instead uses a traditional byte scheme when presented with byte
1701     data.
1702    
1703     Sadly, a wide octal escape does not cause a switch, and in a string with
1704     no other multibyte characters, these octal escapes are treated as before.
1705     Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1706     /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1707     Unicode string.
1708    
1709     I have not perpetrated such confusion in PCRE. Up till now, it took just
1710     the bottom 8 bits, as in old Perl. I have now made octal escapes with
1711     values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1712     translate to the appropriate multibyte character.
1713    
1714     29. Applied some refactoring to reduce the number of warnings from Microsoft
1715     and Borland compilers. This has included removing the fudge introduced
1716     seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1717     a warning about an unused variable.
1718    
1719     21. PCRE has not included VT (character 0x0b) in the set of whitespace
1720     characters since release 4.0, because Perl (from release 5.004) does not.
1721     [Or at least, is documented not to: some releases seem to be in conflict
1722     with the documentation.] However, when a pattern was studied with
1723     pcre_study() and all its branches started with \s, PCRE still included VT
1724     as a possible starting character. Of course, this did no harm; it just
1725     caused an unnecessary match attempt.
1726    
1727     22. Removed a now-redundant internal flag bit that recorded the fact that case
1728     dependency changed within the pattern. This was once needed for "required
1729     byte" processing, but is no longer used. This recovers a now-scarce options
1730     bit. Also moved the least significant internal flag bit to the most-
1731     significant bit of the word, which was not previously used (hangover from
1732     the days when it was an int rather than a uint) to free up another bit for
1733     the future.
1734    
1735     23. Added support for CRLF line endings as well as CR and LF. As well as the
1736     default being selectable at build time, it can now be changed at runtime
1737     via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1738     specify that it is scanning data with non-default line endings.
1739    
1740     24. Changed the definition of CXXLINK to make it agree with the definition of
1741     LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1742    
1743     25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1744     recursions. This makes a big different to stack usage for some patterns.
1745    
1746     26. If a subpattern containing a named recursion or subroutine reference such
1747     as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1748     the space required for the compiled pattern went wrong and gave too small a
1749     value. Depending on the environment, this could lead to "Failed: internal
1750     error: code overflow at offset 49" or "glibc detected double free or
1751     corruption" errors.
1752    
1753     27. Applied patches from Google (a) to support the new newline modes and (b) to
1754     advance over multibyte UTF-8 characters in GlobalReplace.
1755    
1756     28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1757     difference for some implementation of PCRE in some Windows version.
1758    
1759     29. Added some extra testing facilities to pcretest:
1760    
1761     \q<number> in a data line sets the "match limit" value
1762     \Q<number> in a data line sets the "match recursion limt" value
1763     -S <number> sets the stack size, where <number> is in megabytes
1764    
1765     The -S option isn't available for Windows.
1766    
1767    
1768 nigel 89 Version 6.6 06-Feb-06
1769     ---------------------
1770    
1771     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1772     in pcreposix.h. I have copied the definition from pcre.h.
1773    
1774     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1775     because pcre.h is no longer a built file.
1776    
1777     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1778     not normally included in the compiled code.
1779    
1780    
1781 nigel 87 Version 6.5 01-Feb-06
1782     ---------------------
1783    
1784     1. When using the partial match feature with pcre_dfa_exec(), it was not
1785     anchoring the second and subsequent partial matches at the new starting
1786     point. This could lead to incorrect results. For example, with the pattern
1787     /1234/, partially matching against "123" and then "a4" gave a match.
1788    
1789     2. Changes to pcregrep:
1790    
1791     (a) All non-match returns from pcre_exec() were being treated as failures
1792     to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1793     error message is output. Some extra information is given for the
1794     PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1795     probably the only errors that are likely to be caused by users (by
1796     specifying a regex that has nested indefinite repeats, for instance).
1797     If there are more than 20 of these errors, pcregrep is abandoned.
1798    
1799     (b) A binary zero was treated as data while matching, but terminated the
1800     output line if it was written out. This has been fixed: binary zeroes
1801     are now no different to any other data bytes.
1802    
1803     (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1804     used to set a locale for matching. The --locale=xxxx long option has
1805     been added (no short equivalent) to specify a locale explicitly on the
1806     pcregrep command, overriding the environment variables.
1807    
1808     (d) When -B was used with -n, some line numbers in the output were one less
1809     than they should have been.
1810    
1811     (e) Added the -o (--only-matching) option.
1812    
1813     (f) If -A or -C was used with -c (count only), some lines of context were
1814     accidentally printed for the final match.
1815    
1816     (g) Added the -H (--with-filename) option.
1817    
1818     (h) The combination of options -rh failed to suppress file names for files
1819     that were found from directory arguments.
1820    
1821     (i) Added the -D (--devices) and -d (--directories) options.
1822    
1823     (j) Added the -F (--fixed-strings) option.
1824    
1825     (k) Allow "-" to be used as a file name for -f as well as for a data file.
1826    
1827     (l) Added the --colo(u)r option.
1828    
1829     (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1830     is not present by default.
1831    
1832     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1833     items such as (?R) or (?1), when the recursion could match a number of
1834     alternatives. If it matched one of the alternatives, but subsequently,
1835     outside the recursion, there was a failure, the code tried to back up into
1836     the recursion. However, because of the way PCRE is implemented, this is not
1837     possible, and the result was an incorrect result from the match.
1838    
1839     In order to prevent this happening, the specification of recursion has
1840     been changed so that all such subpatterns are automatically treated as
1841     atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1842    
1843     4. I had overlooked the fact that, in some locales, there are characters for
1844     which isalpha() is true but neither isupper() nor islower() are true. In
1845     the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1846     and ordfeminine) are like this. This affected the treatment of \w and \W
1847     when they appeared in character classes, but not when they appeared outside
1848     a character class. The bit map for "word" characters is now created
1849     separately from the results of isalnum() instead of just taking it from the
1850     upper, lower, and digit maps. (Plus the underscore character, of course.)
1851    
1852     5. The above bug also affected the handling of POSIX character classes such as
1853     [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1854     permanent tables. Instead, the bit maps for such a class were previously
1855     created as the appropriate unions of the upper, lower, and digit bitmaps.
1856     Now they are created by subtraction from the [[:word:]] class, which has
1857     its own bitmap.
1858    
1859     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1860     It is created by subtracting the vertical space characters (\x09, \x0a,
1861     \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1862     subtraction was done in the overall bitmap for a character class, meaning
1863     that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1864     be recognized. This bug has been fixed.
1865    
1866     7. Patches from the folks at Google:
1867    
1868     (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1869     real life, but is still worth protecting against".
1870    
1871     (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1872     regular expressions".
1873    
1874     (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1875     have it.
1876    
1877     (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1878     "configure" and the latter not, in order to fix a problem somebody had
1879     with compiling the Arg class on HP-UX.
1880    
1881     (e) Improve the error-handling of the C++ wrapper a little bit.
1882    
1883     (f) New tests for checking recursion limiting.
1884    
1885     8. The pcre_memmove() function, which is used only if the environment does not
1886     have a standard memmove() function (and is therefore rarely compiled),
1887     contained two bugs: (a) use of int instead of size_t, and (b) it was not
1888     returning a result (though PCRE never actually uses the result).
1889    
1890     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1891     large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1892     returned instead of calling malloc() with an overflowing number that would
1893     most likely cause subsequent chaos.
1894    
1895     10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1896    
1897     11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1898     with this option is matched, the nmatch and pmatch options of regexec() are
1899     ignored.
1900    
1901     12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1902     provided in case anyone wants to the the POSIX interface with UTF-8
1903     strings.
1904    
1905     13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1906     C++ linking (needed for some HP-UX environments).
1907    
1908     14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1909     (unused parameter) and in the pcre_printint() function (omitted "default"
1910     switch label when the default is to do nothing).
1911    
1912     15. Added some code to make it possible, when PCRE is compiled as a C++
1913     library, to replace subject pointers for pcre_exec() with a smart pointer
1914     class, thus making it possible to process discontinuous strings.
1915    
1916     16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1917     much the same function. They were added by different people who were trying
1918     to make PCRE easy to compile on non-Unix systems. It has been suggested
1919     that PCRE_EXPORT be abolished now that there is more automatic apparatus
1920     for compiling on Windows systems. I have therefore replaced it with
1921     PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1922     defaults to "extern" for C or "extern C" for C++, which works fine on
1923     Unix-like systems. It is now possible to override the value of PCRE_DATA_
1924     SCOPE with something explicit in config.h. In addition:
1925    
1926     (a) pcreposix.h still had just "extern" instead of either of these macros;
1927     I have replaced it with PCRE_DATA_SCOPE.
1928    
1929     (b) Functions such as _pcre_xclass(), which are internal to the library,
1930     but external in the C sense, all had PCRE_EXPORT in their definitions.
1931     This is apparently wrong for the Windows case, so I have removed it.
1932     (It makes no difference on Unix-like systems.)
1933    
1934     17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1935     of recursive calls to match(). This is different to MATCH_LIMIT because
1936     that limits the total number of calls to match(), not all of which increase
1937     the depth of recursion. Limiting the recursion depth limits the amount of
1938     stack (or heap if NO_RECURSE is set) that is used. The default can be set
1939     when PCRE is compiled, and changed at run time. A patch from Google adds
1940     this functionality to the C++ interface.
1941    
1942     18. Changes to the handling of Unicode character properties:
1943    
1944     (a) Updated the table to Unicode 4.1.0.
1945    
1946     (b) Recognize characters that are not in the table as "Cn" (undefined).
1947    
1948     (c) I revised the way the table is implemented to a much improved format
1949     which includes recognition of ranges. It now supports the ranges that
1950     are defined in UnicodeData.txt, and it also amalgamates other
1951     characters into ranges. This has reduced the number of entries in the
1952     table from around 16,000 to around 3,000, thus reducing its size
1953     considerably. I realized I did not need to use a tree structure after
1954     all - a binary chop search is just as efficient. Having reduced the
1955     number of entries, I extended their size from 6 bytes to 8 bytes to
1956     allow for more data.
1957    
1958     (d) Added support for Unicode script names via properties such as \p{Han}.
1959    
1960     19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1961     matching that character.
1962    
1963     20. When matching a repeated Unicode property with a minimum greater than zero,
1964     (for example \pL{2,}), PCRE could look past the end of the subject if it
1965     reached it while seeking the minimum number of characters. This could
1966     happen only if some of the characters were more than one byte long, because
1967     there is a check for at least the minimum number of bytes.
1968    
1969     21. Refactored the implementation of \p and \P so as to be more general, to
1970     allow for more different types of property in future. This has changed the
1971     compiled form incompatibly. Anybody with saved compiled patterns that use
1972     \p or \P will have to recompile them.
1973    
1974     22. Added "Any" and "L&" to the supported property types.
1975    
1976     23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1977     but give a compile time error if the value is greater than 0xff.
1978    
1979     24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1980     accidentally not being installed or uninstalled.
1981    
1982     25. The pcre.h file was built from pcre.h.in, but the only changes that were
1983     made were to insert the current release number. This seemed silly, because
1984     it made things harder for people building PCRE on systems that don't run
1985     "configure". I have turned pcre.h into a distributed file, no longer built
1986     by "configure", with the version identification directly included. There is
1987     no longer a pcre.h.in file.
1988    
1989     However, this change necessitated a change to the pcre-config script as
1990     well. It is built from pcre-config.in, and one of the substitutions was the
1991     release number. I have updated configure.ac so that ./configure now finds
1992     the release number by grepping pcre.h.
1993    
1994     26. Added the ability to run the tests under valgrind.
1995    
1996    
1997 nigel 85 Version 6.4 05-Sep-05
1998     ---------------------
1999    
2000     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2001     "--" to be printed when multiple files were scanned, even when none of the
2002     -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2003     consider it to be a bug, and have restored the previous behaviour.
2004    
2005     2. A couple of code tidies to get rid of compiler warnings.
2006    
2007     3. The pcretest program used to cheat by referring to symbols in the library
2008     whose names begin with _pcre_. These are internal symbols that are not
2009     really supposed to be visible externally, and in some environments it is
2010     possible to suppress them. The cheating is now confined to including
2011     certain files from the library's source, which is a bit cleaner.
2012    
2013     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2014     file's purpose clearer.
2015    
2016     5. Reorganized pcre_ucp_findchar().
2017    
2018    
2019 nigel 83 Version 6.3 15-Aug-05
2020     ---------------------
2021    
2022     1. The file libpcre.pc.in did not have general read permission in the tarball.
2023    
2024     2. There were some problems when building without C++ support:
2025    
2026     (a) If C++ support was not built, "make install" and "make test" still
2027     tried to test it.
2028    
2029     (b) There were problems when the value of CXX was explicitly set. Some
2030     changes have been made to try to fix these, and ...
2031    
2032     (c) --disable-cpp can now be used to explicitly disable C++ support.
2033    
2034     (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2035     backslash in a target when C++ was disabled. This confuses some
2036     versions of "make", apparently. Using an intermediate variable solves
2037     this. (Same for CPP_LOBJ.)
2038    
2039     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2040     (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2041     necessary on certain architectures.
2042    
2043     4. Added a setting of -export-symbols-regex to the link command to remove
2044     those symbols that are exported in the C sense, but actually are local
2045     within the library, and not documented. Their names all begin with
2046     "_pcre_". This is not a perfect job, because (a) we have to except some
2047     symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2048     available (and never for static libraries). I have made a note to try to
2049     find a way round (a) in the future.
2050    
2051    
2052 nigel 81 Version 6.2 01-Aug-05
2053     ---------------------
2054    
2055     1. There was no test for integer overflow of quantifier values. A construction
2056     such as {1111111111111111} would give undefined results. What is worse, if
2057     a minimum quantifier for a parenthesized subpattern overflowed and became
2058     negative, the calculation of the memory size went wrong. This could have
2059     led to memory overwriting.
2060    
2061     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2062    
2063     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2064     operating environments where this matters.
2065    
2066     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2067     PCRE options from within the C++ wrapper.
2068    
2069     5. Named capturing subpatterns were not being correctly counted when a pattern
2070     was compiled. This caused two problems: (a) If there were more than 100
2071     such subpatterns, the calculation of the memory needed for the whole
2072     compiled pattern went wrong, leading to an overflow error. (b) Numerical
2073     back references of the form \12, where the number was greater than 9, were
2074     not recognized as back references, even though there were sufficient
2075     previous subpatterns.
2076    
2077     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2078     versions of gcc, e.g. 2.95.4.
2079    
2080    
2081 nigel 79 Version 6.1 21-Jun-05
2082     ---------------------
2083    
2084     1. There was one reference to the variable "posix" in pcretest.c that was not
2085     surrounded by "#if !defined NOPOSIX".
2086    
2087     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2088     the cross-check on the old pcre_info() function, for the benefit of the
2089     cut-down version of PCRE that is currently imported into Exim.
2090    
2091     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2092     allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2093     patterns that start (?i) but I don't think that matters. The use of (?i) is
2094     just an example; this all applies to the other options as well.
2095    
2096     4. Since libtool seems to echo the compile commands it is issuing, the output
2097     from "make" can be reduced a bit by putting "@" in front of each libtool
2098     compile command.
2099    
2100     5. Patch from the folks at Google for configure.in to be a bit more thorough
2101     in checking for a suitable C++ installation before trying to compile the
2102     C++ stuff. This should fix a reported problem when a compiler was present,
2103     but no suitable headers.
2104    
2105     6. The man pages all had just "PCRE" as their title. I have changed them to
2106     be the relevant file name. I have also arranged that these names are
2107     retained in the file doc/pcre.txt, which is a concatenation in text format
2108     of all the man pages except the little individual ones for each function.
2109    
2110     7. The NON-UNIX-USE file had not been updated for the different set of source
2111     files that come with release 6. I also added a few comments about the C++
2112     wrapper.
2113    
2114    
2115 nigel 77 Version 6.0 07-Jun-05
2116     ---------------------
2117    
2118     1. Some minor internal re-organization to help with my DFA experiments.
2119    
2120     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2121     didn't matter for the library itself when fully configured, but did matter
2122     when compiling without UCP support, or within Exim, where the ucp files are
2123     not imported.
2124    
2125     3. Refactoring of the library code to split up the various functions into
2126     different source modules. The addition of the new DFA matching code (see
2127     below) to a single monolithic source would have made it really too
2128     unwieldy, quite apart from causing all the code to be include in a
2129     statically linked application, when only some functions are used. This is
2130     relevant even without the DFA addition now that patterns can be compiled in
2131     one application and matched in another.
2132    
2133     The downside of splitting up is that there have to be some external
2134     functions and data tables that are used internally in different modules of
2135     the library but which are not part of the API. These have all had their
2136     names changed to start with "_pcre_" so that they are unlikely to clash
2137     with other external names.
2138    
2139     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2140     a different (DFA) algorithm. Although it is slower than the original
2141     function, it does have some advantages for certain types of matching
2142     problem.
2143    
2144     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2145     including restarting after a partial match.
2146    
2147     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2148     defined when compiling for Windows was sent to me. I have put it into the
2149     code, though I have no means of testing or verifying it.
2150    
2151     7. Added the pcre_refcount() auxiliary function.
2152    
2153     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2154     match before or at the first newline in the subject string. In pcretest,
2155     the /f option on a pattern can be used to set this.
2156    
2157     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2158     would behave wrongly. This has been present in PCRE since release 4.0.
2159    
2160     10. A number of changes to the pcregrep command:
2161    
2162     (a) Refactored how -x works; insert ^(...)$ instead of setting
2163     PCRE_ANCHORED and checking the length, in preparation for adding
2164     something similar for -w.
2165    
2166     (b) Added the -w (match as a word) option.
2167    
2168     (c) Refactored the way lines are read and buffered so as to have more
2169     than one at a time available.
2170    
2171     (d) Implemented a pcregrep test script.
2172    
2173     (e) Added the -M (multiline match) option. This allows patterns to match
2174     over several lines of the subject. The buffering ensures that at least
2175     8K, or the rest of the document (whichever is the shorter) is available
2176     for matching (and similarly the previous 8K for lookbehind assertions).
2177    
2178     (f) Changed the --help output so that it now says
2179    
2180     -w, --word-regex(p)
2181    
2182     instead of two lines, one with "regex" and the other with "regexp"
2183     because that confused at least one person since the short forms are the
2184     same. (This required a bit of code, as the output is generated
2185     automatically from a table. It wasn't just a text change.)
2186    
2187     (g) -- can be used to terminate pcregrep options if the next thing isn't an
2188     option but starts with a hyphen. Could be a pattern or a path name
2189     starting with a hyphen, for instance.
2190    
2191     (h) "-" can be given as a file name to represent stdin.
2192    
2193     (i) When file names are being printed, "(standard input)" is used for
2194     the standard input, for compatibility with GNU grep. Previously
2195     "<stdin>" was used.
2196    
2197     (j) The option --label=xxx can be used to supply a name to be used for
2198     stdin when file names are being printed. There is no short form.
2199    
2200     (k) Re-factored the options decoding logic because we are going to add
2201     two more options that take data. Such options can now be given in four
2202     different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2203    
2204     (l) Added the -A, -B, and -C options for requesting that lines of context
2205     around matches be printed.
2206    
2207     (m) Added the -L option to print the names of files that do not contain
2208     any matching lines, that is, the complement of -l.
2209    
2210     (n) The return code is 2 if any file cannot be opened, but pcregrep does
2211     continue to scan other files.
2212    
2213     (o) The -s option was incorrectly implemented. For compatibility with other
2214     greps, it now suppresses the error message for a non-existent or non-
2215     accessible file (but not the return code). There is a new option called
2216     -q that suppresses the output of matching lines, which was what -s was
2217     previously doing.
2218    
2219     (p) Added --include and --exclude options to specify files for inclusion
2220     and exclusion when recursing.
2221    
2222     11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2223     Hopefully, it now does.
2224    
2225     12. Missing cast in pcre_study().
2226    
2227     13. Added an "uninstall" target to the makefile.
2228    
2229     14. Replaced "extern" in the function prototypes in Makefile.in with
2230     "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2231     world, but is set differently for Windows.
2232    
2233     15. Added a second compiling function called pcre_compile2(). The only
2234     difference is that it has an extra argument, which is a pointer to an
2235     integer error code. When there is a compile-time failure, this is set
2236     non-zero, in addition to the error test pointer being set to point to an
2237     error message. The new argument may be NULL if no error number is required
2238     (but then you may as well call pcre_compile(), which is now just a
2239     wrapper). This facility is provided because some applications need a
2240     numeric error indication, but it has also enabled me to tidy up the way
2241     compile-time errors are handled in the POSIX wrapper.
2242    
2243     16. Added VPATH=.libs to the makefile; this should help when building with one
2244     prefix path and installing with another. (Or so I'm told by someone who
2245     knows more about this stuff than I do.)
2246    
2247     17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2248     passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2249     match everything, including newlines. This is not POSIX-compatible, but
2250     somebody wanted the feature. From pcretest it can be activated by using
2251     both the P and the s flags.
2252    
2253     18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2254    
2255     19. libpcre.pc was being incorrectly installed as executable.
2256    
2257     20. A couple of places in pcretest check for end-of-line by looking for '\n';
2258     it now also looks for '\r' so that it will work unmodified on Windows.
2259    
2260     21. Added Google's contributed C++ wrapper to the distribution.
2261    
2262     22. Added some untidy missing memory free() calls in pcretest, to keep
2263     Electric Fence happy when testing.
2264    
2265    
2266    
2267 nigel 75 Version 5.0 13-Sep-04
2268     ---------------------
2269    
2270     1. Internal change: literal characters are no longer packed up into items
2271     containing multiple characters in a single byte-string. Each character
2272     is now matched using a separate opcode. However, there may be more than one
2273     byte in the character in UTF-8 mode.
2274    
2275     2. The pcre_callout_block structure has two new fields: pattern_position and
2276     next_item_length. These contain the offset in the pattern to the next match
2277     item, and its length, respectively.
2278    
2279     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2280     insertion of callouts before each pattern item. Added the /C option to
2281     pcretest to make use of this.
2282    
2283     4. On the advice of a Windows user, the lines
2284    
2285     #if defined(_WIN32) || defined(WIN32)
2286     _setmode( _fileno( stdout ), 0x8000 );
2287     #endif /* defined(_WIN32) || defined(WIN32) */
2288    
2289     have been added to the source of pcretest. This apparently does useful
2290     magic in relation to line terminators.
2291    
2292     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2293     for the benefit of those environments where the "b" makes a difference.
2294    
2295     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2296     to know about it. I have put a hack into configure.in that adds in code
2297     to set GCC=yes if CC=icc. This seems to end up at a point in the
2298     generated configure script that is early enough to affect the setting of
2299     compiler options, which is what is needed, but I have no means of testing
2300     whether it really works. (The user who reported this had patched the
2301     generated configure script, which of course I cannot do.)
2302    
2303     LATER: After change 22 below (new libtool files), the configure script
2304     seems to know about icc (and also ecc). Therefore, I have commented out
2305     this hack in configure.in.
2306    
2307     7. Added support for pkg-config (2 patches were sent in).
2308    
2309     8. Negated POSIX character classes that used a combination of internal tables
2310     were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2311     [[:^ascii]]. Typically, they would match almost any characters. The other
2312     POSIX classes were not broken in this way.
2313    
2314     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2315     to find the match, as PCRE was deluded into thinking that the match had to
2316     start at the start point or following a newline. The same bug applied to
2317     patterns with negative forward assertions or any backward assertions
2318     preceding ".*" at the start, unless the pattern required a fixed first
2319     character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2320    
2321     10. In UTF-8 mode, when moving forwards in the subject after a failed match
2322     starting at the last subject character, bytes beyond the end of the subject
2323     string were read.
2324    
2325     11. Renamed the variable "class" as "classbits" to make life easier for C++
2326     users. (Previously there was a macro definition, but it apparently wasn't
2327     enough.)
2328    
2329     12. Added the new field "tables" to the extra data so that tables can be passed
2330     in at exec time, or the internal tables can be re-selected. This allows
2331     a compiled regex to be saved and re-used at a later time by a different
2332     program that might have everything at different addresses.
2333    
2334     13. Modified the pcre-config script so that, when run on Solaris, it shows a
2335     -R library as well as a -L library.
2336    
2337     14. The debugging options of pcretest (-d on the command line or D on a
2338     pattern) showed incorrect output for anything following an extended class
2339     that contained multibyte characters and which was followed by a quantifier.
2340    
2341     15. Added optional support for general category Unicode character properties
2342     via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2343     support. It adds about 90K to the size of the library. The meanings of the
2344     inbuilt class escapes such as \d and \s have NOT been changed.
2345    
2346     16. Updated pcredemo.c to include calls to free() to release the memory for the
2347     compiled pattern.
2348    
2349     17. The generated file chartables.c was being created in the source directory
2350     instead of in the building directory. This caused the build to fail if the
2351     source directory was different from the building directory, and was
2352     read-only.
2353    
2354     18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2355     file. No doubt somebody will tell me if they don't make sense... Also added
2356     Dan Mooney's comments about building on OpenVMS.
2357    
2358     19. Added support for partial matching via the PCRE_PARTIAL option for
2359     pcre_exec() and the \P data escape in pcretest.
2360    
2361     20. Extended pcretest with 3 new pattern features:
2362    
2363     (i) A pattern option of the form ">rest-of-line" causes pcretest to
2364     write the compiled pattern to the file whose name is "rest-of-line".
2365     This is a straight binary dump of the data, with the saved pointer to
2366     the character tables forced to be NULL. The study data, if any, is
2367     written too. After writing, pcretest reads a new pattern.
2368    
2369     (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2370     compiled pattern from the given file. There must not be any
2371     occurrences of "<" in the file name (pretty unlikely); if there are,
2372     pcretest will instead treat the initial "<" as a pattern delimiter.
2373     After reading in the pattern, pcretest goes on to read data lines as
2374     usual.
2375    
2376     (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2377     and 16-bit fields in a compiled pattern, to simulate a pattern that
2378     was compiled on a host of opposite endianness.
2379    
2380     21. The pcre-exec() function can now cope with patterns that were compiled on
2381     hosts of opposite endianness, with this restriction:
2382    
2383     As for any compiled expression that is saved and used later, the tables
2384     pointer field cannot be preserved; the extra_data field in the arguments
2385     to pcre_exec() should be used to pass in a tables address if a value
2386     other than the default internal tables were used at compile time.
2387    
2388     22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2389     now diagnosed as an error. Previously, most of the time, a negative number
2390     would have been treated as zero, but if in addition "ovector" was passed as
2391     NULL, a crash could occur.
2392    
2393     23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2394     new versions from the libtool 1.5 distribution (the last one is a copy of
2395     a file called libtool.m4). This seems to have fixed the need to patch
2396     "configure" to support Darwin 1.3 (which I used to do). However, I still
2397     had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2398     workstation).
2399    
2400     24. Changed the PCRE licence to be the more standard "BSD" licence.
2401    
2402    
2403 nigel 73 Version 4.5 01-Dec-03
2404     ---------------------
2405    
2406     1. There has been some re-arrangement of the code for the match() function so
2407     that it can be compiled in a version that does not call itself recursively.
2408     Instead, it keeps those local variables that need separate instances for
2409     each "recursion" in a frame on the heap, and gets/frees frames whenever it
2410     needs to "recurse". Keeping track of where control must go is done by means
2411     of setjmp/longjmp. The whole thing is implemented by a set of macros that
2412     hide most of the details from the main code, and operates only if
2413     NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2414     "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2415     operating.
2416    
2417     To make it easier for callers to provide specially tailored get/free
2418     functions for this usage, two new functions, pcre_stack_malloc, and
2419     pcre_stack_free, are used. They are always called in strict stacking order,
2420     and the size of block requested is always the same.
2421    
2422     The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2423     PCRE has been compiled to use the stack or the heap for recursion. The
2424     -C option of pcretest uses this to show which version is compiled.
2425    
2426     A new data escape \S, is added to pcretest; it causes the amounts of store
2427     obtained and freed by both kinds of malloc/free at match time to be added
2428     to the output.
2429    
2430     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2431     what's available on my current Linux desktop machine.
2432    
2433     3. When matching a UTF-8 string, the test for a valid string at the start has
2434     been extended. If start_offset is not zero, PCRE now checks that it points
2435     to a byte that is the start of a UTF-8 character. If not, it returns
2436     PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2437     this is necessary because there may be backward assertions in the pattern.
2438     When matching the same subject several times, it may save resources to use
2439     PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2440    
2441     4. The code for checking the validity of UTF-8 strings has been tightened so
2442     that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2443     containing "overlong sequences".
2444    
2445     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2446     I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2447     should have been "&", but it just so happened that all the cases this let
2448     through by mistake were picked up later in the function.
2449    
2450     6. I had used a variable called "isblank" - this is a C99 function, causing
2451     some compilers to warn. To avoid this, I renamed it (as "blankclass").
2452    
2453     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2454     prompting; (b) run "./pcretest /dev/null" at the start of the test script
2455     so the version is shown; (c) stop "make test" echoing "./RunTest".
2456    
2457     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2458    
2459     9. The prototype for memmove() for systems that don't have it was using
2460     size_t, but the inclusion of the header that defines size_t was later. I've
2461     moved the #includes for the C headers earlier to avoid this.
2462    
2463     10. Added some adjustments to the code to make it easier to compiler on certain
2464     special systems:
2465    
2466     (a) Some "const" qualifiers were missing.
2467     (b) Added the macro EXPORT before all exported functions; by default this
2468     is defined to be empty.
2469     (c) Changed the dftables auxiliary program (that builds chartables.c) so
2470     that it reads its output file name as an argument instead of writing
2471     to the standard output and assuming this can be redirected.
2472    
2473     11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2474     class containing characters with values greater than 255, PCRE compilation
2475     went into a loop.
2476    
2477     12. A recursive reference to a subpattern that was within another subpattern
2478     that had a minimum quantifier of zero caused PCRE to crash. For example,
2479     (x(y(?2))z)? provoked this bug with a subject that got as far as the
2480     recursion. If the recursively-called subpattern itself had a zero repeat,
2481     that was OK.
2482    
2483     13. In pcretest, the buffer for reading a data line was set at 30K, but the
2484     buffer into which it was copied (for escape processing) was still set at
2485     1024, so long lines caused crashes.
2486    
2487     14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2488     "internal error: code overflow...". This applied to any character class
2489     that was followed by a possessive quantifier.
2490    
2491     15. Modified the Makefile to add libpcre.la as a prerequisite for
2492     libpcreposix.la because I was told this is needed for a parallel build to
2493     work.
2494    
2495     16. If a pattern that contained .* following optional items at the start was
2496     studied, the wrong optimizing data was generated, leading to matching
2497     errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2498     matching string must start with a or b or c. The correct conclusion for
2499     this pattern is that a match can start with any character.
2500    
2501    
2502 nigel 71 Version 4.4 13-Aug-03
2503     ---------------------
2504    
2505     1. In UTF-8 mode, a character class containing characters with values between
2506     127 and 255 was not handled correctly if the compiled pattern was studied.
2507     In fixing this, I have also improved the studying algorithm for such
2508     classes (slightly).
2509    
2510     2. Three internal functions had redundant arguments passed to them. Removal
2511     might give a very teeny performance improvement.
2512    
2513     3. Documentation bug: the value of the capture_top field in a callout is *one
2514     more than* the number of the hightest numbered captured substring.
2515    
2516     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2517     in incorrectly linking with a previously installed version. They now link
2518     explicitly with libpcre.la.
2519    
2520     5. configure.in no longer needs to recognize Cygwin specially.
2521    
2522     6. A problem in pcre.in for Windows platforms is fixed.
2523    
2524     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2525     pcretest, it used to include the size of the study block as part of its
2526     output. Unfortunately, the structure contains a field that has a different
2527     size on different hardware architectures. This meant that the tests that
2528     showed this size failed. As the block is currently always of a fixed size,
2529     this information isn't actually particularly useful in pcretest output, so
2530     I have just removed it.
2531    
2532     8. Three pre-processor statements accidentally did not start in column 1.
2533     Sadly, there are *still* compilers around that complain, even though
2534     standard C has not required this for well over a decade. Sigh.
2535    
2536     9. In pcretest, the code for checking callouts passed small integers in the
2537     callout_data field, which is a void * field. However, some picky compilers
2538     complained about the casts involved for this on 64-bit systems. Now
2539     pcretest passes the address of the small integer instead, which should get
2540     rid of the warnings.
2541    
2542     10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2543     both compile and run time, and gives an error if an invalid UTF-8 sequence
2544     is found. There is a option for disabling this check in cases where the
2545     string is known to be correct and/or the maximum performance is wanted.
2546    
2547     11. In response to a bug report, I changed one line in Makefile.in from
2548    
2549     -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2550     to
2551     -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2552    
2553     to look similar to other lines, but I have no way of telling whether this
2554     is the right thing to do, as I do not use Windows. No doubt I'll get told
2555     if it's wrong...
2556    
2557    
2558 nigel 69 Version 4.3 21-May-03
2559     ---------------------
2560    
2561     1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2562     Makefile.
2563    
2564     2. Some refactoring to improve the quality of the code:
2565    
2566     (i) The utf8_table... variables are now declared "const".
2567    
2568     (ii) The code for \cx, which used the "case flipping" table to upper case
2569     lower case letters, now just substracts 32. This is ASCII-specific,
2570     but the whole concept of \cx is ASCII-specific, so it seems
2571     reasonable.
2572    
2573     (iii) PCRE was using its character types table to recognize decimal and
2574     hexadecimal digits in the pattern. This is silly, because it handles
2575     only 0-9, a-f, and A-F, but the character types table is locale-
2576     specific, which means strange things might happen. A private
2577     table is now used for this - though it costs 256 bytes, a table is
2578     much faster than multiple explicit tests. Of course, the standard
2579     character types table is still used for matching digits in subject
2580     strings against \d.
2581    
2582     (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2583     ending in _t are). So I've renamed it as ESC_tee.
2584    
2585     3. The first argument for regexec() in the POSIX wrapper should have been
2586     defined as "const".
2587    
2588     4. Changed pcretest to use malloc() for its buffers so that they can be
2589     Electric Fenced for debugging.
2590    
2591     5. There were several places in the code where, in UTF-8 mode, PCRE would try
2592     to read one or more bytes before the start of the subject string. Often this
2593     had no effect on PCRE's behaviour, but in some circumstances it could
2594     provoke a segmentation fault.
2595    
2596     6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2597     to try to read one or more bytes before the start of the subject string.
2598    
2599     7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2600     UTF-8 support could misbehave in various ways if the subject string
2601     contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2602     area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2603     back over UTF-8 characters.)
2604    
2605    
2606 nigel 67 Version 4.2 14-Apr-03
2607     ---------------------
2608    
2609     1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2610    
2611     2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2612     [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2613     [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2614     [WIN_PREFIX]: new variable, "cyg" for Cygwin
2615     * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2616     and BUILD_EXEEXT
2617     Note: automatic setting of the BUILD variables is not yet working
2618     set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2619     compile-time but not at link-time
2620     [LINK]: use for linking executables only
2621     make different versions for Windows and non-Windows
2622     [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2623     libraries
2624     [LINK_FOR_BUILD]: new variable
2625     [OBJEXT]: use throughout
2626     [EXEEXT]: use throughout
2627     <winshared>: new target
2628     <wininstall>: new target
2629     <dftables.o>: use native compiler
2630     <dftables>: use native linker
2631     <install>: handle Windows platform correctly
2632     <clean>: ditto
2633     <check>: ditto
2634     copy DLL to top builddir before testing
2635    
2636     As part of these changes, -no-undefined was removed again. This was reported
2637     to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2638     in any case.
2639    
2640     3. Some tidies to get rid of compiler warnings:
2641    
2642     . In the match_data structure, match_limit was an unsigned long int, whereas
2643     match_call_count was an int. I've made them both unsigned long ints.
2644    
2645     . In pcretest the fact that a const uschar * doesn't automatically cast to
2646     a void * provoked a warning.
2647    
2648     . Turning on some more compiler warnings threw up some "shadow" variables
2649     and a few more missing casts.
2650    
2651     4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2652     option, a class that contained a single character with a value between 128
2653     and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2654    
2655     5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2656     option, a class that contained several characters, but with at least one
2657     whose value was between 128 and 255 caused PCRE to crash.
2658    
2659    
2660 nigel 65 Version 4.1 12-Mar-03
2661     ---------------------
2662 nigel 63
2663 nigel 65 1. Compiling with gcc -pedantic found a couple of places where casts were
2664     needed, and a string in dftables.c that was longer than standard compilers are
2665     required to support.
2666    
2667     2. Compiling with Sun's compiler found a few more places where the code could
2668     be tidied up in order to avoid warnings.
2669    
2670     3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2671     first of these names is deprecated in the latest Autoconf in favour of the name
2672     CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2673     compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2674     analogy I have changed it to CFLAGS_FOR_BUILD.
2675    
2676     4. Added -no-undefined to the linking command in the Makefile, because this is
2677     apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2678     linking step for the pcreposix library.
2679    
2680     5. PCRE was failing to diagnose the case of two named groups with the same
2681     name.
2682    
2683     6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2684     literal character that is needed in the subject for a match, and scans along to
2685     ensure that it is present before embarking on the full matching process. This
2686     saves time in cases of nested unlimited repeats that are never going to match.
2687     Problem: the scan can take a lot of time if the subject is very long (e.g.
2688     megabytes), thus penalizing straightforward matches. It is now done only if the
2689     amount of subject to be scanned is less than 1000 bytes.
2690    
2691     7. A lesser problem with the same optimization is that it was recording the
2692     first character of an anchored pattern as "needed", thus provoking a search
2693     right along the subject, even when the first match of the pattern was going to
2694     fail. The "needed" character is now not set for anchored patterns, unless it
2695     follows something in the pattern that is of non-fixed length. Thus, it still
2696     fulfils its original purpose of finding quick non-matches in cases of nested
2697     unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2698    
2699    
2700     Version 4.0 17-Feb-03
2701     ---------------------
2702    
2703 nigel 63 1. If a comment in an extended regex that started immediately after a meta-item
2704     extended to the end of string, PCRE compiled incorrect data. This could lead to
2705     all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2706    
2707     2. Moved to autoconf 2.53 and libtool 1.4.2.
2708    
2709     3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2710     the special perltest8 script is no longer needed - all the tests can be run
2711     from a single perltest script.
2712    
2713     4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2714     by \s. It has now been removed in PCRE. This means it isn't recognized as
2715     whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2716     class [:space:] *does* include VT, thereby creating a mess.
2717    
2718     5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2719     space and tab.
2720    
2721     6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2722     its new features into the main test script, reducing the number of scripts.
2723    
2724     7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2725     were backward compatible, and made the (?i) apply to the whole pattern, as if
2726     /i were given. Now it behaves more logically, and applies the option setting
2727     only to what follows. PCRE has been changed to follow suit. However, if it
2728     finds options settings right at the start of the pattern, it extracts them into
2729     the global options, as before. Thus, they show up in the info data.
2730    
2731     8. Added support for the \Q...\E escape sequence. Characters in between are
2732     treated as literals. This is slightly different from Perl in that $ and @ are
2733     also handled as literals inside the quotes. In Perl, they will cause variable
2734     interpolation. Note the following examples:
2735    
2736     Pattern PCRE matches Perl matches
2737    
2738     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2739     \Qabc\$xyz\E abc\$xyz abc\$xyz
2740     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2741    
2742     For compatibility with Perl, \Q...\E sequences are recognized inside character
2743     classes as well as outside them.
2744    
2745     9. Re-organized 3 code statements in pcretest to avoid "overflow in
2746     floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2747     (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2748     signed/unsigned warnings.
2749    
2750     10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2751     option for pcretest, so I've replaced it by a simple function that does just
2752     that job.
2753    
2754     11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2755     "pcregrep -".
2756    
2757     12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2758     Java package. This provides some syntactic sugar for simple cases of what my
2759     documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2760     as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2761     item, you can use this simplified notation. Note that only makes sense with
2762     greedy quantifiers. Consequently, the use of the possessive quantifier forces
2763     greediness, whatever the setting of the PCRE_UNGREEDY option.
2764    
2765     13. A change of greediness default within a pattern was not taking effect at
2766     the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2767     subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2768     was abstracted outside.
2769    
2770     14. PCRE now supports the \G assertion. It is true when the current matching
2771     position is at the start point of the match. This differs from \A when the
2772     starting offset is non-zero. Used with the /g option of pcretest (or similar
2773     code), it works in the same way as it does for Perl's /g option. If all
2774     alternatives of a regex begin with \G, the expression is anchored to the start
2775     match position, and the "anchored" flag is set in the compiled expression.
2776    
2777     15. Some bugs concerning the handling of certain option changes within patterns
2778     have been fixed. These applied to options other than (?ims). For example,
2779     "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2780     been the other way round. Some of this was related to change 7 above.
2781    
2782     16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2783     features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2784     and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2785     POSIX classes only within a class (e.g. /[[:alpha:]]/).
2786    
2787     17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2788     mode. Unlike ".", it always matches newline, whatever the setting of
2789     PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2790     assertions. Perl allows it, but it doesn't (in general) work because it can't
2791     calculate the length of the lookbehind. At least, that's the case for Perl
2792     5.8.0 - I've been told they are going to document that it doesn't work in
2793     future.
2794    
2795     18. Added an error diagnosis for escapes that PCRE does not support: these are
2796     \L, \l, \N, \P, \p, \U, \u, and \X.
2797    
2798     19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2799     reading past the end of the pattern in cases such as /[abcd/.
2800    
2801     20. PCRE was getting more memory than necessary for patterns with classes that
2802     contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2803    
2804     21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2805     compiling PCRE for use with Virtual Pascal.
2806    
2807     22. Small fix to the Makefile to make it work properly if the build is done
2808     outside the source tree.
2809    
2810     23. Added a new extension: a condition to go with recursion. If a conditional
2811     subpattern starts with (?(R) the "true" branch is used if recursion has
2812     happened, whereas the "false" branch is used only at the top level.
2813    
2814     24. When there was a very long string of literal characters (over 255 bytes
2815     without UTF support, over 250 bytes with UTF support), the computation of how
2816     much memory was required could be incorrect, leading to segfaults or other
2817     strange effects.
2818    
2819     25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2820     start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2821     there was a subsequent back reference to those brackets. This meant that, for
2822     example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2823     possible to check for precisely this case. All we can do is abandon the
2824     optimization if .* occurs inside capturing brackets when there are any back
2825     references whatsoever. (See below for a better fix that came later.)
2826    
2827     26. The handling of the optimization for finding the first character of a
2828     non-anchored pattern, and for finding a character that is required later in the
2829     match were failing in some cases. This didn't break the matching; it just
2830     failed to optimize when it could. The way this is done has been re-implemented.
2831    
2832     27. Fixed typo in error message for invalid (?R item (it said "(?p").
2833    
2834     28. Added a new feature that provides some of the functionality that Perl
2835     provides with (?{...}). The facility is termed a "callout". The way it is done
2836     in PCRE is for the caller to provide an optional function, by setting
2837     pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2838     global variable. By default it is unset, which disables all calling out. To get
2839     the function called, the regex must include (?C) at appropriate points. This
2840     is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2841     This provides a means of identifying different callout points. When PCRE
2842     reaches such a point in the regex, if pcre_callout has been set, the external
2843     function is called. It is provided with data in a structure called
2844     pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2845     matching continues; if it returns a non-zero value, the match at the current
2846     point fails. However, backtracking will occur if possible. [This was changed
2847     later and other features added - see item 49 below.]
2848    
2849     29. pcretest is upgraded to test the callout functionality. It provides a
2850     callout function that displays information. By default, it shows the start of
2851     the match and the current position in the text. There are some new data escapes
2852     to vary what happens:
2853    
2854     \C+ in addition, show current contents of captured substrings
2855     \C- do not supply a callout function
2856     \C!n return 1 when callout number n is reached
2857     \C!n!m return 1 when callout number n is reached for the mth time
2858    
2859     30. If pcregrep was called with the -l option and just a single file name, it
2860     output "<stdin>" if a match was found, instead of the file name.
2861    
2862     31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2863     slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2864     pcre_exec(). This saves a malloc/free per call. The default value of
2865     POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2866     when configuring.
2867    
2868     32. The default maximum size of a compiled pattern is 64K. There have been a
2869     few cases of people hitting this limit. The code now uses macros to handle the
2870     storing of links as offsets within the compiled pattern. It defaults to 2-byte
2871     links, but this can be changed to 3 or 4 bytes by --with-link-size when
2872     configuring. Tests 2 and 5 work only with 2-byte links because they output
2873     debugging information about compiled patterns.
2874    
2875     33. Internal code re-arrangements:
2876    
2877     (a) Moved the debugging function for printing out a compiled regex into
2878     its own source file (printint.c) and used #include to pull it into
2879     pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2880     separate copies.
2881    
2882     (b) Defined the list of op-code names for debugging as a macro in
2883     internal.h so that it is next to the definition of the opcodes.
2884    
2885     (c) Defined a table of op-code lengths for simpler skipping along compiled
2886     code. This is again a macro in internal.h so that it is next to the
2887     definition of the opcodes.
2888    
2889     34. Added support for recursive calls to individual subpatterns, along the
2890     lines of Robin Houston's patch (but implemented somewhat differently).
2891    
2892     35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2893     allow it to read and process whole directories in Win32. This code was
2894     contributed by Lionel Fourquaux; it has not been tested by me.
2895    
2896     36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2897     used to name a group. Names consist of alphanumerics and underscores, and must
2898     be unique. Back references use the syntax (?P=name) and recursive calls use
2899     (?P>name) which is a PCRE extension to the Python extension. Groups still have
2900     numbers. The function pcre_fullinfo() can be used after compilation to extract
2901     a name/number map. There are three relevant calls:
2902    
2903     PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2904     PCRE_INFO_NAMECOUNT yields the number of entries
2905     PCRE_INFO_NAMETABLE yields a pointer to the map.
2906    
2907     The map is a vector of fixed-size entries. The size of each entry depends on
2908     the length of the longest name used. The first two bytes of each entry are the
2909     group number, most significant byte first. There follows the corresponding
2910     name, zero terminated. The names are in alphabetical order.
2911    
2912     37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2913     case instead of 255. Making it the same both with and without UTF-8 support
2914     means that the same test output works with both.
2915    
2916     38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2917     calling malloc() with a zero argument.
2918    
2919     39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2920     optimization. I've improved things by keeping a bitmap of backreferences with
2921     numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2922     fact referenced, the optimization can be applied. It is unlikely that a
2923     relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2924     the match to follow \n) will appear inside brackets with a number greater than
2925     31, but if it does, any back reference > 31 suppresses the optimization.
2926    
2927     40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2928     of disabling numbered capturing parentheses. Any opening parenthesis that is
2929     not followed by ? behaves as if it were followed by ?: but named parentheses
2930     can still be used for capturing (and they will acquire numbers in the usual
2931     way).
2932    
2933     41. Redesigned the return codes from the match() function into yes/no/error so
2934     that errors can be passed back from deep inside the nested calls. A malloc
2935     failure while inside a recursive subpattern call now causes the
2936     PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2937    
2938     42. It is now possible to set a limit on the number of times the match()
2939     function is called in a call to pcre_exec(). This facility makes it possible to
2940     limit the amount of recursion and backtracking, though not in a directly
2941     obvious way, because the match() function is used in a number of different
2942     circumstances. The count starts from zero for each position in the subject
2943     string (for non-anchored patterns). The default limit is, for compatibility, a
2944     large number, namely 10 000 000. You can change this in two ways:
2945    
2946     (a) When configuring PCRE before making, you can use --with-match-limit=n
2947     to set a default value for the compiled library.
2948    
2949     (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2950     a different value is set. See 45 below.
2951    
2952     If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2953    
2954     43. Added a new function pcre_config(int, void *) to enable run-time extraction
2955     of things that can be changed at compile time. The first argument specifies
2956     what is wanted and the second points to where the information is to be placed.
2957     The current list of available information is:
2958    
2959     PCRE_CONFIG_UTF8
2960    
2961     The output is an integer that is set to one if UTF-8 support is available;
2962     otherwise it is set to zero.
2963    
2964     PCRE_CONFIG_NEWLINE
2965    
2966     The output is an integer that it set to the value of the code that is used for
2967     newline. It is either LF (10) or CR (13).
2968    
2969     PCRE_CONFIG_LINK_SIZE
2970    
2971     The output is an integer that contains the number of bytes used for internal
2972     linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2973    
2974     PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2975    
2976     The output is an integer that contains the threshold above which the POSIX
2977     interface uses malloc() for output vectors. See item 31 above.
2978    
2979     PCRE_CONFIG_MATCH_LIMIT
2980    
2981     The output is an unsigned integer that contains the default limit of the number
2982     of match() calls in a pcre_exec() execution. See 42 above.
2983    
2984     44. pcretest has been upgraded by the addition of the -C option. This causes it
2985     to extract all the available output from the new pcre_config() function, and to
2986     output it. The program then exits immediately.
2987    
2988     45. A need has arisen to pass over additional data with calls to pcre_exec() in
2989     order to support additional features. One way would have been to define
2990     pcre_exec2() (for example) with extra arguments, but this would not have been
2991     extensible, and would also have required all calls to the original function to
2992     be mapped to the new one. Instead, I have chosen to extend the mechanism that
2993     is used for passing in "extra" data from pcre_study().
2994    
2995     The pcre_extra structure is now exposed and defined in pcre.h. It currently
2996     contains the following fields:
2997    
2998     flags a bitmap indicating which of the following fields are set
2999     study_data opaque data from pcre_study()
3000     match_limit a way of specifying a limit on match() calls for a specific
3001     call to pcre_exec()
3002     callout_data data for callouts (see 49 below)
3003    
3004     The flag bits are also defined in pcre.h, and are
3005    
3006     PCRE_EXTRA_STUDY_DATA
3007     PCRE_EXTRA_MATCH_LIMIT
3008     PCRE_EXTRA_CALLOUT_DATA
3009    
3010     The pcre_study() function now returns one of these new pcre_extra blocks, with
3011     the actual study data pointed to by the study_data field, and the
3012     PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3013     before. That is, this change is entirely upwards-compatible and requires no
3014     change to existing code.
3015    
3016     If you want to pass in additional data to pcre_exec(), you can either place it
3017     in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3018     block.
3019    
3020     46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3021     data string contains the escape sequence \M, pcretest calls pcre_exec() several
3022     times with different match limits, until it finds the minimum value needed for
3023     pcre_exec() to complete. The value is then output. This can be instructive; for
3024     most simple matches the number is quite small, but for pathological cases it
3025     gets very large very quickly.
3026    
3027     47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3028     returns the size of the data block pointed to by the study_data field in a
3029     pcre_extra block, that is, the value that was passed as the argument to
3030     pcre_malloc() when PCRE was getting memory in which to place the information
3031     created by pcre_study(). The fourth argument should point to a size_t variable.
3032     pcretest has been extended so that this information is shown after a successful
3033     pcre_study() call when information about the compiled regex is being displayed.
3034    
3035     48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3036     because what follows is always an absolute path. (Later: it turns out that this
3037     is more than cosmetic for MinGW, because it doesn't like empty path
3038     components.)
3039    
3040     49. Some changes have been made to the callout feature (see 28 above):
3041    
3042     (i) A callout function now has three choices for what it returns:
3043    
3044     0 => success, carry on matching
3045     > 0 => failure at this point, but backtrack if possible
3046     < 0 => serious error, return this value from pcre_exec()
3047    
3048     Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3049     values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3050     "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3051     use by callout functions. It will never be used by PCRE itself.
3052    
3053     (ii) The pcre_extra structure (see 45 above) has a void * field called
3054     callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3055     pcre_callout_block structure has a field of the same name. The contents of
3056     the field passed in the pcre_extra structure are passed to the callout
3057     function in the corresponding field in the callout block. This makes it
3058     easier to use the same callout-containing regex from multiple threads. For
3059     testing, the pcretest program has a new data escape
3060    
3061     \C*n pass the number n (may be negative) as callout_data
3062    
3063     If the callout function in pcretest receives a non-zero value as
3064     callout_data, it returns that value.
3065    
3066     50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3067     there were some redundant $(CFLAGS) in commands that are now specified as
3068     $(LINK), which already includes $(CFLAGS).
3069    
3070     51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3071     has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3072     with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3073     one-byte characters throughout. Note that case-insensitive matching applies
3074     only to characters whose values are less than 256. PCRE doesn't support the
3075     notion of cases for higher-valued characters.
3076    
3077     (i) A character class whose characters are all within 0-255 is handled as
3078     a bit map, and the map is inverted for negative classes. Previously, a
3079     character > 255 always failed to match such a class; however it should
3080     match if the class was a negative one (e.g. [^ab]). This has been fixed.
3081    
3082     (ii) A negated character class with a single character < 255 is coded as
3083     "not this character" (OP_NOT). This wasn't working properly when the test
3084     character was multibyte, either singly or repeated.
3085    
3086     (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3087     mode, for example: \x{100}{2,3}.
3088    
3089     (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3090     singly or repeated) now correctly test multibyte characters. However,
3091     PCRE doesn't recognize any characters with values greater than 255 as
3092     digits, spaces, or word characters. Such characters always match \D, \S,
3093     and \W, and never match \d, \s, or \w.
3094    
3095     (v) Classes may now contain characters and character ranges with values
3096     greater than 255. For example: [ab\x{100}-\x{400}].
3097    
3098     (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3099     PCRE in UTF-8 mode.
3100    
3101     52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3102     PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3103     retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3104     value.)
3105    
3106     53. The single man page has become too large. I have therefore split it up into
3107     a number of separate man pages. These also give rise to individual HTML pages;
3108     these are now put in a separate directory, and there is an index.html page that
3109     lists them all. Some hyperlinking between the pages has been installed.
3110    
3111     54. Added convenience functions for handling named capturing parentheses.
3112    
3113     55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3114     aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3115     true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3116     are faulted.
3117    
3118     56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3119     calling configure. These values are used when compiling the dftables.c program
3120     which is run to generate the source of the default character tables. They
3121     default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3122     you will need to set these values.
3123    
3124     57. Updated the building process for Windows DLL, as provided by Fred Cox.
3125    
3126    
3127     Version 3.9 02-Jan-02
3128 nigel 61 ---------------------
3129    
3130     1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3131    
3132     2. If --disable-static was given, the building process failed when trying to
3133     build pcretest and pcregrep. (For some reason it was using libtool to compile
3134     them, which is not right, as they aren't part of the library.)
3135    
3136    
3137 nigel 59 Version 3.8 18-Dec-01
3138     ---------------------
3139    
3140     1. The experimental UTF-8 code was completely screwed up. It was packing the
3141     bytes in the wrong order. How dumb can you get?
3142    
3143    
3144 nigel 57 Version 3.7 29-Oct-01
3145     ---------------------
3146    
3147     1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3148     This caused pcretest, when used on the test data, to segfault. Unfortunately,
3149     this didn't happen under Solaris 8, where I normally test things.
3150    
3151 nigel 59 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3152     doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3153     isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3154     this fix an hour or so after the initial 3.7 release.)
3155 nigel 57
3156 nigel 59
3157 nigel 55 Version 3.6 23-Oct-01
3158     ---------------------
3159    
3160     1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3161     offsets passed as NULL with zero offset count.
3162    
3163     2. The config.guess and config.sub files had not been updated when I moved to
3164     the latest autoconf.
3165    
3166    
3167 nigel 53 Version 3.5 15-Aug-01
3168     ---------------------
3169 nigel 5
3170 nigel 53 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3171     had been forgotten.
3172    
3173     2. By using declared but undefined structures, we can avoid using "void"
3174     definitions in pcre.h while keeping the internal definitions of the structures
3175     private.
3176    
3177     3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3178     user point of view, this means that both static and shared libraries are built
3179     by default, but this can be individually controlled. More of the work of
3180     handling this static/shared cases is now inside libtool instead of PCRE's make
3181     file.
3182    
3183     4. The pcretest utility is now installed along with pcregrep because it is
3184     useful for users (to test regexs) and by doing this, it automatically gets
3185     relinked by libtool. The documentation has been turned into a man page, so
3186     there are now .1, .txt, and .html versions in /doc.
3187    
3188     5. Upgrades to pcregrep:
3189     (i) Added long-form option names like gnu grep.
3190     (ii) Added --help to list all options with an explanatory phrase.
3191     (iii) Added -r, --recursive to recurse into sub-directories.
3192     (iv) Added -f, --file to read patterns from a file.
3193    
3194     6. pcre_exec() was referring to its "code" argument before testing that
3195     argument for NULL (and giving an error if it was NULL).
3196    
3197     7. Upgraded Makefile.in to allow for compiling in a different directory from
3198     the source directory.
3199    
3200     8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3201     options bits, the pointer it was passed was to an int instead of to an unsigned
3202     long int. This mattered only on 64-bit systems.
3203    
3204     9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3205     generated) instead of pcre.in, which it its source. Also made the same change
3206     in several of the .c files.
3207    
3208     10. A new release of gcc defines printf() as a macro, which broke pcretest
3209     because it had an ifdef in the middle of a string argument for printf(). Fixed
3210     by using separate calls to printf().
3211    
3212     11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3213     script, to force use of CR or LF instead of \n in the source. On non-Unix
3214     systems, the value can be set in config.h.
3215    
3216     12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3217     absolute limit. Changed the text of the error message to make this clear, and
3218     likewise updated the man page.
3219    
3220     13. The limit of 99 on the number of capturing subpatterns has been removed.
3221     The new limit is 65535, which I hope will not be a "real" limit.
3222    
3223    
3224 nigel 51 Version 3.4 22-Aug-00
3225     ---------------------
3226    
3227     1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3228    
3229     2. Diagnose condition (?(0) as an error instead of crashing on matching.
3230    
3231    
3232 nigel 49 Version 3.3 01-Aug-00
3233     ---------------------
3234    
3235     1. If an octal character was given, but the value was greater than \377, it
3236     was not getting masked to the least significant bits, as documented. This could
3237     lead to crashes in some systems.
3238    
3239     2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3240     the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3241    
3242     3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3243     These just pass their arguments on to (pcre_free)(), but they are provided
3244     because some uses of PCRE bind it to non-C systems that can call its functions,
3245     but cannot call free() or pcre_free() directly.
3246    
3247     4. Add "make test" as a synonym for "make check". Corrected some comments in
3248     the Makefile.
3249    
3250     5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3251     Makefile.
3252    
3253     6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3254     command called pgrep for grepping around the active processes.
3255    
3256     7. Added the beginnings of support for UTF-8 character strings.
3257    
3258     8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3259     RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3260     the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3261     out for the ar command.)
3262    
3263    
3264 nigel 47 Version 3.2 12-May-00
3265     ---------------------
3266    
3267     This is purely a bug fixing release.
3268    
3269     1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3270     of ZA. This was just one example of several cases that could provoke this bug,
3271     which was introduced by change 9 of version 2.00. The code for breaking
3272     infinite loops after an iteration that matches an empty string was't working
3273     correctly.
3274    
3275     2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3276     when matched against abbab (for example). After matching an empty string, it
3277     wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3278     caused it to match further down the string than it should.
3279    
3280     3. The code contained an inclusion of sys/types.h. It isn't clear why this
3281     was there because it doesn't seem to be needed, and it causes trouble on some
3282     systems, as it is not a Standard C header. It has been removed.
3283    
3284     4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3285     were reported on the Macintosh. The changes were from
3286    
3287     while ((c = *(++ptr)) != 0 && c != '\n');
3288     to
3289     while ((c = *(++ptr)) != 0 && c != '\n') ;
3290    
3291     Totally extraordinary, but if that's what it takes...
3292    
3293     5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3294     available. Added HAVE_BCOPY and an autoconf test for it; if neither
3295     HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3296     assumes the way PCRE uses memmove() (always moving upwards).
3297    
3298     6. PCRE is being used in one environment where strchr() is not available. There
3299     was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3300     faster code anyway.
3301    
3302    
3303 nigel 45 Version 3.1 09-Feb-00
3304     ---------------------
3305    
3306     The only change in this release is the fixing of some bugs in Makefile.in for
3307     the "install" target:
3308    
3309     (1) It was failing to install pcreposix.h.
3310    
3311     (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3312    
3313    
3314 nigel 43 Version 3.0 01-Feb-00
3315     ---------------------
3316 nigel 41
3317     1. Add support for the /+ modifier to perltest (to output $` like it does in
3318     pcretest).
3319    
3320     2. Add support for the /g modifier to perltest.
3321    
3322     3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3323     matches null strings.
3324    
3325     4. Fix perltest so that it doesn't do unwanted things when fed an empty
3326     pattern. Perl treats empty patterns specially - it reuses the most recent
3327     pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3328     effect.
3329    
3330     5. The POSIX interface was broken in that it was just handing over the POSIX
3331     captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3332     required a bigger vector, with some working space on the end. This means that
3333     the POSIX wrapper now has to get and free some memory, and copy the results.
3334    
3335 nigel 43 6. Added some simple autoconf support, placing the test data and the
3336     documentation in separate directories, re-organizing some of the
3337     information files, and making it build pcre-config (a GNU standard). Also added
3338     libtool support for building PCRE as a shared library, which is now the
3339     default.
3340 nigel 41
3341 nigel 43 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3342     09 are not valid octal constants. Single digits will be used for minor values
3343     less than 10.
3344    
3345     8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3346     existing programs that set these in the POSIX interface can use PCRE without
3347     modification.
3348    
3349     9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3350     return all that pcre_info() returns, plus additional data. The pcre_info()
3351     function is retained for compatibility, but is considered to be obsolete.
3352    
3353     10. Added experimental recursion feature (?R) to handle one common case that
3354     Perl 5.6 will be able to do with (?p{...}).
3355    
3356     11. Added support for POSIX character classes like [:alpha:], which Perl is
3357     adopting.
3358    
3359    
3360 nigel 39 Version 2.08 31-Aug-99
3361     ----------------------
3362    
3363     1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3364     trying to match at the startoffset position, but instead was moving forward to
3365     the next newline as if a previous match had failed.
3366    
3367     2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3368     and could get into a loop if a null string was matched other than at the start
3369     of the subject.
3370    
3371     3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3372     be distinguished at compile time, and for completeness also added PCRE_DATE.
3373    
3374     5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3375     in GnuWin32 environments.
3376    
3377    
3378 nigel 37 Version 2.07 29-Jul-99
3379     ----------------------
3380    
3381     1. The documentation is now supplied in plain text form and HTML as well as in
3382     the form of man page sources.
3383    
3384     2. C++ compilers don't like assigning (void *) values to other pointer types.
3385     In particular this affects malloc(). Although there is no problem in Standard
3386     C, I've put in casts to keep C++ compilers happy.
3387    
3388     3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3389     should be (const char *).
3390    
3391     4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3392     be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3393     However, I haven't made this a standard facility. The documentation doesn't
3394     mention it, and the Makefile doesn't support it.
3395    
3396     5. The Makefile now contains an "install" target, with editable destinations at
3397     the top of the file. The pcretest program is not installed.
3398    
3399     6. pgrep -V now gives the PCRE version number and date.
3400    
3401     7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3402     causing the entire string to be ignored, instead of just the last character.
3403    
3404     8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3405     non-matching string, it can take a very, very long time, even for strings of
3406     quite modest length, because of the nested recursion. PCRE now does better in
3407     some of these cases. It does this by remembering the last required literal
3408     character in the pattern, and pre-searching the subject to ensure it is present
3409     before running the real match. In other words, it applies a heuristic to detect
3410     some types of certain failure quickly, and in the above example, if presented
3411     with a string that has no trailing " it gives "no match" very quickly.
3412    
3413     9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3414     other alternatives are tried instead.
3415    
3416    
3417 nigel 35 Version 2.06 09-Jun-99
3418     ----------------------
3419    
3420     1. Change pcretest's output for amount of store used to show just the code
3421     space, because the remainder (the data block) varies in size between 32-bit and
3422     64-bit systems.
3423    
3424     2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3425     start matching at. This allows lookbehinds to work when searching for multiple
3426     occurrences in a string.
3427    
3428     3. Added additional options to pcretest for testing multiple occurrences:
3429    
3430     /+ outputs the rest of the string that follows a match
3431     /g loops for multiple occurrences, using the new startoffset argument
3432     /G loops for multiple occurrences by passing an incremented pointer
3433    
3434     4. PCRE wasn't doing the "first character" optimization for patterns starting
3435     with \b or \B, though it was doing it for other lookbehind assertions. That is,
3436     it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3437     the letter 'x'. On long subject strings, this gives a significant speed-up.
3438    
3439    
3440 nigel 33 Version 2.05 21-Apr-99
3441     ----------------------
3442    
3443     1. Changed the type of magic_number from int to long int so that it works
3444     properly on 16-bit systems.
3445    
3446     2. Fixed a bug which caused patterns starting with .* not to work correctly
3447     when the subject string contained newline characters. PCRE was assuming
3448     anchoring for such patterns in all cases, which is not correct because .* will
3449     not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3450     DOTALL is set at top level; otherwise it knows that patterns starting with .*
3451     must be retried after every newline in the subject.
3452    
3453    
3454 nigel 31 Version 2.04 18-Feb-99
3455     ----------------------
3456    
3457     1. For parenthesized subpatterns with repeats whose minimum was zero, the
3458     computation of the store needed to hold the pattern was incorrect (too large).
3459     If such patterns were nested a few deep, this could multiply and become a real
3460     problem.
3461    
3462     2. Added /M option to pcretest to show the memory requirement of a specific
3463     pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3464    
3465     3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3466     compiled in such a way that the backtracking after subsequent failure was
3467     pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3468     ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3469    
3470    
3471 nigel 29 Version 2.03 02-Feb-99
3472     ----------------------
3473    
3474     1. Fixed typo and small mistake in man page.
3475    
3476 nigel 31 2. Added 4th condition (GPL supersedes if conflict) and created separate
3477     LICENCE file containing the conditions.
3478 nigel 29
3479     3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3480     Perl, that is the internal \ allows the delimiter to be included in the
3481     pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3482     the final delimiter, add \ to the end of the pattern (to test the error).
3483    
3484     4. Added the convenience functions for extracting substrings after a successful
3485     match. Updated pcretest to make it able to test these functions.
3486    
3487    
3488 nigel 27 Version 2.02 14-Jan-99
3489     ----------------------
3490    
3491     1. Initialized the working variables associated with each extraction so that
3492     their saving and restoring doesn't refer to uninitialized store.
3493    
3494     2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3495     compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3496     fix the problem.
3497    
3498     3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3499     calls, and wasn't printing the correct value for compiling calls. Increased the
3500     default value of LOOPREPEAT, and the number of significant figures in the
3501     times.
3502    
3503     4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3504    
3505     5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3506     a building problem on Windows NT with a FAT file system.
3507    
3508    
3509 nigel 25 Version 2.01 21-Oct-98
3510     ----------------------
3511    
3512     1. Changed the API for pcre_compile() to allow for the provision of a pointer
3513     to character tables built by pcre_maketables() in the current locale. If NULL
3514     is passed, the default tables are used.
3515    
3516    
3517 nigel 23 Version 2.00 24-Sep-98
3518 nigel 21 ----------------------
3519    
3520 nigel 23 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3521     it any more.
3522    
3523     2. Allow quantification of (?>) groups, and make it work correctly.
3524    
3525     3. The first character computation wasn't working for (?>) groups.
3526    
3527     4. Correct the implementation of \Z (it is permitted to match on the \n at the
3528     end of the subject) and add 5.005's \z, which really does match only at the
3529     very end of the subject.
3530    
3531     5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3532    
3533     6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3534     DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3535     localized options. All options to pcre_study() were also removed.
3536    
3537     7. Add other new features from 5.005:
3538    
3539     $(?<= positive lookbehind
3540     $(?<! negative lookbehind
3541     (?imsx-imsx) added the unsetting capability
3542     such a setting is global if at outer level; local otherwise
3543     (?imsx-imsx:) non-capturing groups with option setting
3544     (?(cond)re|re) conditional pattern matching
3545    
3546     A backreference to itself in a repeated group matches the previous
3547     captured string.
3548    
3549     8. General tidying up of studying (both automatic and via "study")
3550     consequential on the addition of new assertions.
3551    
3552     9. As in 5.005, unlimited repeated groups that could match an empty substring
3553     are no longer faulted at compile time. Instead, the loop is forcibly broken at
3554     runtime if any iteration does actually match an empty substring.
3555    
3556     10. Include the RunTest script in the distribution.
3557    
3558     11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3559     discrepancies, some of which were old and were also with respect to 5.004. They
3560     have now been fixed.
3561    
3562    
3563     Version 1.09 28-Apr-98
3564     ----------------------
3565    
3566 nigel 21 1. A negated single character class followed by a quantifier with a minimum
3567     value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3568     program crashes, or just wrong answers. This did not apply to negated classes
3569     containing more than one character, or to minima other than one.
3570    
3571    
3572 nigel 19 Version 1.08 27-Mar-98
3573     ----------------------
3574    
3575     1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3576    
3577     2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3578     latter must appear before anything that relies on it in the pattern.
3579    
3580    
3581 nigel 17 Version 1.07 16-Feb-98
3582     ----------------------
3583    
3584     1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3585     repeat of a potentially empty string).
3586    
3587    
3588 nigel 15 Version 1.06 23-Jan-98
3589     ----------------------
3590    
3591     1. Added Markus Oberhumer's little patches for C++.
3592    
3593     2. Literal strings longer than 255 characters were broken.
3594    
3595    
3596 nigel 13 Version 1.05 23-Dec-97
3597     ----------------------
3598    
3599     1. Negated character classes containing more than one character were failing if
3600     PCRE_CASELESS was set at run time.
3601    
3602    
3603 nigel 11 Version 1.04 19-Dec-97
3604     ----------------------
3605    
3606     1. Corrected the man page, where some "const" qualifiers had been omitted.
3607    
3608     2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3609     input syntax.
3610    
3611     3. Fixed memory leak which occurred when a regex with back references was
3612     matched with an offsets vector that wasn't big enough. The temporary memory
3613     that is used in this case wasn't being freed if the match failed.
3614    
3615     4. Tidied pcretest to ensure it frees memory that it gets.
3616    
3617     5. Temporary memory was being obtained in the case where the passed offsets
3618     vector was exactly big enough.
3619    
3620     6. Corrected definition of offsetof() from change 5 below.
3621    
3622     7. I had screwed up change 6 below and broken the rules for the use of
3623     setjmp(). Now fixed.
3624    
3625    
3626 nigel 9 Version 1.03 18-Dec-97
3627     ----------------------
3628    
3629     1. A erroneous regex with a missing opening parenthesis was correctly
3630     diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3631     on some systems.
3632    
3633     2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3634     it was reported that one broken compiler failed on the former because "code" is
3635     also an independent variable.
3636    
3637     3. The erroneous regex a[]b caused an array overrun reference.
3638    
3639     4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3640     fail on data ending with that character. (It was going on too far, and checking
3641     the next character, typically a binary zero.) This was specific to the
3642     optimized code for single-character negative classes.
3643    
3644     5. Added a contributed patch from the TIN world which does the following:
3645    
3646     + Add an undef for memmove, in case the the system defines a macro for it.
3647    
3648     + Add a definition of offsetof(), in case there isn't one. (I don't know
3649     the reason behind this - offsetof() is part of the ANSI standard - but
3650     it does no harm).
3651    
3652     + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3653     most of the places where whitespace preceded '#'. I have given up and
3654     allowed the remaining 2 cases to be at the margin.
3655    
3656     + Rename some variables in pcre to eliminate shadowing. This seems very
3657     pedantic, but does no harm, of course.
3658    
3659     6. Moved the call to setjmp() into its own function, to get rid of warnings
3660     from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3661    
3662     7. Constructs such as \d{8,} were compiling into the equivalent of
3663 nigel 11 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3664 nigel 9 outcome, but in this particular case used more store than had been allocated,
3665     which caused the bug to be discovered because it threw up an internal error.
3666    
3667     8. The debugging code in both pcre and pcretest for outputting the compiled
3668     form of a regex was going wrong in the case of back references followed by
3669     curly-bracketed repeats.
3670    
3671    
3672 nigel 7 Version 1.02 12-Dec-97
3673     ----------------------
3674    
3675     1. Typos in pcre.3 and comments in the source fixed.
3676    
3677     2. Applied a contributed patch to get rid of places where it used to remove
3678     'const' from variables, and fixed some signed/unsigned and uninitialized
3679     variable warnings.
3680    
3681     3. Added the "runtest" target to Makefile.
3682    
3683     4. Set default compiler flag to -O2 rather than just -O.
3684    
3685    
3686 nigel 5 Version 1.01 19-Nov-97
3687     ----------------------
3688    
3689     1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3690     like /([ab]*)*/, that is, for classes with more than one character in them.
3691    
3692     2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3693     as /((?>a*))*/ (a PCRE_EXTRA facility).
3694    
3695    
3696     Version 1.00 18-Nov-97
3697     ----------------------
3698    
3699     1. Added compile-time macros to support systems such as SunOS4 which don't have
3700     memmove() or strerror() but have other things that can be used instead.
3701    
3702     2. Arranged that "make clean" removes the executables.
3703    
3704    
3705 nigel 3 Version 0.99 27-Oct-97
3706     ----------------------
3707    
3708     1. Fixed bug in code for optimizing classes with only one character. It was
3709     initializing a 32-byte map regardless, which could cause it to run off the end
3710     of the memory it had got.
3711    
3712     2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3713    
3714    
3715     Version 0.98 22-Oct-97
3716     ----------------------
3717    
3718     1. Fixed bug in code for handling temporary memory usage when there are more
3719     back references than supplied space in the ovector. This could cause segfaults.
3720    
3721    
3722     Version 0.97 21-Oct-97
3723     ----------------------
3724    
3725     1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3726    
3727     2. Optimized negated single characters not to use a bit map.
3728    
3729     3. Brought error texts together as macro definitions; clarified some of them;
3730     fixed one that was wrong - it said "range out of order" when it meant "invalid
3731     escape sequence".
3732    
3733     4. Changed some char * arguments to const char *.
3734    
3735     5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3736    
3737     6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3738     pcretest.
3739    
3740    
3741     Version 0.96 16-Oct-97
3742     ----------------------
3743    
3744     1. Added a simple "pgrep" utility to the distribution.
3745    
3746     2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3747     unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3748     where "ddd" means "one or more decimal digits".
3749    
3750     3. Fixed serious bug. If a pattern had a back reference, but the call to
3751     pcre_exec() didn't supply a large enough ovector to record the related
3752     identifying subpattern, the match always failed. PCRE now remembers the number
3753     of the largest back reference, and gets some temporary memory in which to save
3754     the offsets during matching if necessary, in order to ensure that
3755     backreferences always work.
3756    
3757     4. Increased the compatibility with Perl in a number of ways:
3758    
3759     (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3760     to request this handling. The option can be set at compile or exec time.
3761    
3762     (b) $ matches before a terminating newline by default; an option
3763     PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3764     mode). The option can be set at compile or exec time.
3765    
3766     (c) The handling of \ followed by a digit other than 0 is now supposed to be
3767     the same as Perl's. If the decimal number it represents is less than 10
3768     or there aren't that many previous left capturing parentheses, an octal
3769     escape is read. Inside a character class, it's always an octal escape,
3770     even if it is a single digit.
3771    
3772     (d) An escaped but undefined alphabetic character is taken as a literal,
3773     unless PCRE_EXTRA is set. Currently this just reserves the remaining
3774     escapes.
3775    
3776     (e) {0} is now permitted. (The previous item is removed from the compiled
3777     pattern).
3778    
3779     5. Changed all the names of code files so that the basic parts are no longer
3780     than 10 characters, and abolished the teeny "globals.c" file.
3781    
3782     6. Changed the handling of character classes; they are now done with a 32-byte
3783     bit map always.
3784