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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 530 - (hide annotations) (download)
Tue Jun 1 13:42:06 2010 UTC (4 years, 3 months ago) by ph10
File size: 188064 byte(s)
Added a lot of (int) casts to avoid compiler warnings in systems where      
size_t is 64-bit.

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