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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 499 - (hide annotations) (download)
Wed Mar 3 20:09:39 2010 UTC (4 years, 1 month ago) by ph10
File size: 183343 byte(s)
Add some checks for the eint vector size and the list of compile-time error 
texts.

1 nigel 3 ChangeLog for PCRE
2     ------------------
3    
4 ph10 491 Version 8.02 01-Mar-2010
5     ------------------------
6    
7     1. The Unicode data tables have been updated to Unicode 5.2.0.
8    
9 ph10 492 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
10     configured.
11 ph10 493
12     3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
13     original author of that file, following a query about its status.
14 ph10 494
15     4. On systems that do not have stdint.h (e.g. Solaris), check for and include
16     inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
17 ph10 496
18     5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
19     quantifier applied to a forward-referencing subroutine call, could compile
20     incorrect code or give the error "internal error: previously-checked
21     referenced subpattern not found".
22 ph10 497
23     6. Both MS Visual Studio and Symbian OS have problems with initializing
24     variables to point to external functions. For these systems, therefore,
25     pcre_malloc etc. are now initialized to local functions that call the
26     relevant global functions.
27 ph10 498
28     7. There were two entries missing in the vectors called coptable and poptable
29     in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
30     I've fixed the data, and added a kludgy way of testing at compile time that
31     the lengths are correct (equal to the number of opcodes).
32 ph10 499
33     8. Following on from 7, I added a similar kludge to check the length of the
34     eint vector in pcreposix.c.
35    
36     9. Error texts for pcre_compile() are held as one long string to avoid too
37     much relocation at load time. To find a text, the string is searched,
38     counting zeros. There was no check for running off the end of the string,
39     which could happen if a new error number was added without updating the
40     string.
41 ph10 491
42 ph10 492
43 ph10 489 Version 8.01 19-Jan-2010
44     ------------------------
45 ph10 471
46     1. If a pattern contained a conditional subpattern with only one branch (in
47 ph10 489 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
48 ph10 473 computed the wrong minimum data length (which is of course zero for such
49 ph10 489 subpatterns). This could cause incorrect "no match" results.
50 ph10 473
51     2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
52 ph10 472 the pattern is reset in the first branch, pcre_compile() failed with
53     "internal error: code overflow at offset...". This happened only when
54 ph10 473 the reset was to the original external option setting. (An optimization
55     abstracts leading options settings into an external setting, which was the
56 ph10 472 cause of this.)
57 ph10 471
58 ph10 473 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
59     of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
60     assertion pattern did not match (meaning that the assertion was true), it
61     was incorrectly treated as false if the SKIP had been reached during the
62     matching. This also applied to assertions used as conditions.
63 ph10 471
64 ph10 473 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
65     assertion subpattern, including such a pattern used as a condition,
66     unpredictable results occurred, instead of the error return
67     PCRE_ERROR_DFA_UITEM.
68 ph10 478
69 ph10 474 5. The C++ GlobalReplace function was not working like Perl for the special
70     situation when an empty string is matched. It now does the fancy magic
71 ph10 478 stuff that is necessary.
72    
73     6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
74 ph10 475 removed. (These were left over from very, very early versions of PCRE.)
75 ph10 478
76 ph10 475 7. Some cosmetic changes to the code to make life easier when compiling it
77     as part of something else:
78 ph10 478
79     (a) Change DEBUG to PCRE_DEBUG.
80    
81     (b) In pcre_compile(), rename the member of the "branch_chain" structure
82     called "current" as "current_branch", to prevent a collision with the
83 ph10 475 Linux macro when compiled as a kernel module.
84 ph10 478
85     (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
86     prevent a collision with the Linux macro when compiled as a kernel
87 ph10 475 module.
88 ph10 478
89 ph10 475 8. In pcre_compile() there are some checks for integer overflows that used to
90     cast potentially large values to (double). This has been changed to that
91 ph10 478 when building, a check for int64_t is made, and if it is found, it is used
92     instead, thus avoiding the use of floating point arithmetic. (There is no
93     other use of FP in PCRE.) If int64_t is not found, the fallback is to
94     double.
95    
96     9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
97 ph10 476 2005 (difference between two addresses compared to an unsigned value).
98 ph10 473
99 ph10 478 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
100     custom one, because of the following reported problem in Windows:
101 ph10 473
102 ph10 478 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
103 ph10 489 under Win32.
104 ph10 478 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
105 ph10 489 therefore missing the function definition.
106 ph10 478 - The compiler thus generates a "C" signature for the test function.
107     - The linker fails to find the "C" function.
108     - PCRE fails to configure if asked to do so against libbz2.
109 ph10 486
110 ph10 479 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
111     messages were output:
112 ph10 478
113 ph10 479 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
114     rerunning libtoolize, to keep the correct libtool macros in-tree.
115     Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
116 ph10 486
117 ph10 479 I have done both of these things.
118 ph10 486
119 ph10 480 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
120 ph10 486 most of the time, it *can* run out if it is given a pattern that contains a
121     runaway infinite recursion. I updated the discussion in the pcrestack man
122     page.
123    
124     13. Now that we have gone to the x.xx style of version numbers, the minor
125     version may start with zero. Using 08 or 09 is a bad idea because users
126     might check the value of PCRE_MINOR in their code, and 08 or 09 may be
127     interpreted as invalid octal numbers. I've updated the previous comment in
128     configure.ac, and also added a check that gives an error if 08 or 09 are
129 ph10 481 used.
130 ph10 486
131 ph10 482 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
132 ph10 483 causing partial matching to fail when the end of the subject matched \W
133     in a UTF-8 pattern where \W was quantified with a minimum of 3.
134 ph10 486
135 ph10 483 15. There were some discrepancies between the declarations in pcre_internal.h
136 ph10 486 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
137 ph10 483 their definitions. The declarations used "const uschar *" and the
138     definitions used USPTR. Even though USPTR is normally defined as "const
139     unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
140     reported that: "This difference in casting confuses some C++ compilers, for
141     example, SunCC recognizes above declarations as different functions and
142 ph10 486 generates broken code for hbpcre." I have changed the declarations to use
143 ph10 483 USPTR.
144 ph10 486
145     16. GNU libtool is named differently on some systems. The autogen.sh script now
146 ph10 484 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
147     (FreeBSD).
148 ph10 478
149 ph10 486 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
150     (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
151     comment: "Figure out how to create a longlong from a string: strtoll and
152     equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
153     instance, but it only takes 2 args instead of 3!"
154 ph10 479
155 ph10 488 18. A subtle bug concerned with back references has been fixed by a change of
156     specification, with a corresponding code fix. A pattern such as
157     ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
158     refers, was giving matches when it shouldn't. For example, xa=xaaa would
159     match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
160     same bug. Such groups have to be quantified to be useful, or contained
161     inside another quantified group. (If there's no repetition, the reference
162     can never match.) The problem arises because, having left the group and
163     moved on to the rest of the pattern, a later failure that backtracks into
164     the group uses the captured value from the final iteration of the group
165     rather than the correct earlier one. I have fixed this in PCRE by forcing
166     any group that contains a reference to itself to be an atomic group; that
167     is, there cannot be any backtracking into it once it has completed. This is
168     similar to recursive and subroutine calls.
169 ph10 486
170 ph10 488
171 ph10 469 Version 8.00 19-Oct-09
172 ph10 418 ----------------------
173    
174     1. The table for translating pcre_compile() error codes into POSIX error codes
175 ph10 461 was out-of-date, and there was no check on the pcre_compile() error code
176     being within the table. This could lead to an OK return being given in
177 ph10 418 error.
178 ph10 461
179     2. Changed the call to open a subject file in pcregrep from fopen(pathname,
180     "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
181     in a Windows environment.
182    
183 ph10 420 3. The pcregrep --count option prints the count for each file even when it is
184     zero, as does GNU grep. However, pcregrep was also printing all files when
185     --files-with-matches was added. Now, when both options are given, it prints
186     counts only for those files that have at least one match. (GNU grep just
187 ph10 461 prints the file name in this circumstance, but including the count seems
188     more useful - otherwise, why use --count?) Also ensured that the
189 ph10 420 combination -clh just lists non-zero counts, with no names.
190 ph10 461
191     4. The long form of the pcregrep -F option was incorrectly implemented as
192     --fixed_strings instead of --fixed-strings. This is an incompatible change,
193     but it seems right to fix it, and I didn't think it was worth preserving
194     the old behaviour.
195    
196     5. The command line items --regex=pattern and --regexp=pattern were not
197 ph10 422 recognized by pcregrep, which required --regex pattern or --regexp pattern
198 ph10 461 (with a space rather than an '='). The man page documented the '=' forms,
199 ph10 422 which are compatible with GNU grep; these now work.
200 ph10 461
201     6. No libpcreposix.pc file was created for pkg-config; there was just
202 ph10 423 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
203 ph10 461
204 ph10 425 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
205 ph10 461 when UCP support is not needed, by modifying the Python script that
206 ph10 425 generates it from Unicode data files. This should not matter if the module
207     is correctly used as a library, but I received one complaint about 50K of
208     unwanted data. My guess is that the person linked everything into his
209     program rather than using a library. Anyway, it does no harm.
210 ph10 461
211 ph10 426 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
212 ph10 461 was a minimum greater than 1 for a wide character in a possessive
213 ph10 431 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
214     which had an unlimited repeat of a nested, fixed maximum repeat of a wide
215     character. Chaos in the form of incorrect output or a compiling loop could
216     result.
217 ph10 461
218 ph10 426 9. The restrictions on what a pattern can contain when partial matching is
219 ph10 461 requested for pcre_exec() have been removed. All patterns can now be
220 ph10 426 partially matched by this function. In addition, if there are at least two
221 ph10 435 slots in the offset vector, the offset of the earliest inspected character
222     for the match and the offset of the end of the subject are set in them when
223 ph10 461 PCRE_ERROR_PARTIAL is returned.
224    
225 ph10 428 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
226     synonymous with PCRE_PARTIAL, for backwards compatibility, and
227     PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
228 ph10 462 and may be more useful for multi-segment matching.
229 ph10 461
230     11. Partial matching with pcre_exec() is now more intuitive. A partial match
231     used to be given if ever the end of the subject was reached; now it is
232     given only if matching could not proceed because another character was
233     needed. This makes a difference in some odd cases such as Z(*FAIL) with the
234     string "Z", which now yields "no match" instead of "partial match". In the
235     case of pcre_dfa_exec(), "no match" is given if every matching path for the
236     final character ended with (*FAIL).
237    
238 ph10 428 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
239 ph10 461 if the pattern had a "must contain" character that was already found in the
240 ph10 428 earlier partial match, unless partial matching was again requested. For
241     example, with the pattern /dog.(body)?/, the "must contain" character is
242     "g". If the first part-match was for the string "dog", restarting with
243 ph10 435 "sbody" failed. This bug has been fixed.
244 ph10 461
245     13. The string returned by pcre_dfa_exec() after a partial match has been
246     changed so that it starts at the first inspected character rather than the
247     first character of the match. This makes a difference only if the pattern
248     starts with a lookbehind assertion or \b or \B (\K is not supported by
249     pcre_dfa_exec()). It's an incompatible change, but it makes the two
250 ph10 435 matching functions compatible, and I think it's the right thing to do.
251 ph10 461
252 ph10 435 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
253 ph10 461 so that the demonstration program is easily available in environments where
254     PCRE has not been installed from source.
255    
256 ph10 435 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
257 ph10 430 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
258     library.
259 ph10 461
260 ph10 435 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
261 ph10 432 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
262 ph10 461 is not the first non-POSIX option to be added. Clearly some people find
263 ph10 433 these options useful.
264 ph10 461
265     17. If a caller to the POSIX matching function regexec() passes a non-zero
266 ph10 438 value for nmatch with a NULL value for pmatch, the value of
267 ph10 461 nmatch is forced to zero.
268    
269 ph10 437 18. RunGrepTest did not have a test for the availability of the -u option of
270 ph10 461 the diff command, as RunTest does. It now checks in the same way as
271 ph10 437 RunTest, and also checks for the -b option.
272 ph10 461
273 ph10 438 19. If an odd number of negated classes containing just a single character
274     interposed, within parentheses, between a forward reference to a named
275 ph10 461 subpattern and the definition of the subpattern, compilation crashed with
276     an internal error, complaining that it could not find the referenced
277 ph10 438 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
278 ph10 461 [The bug was that it was starting one character too far in when skipping
279     over the character class, thus treating the ] as data rather than
280     terminating the class. This meant it could skip too much.]
281    
282 ph10 442 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
283 ph10 461 /g option in pcretest when the pattern contains \K, which makes it possible
284 ph10 442 to have an empty string match not at the start, even when the pattern is
285 ph10 461 anchored. Updated pcretest and pcredemo to use this option.
286    
287 ph10 446 21. If the maximum number of capturing subpatterns in a recursion was greater
288 ph10 461 than the maximum at the outer level, the higher number was returned, but
289     with unset values at the outer level. The correct (outer level) value is
290 ph10 446 now given.
291 ph10 461
292 ph10 447 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
293     PCRE did not set those parentheses (unlike Perl). I have now found a way to
294     make it do so. The string so far is captured, making this feature
295     compatible with Perl.
296 ph10 461
297     23. The tests have been re-organized, adding tests 11 and 12, to make it
298 ph10 448 possible to check the Perl 5.10 features against Perl 5.10.
299 ph10 461
300 ph10 454 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
301 ph10 461 pattern matches a fixed length string. PCRE did not allow this; now it
302     does. Neither allows recursion.
303    
304     25. I finally figured out how to implement a request to provide the minimum
305     length of subject string that was needed in order to match a given pattern.
306     (It was back references and recursion that I had previously got hung up
307     on.) This code has now been added to pcre_study(); it finds a lower bound
308 ph10 455 to the length of subject needed. It is not necessarily the greatest lower
309     bound, but using it to avoid searching strings that are too short does give
310     some useful speed-ups. The value is available to calling programs via
311     pcre_fullinfo().
312 ph10 461
313 ph10 455 26. While implementing 25, I discovered to my embarrassment that pcretest had
314     not been passing the result of pcre_study() to pcre_dfa_exec(), so the
315     study optimizations had never been tested with that matching function.
316     Oops. What is worse, even when it was passed study data, there was a bug in
317     pcre_dfa_exec() that meant it never actually used it. Double oops. There
318     were also very few tests of studied patterns with pcre_dfa_exec().
319 ph10 461
320 ph10 457 27. If (?| is used to create subpatterns with duplicate numbers, they are now
321     allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
322     on the other side of the coin, they are no longer allowed to have different
323     names, because these cannot be distinguished in PCRE, and this has caused
324     confusion. (This is a difference from Perl.)
325 ph10 461
326     28. When duplicate subpattern names are present (necessarily with different
327     numbers, as required by 27 above), and a test is made by name in a
328     conditional pattern, either for a subpattern having been matched, or for
329     recursion in such a pattern, all the associated numbered subpatterns are
330 ph10 459 tested, and the overall condition is true if the condition is true for any
331     one of them. This is the way Perl works, and is also more like the way
332     testing by number works.
333 ph10 418
334 ph10 461
335 ph10 412 Version 7.9 11-Apr-09
336 ph10 376 ---------------------
337    
338 ph10 392 1. When building with support for bzlib/zlib (pcregrep) and/or readline
339 ph10 376 (pcretest), all targets were linked against these libraries. This included
340     libpcre, libpcreposix, and libpcrecpp, even though they do not use these
341     libraries. This caused unwanted dependencies to be created. This problem
342 ph10 392 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
343 ph10 376 pcretest is linked with readline.
344 ph10 392
345 ph10 376 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
346     "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
347     moved outside it again, because FALSE and TRUE are already defined in AIX,
348     but BOOL is not.
349 ph10 392
350     3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
351 ph10 409 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
352 ph10 392
353 ph10 376 4. The pcregrep documentation said spaces were inserted as well as colons (or
354     hyphens) following file names and line numbers when outputting matching
355 ph10 392 lines. This is not true; no spaces are inserted. I have also clarified the
356 ph10 376 wording for the --colour (or --color) option.
357 ph10 392
358 ph10 378 5. In pcregrep, when --colour was used with -o, the list of matching strings
359     was not coloured; this is different to GNU grep, so I have changed it to be
360     the same.
361 ph10 392
362 ph10 378 6. When --colo(u)r was used in pcregrep, only the first matching substring in
363 ph10 392 each matching line was coloured. Now it goes on to look for further matches
364     of any of the test patterns, which is the same behaviour as GNU grep.
365    
366     7. A pattern that could match an empty string could cause pcregrep to loop; it
367     doesn't make sense to accept an empty string match in pcregrep, so I have
368 ph10 379 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
369     seems to be how GNU grep behaves.
370 ph10 392
371 ph10 380 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
372 ph10 392 start or after a newline", because the conditional assertion was not being
373     correctly handled. The rule now is that both the assertion and what follows
374     in the first alternative must satisfy the test.
375    
376 ph10 405 9. If auto-callout was enabled in a pattern with a conditional group whose
377 ph10 399 condition was an assertion, PCRE could crash during matching, both with
378     pcre_exec() and pcre_dfa_exec().
379 ph10 392
380     10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
381     used for matching.
382    
383     11. Unicode property support in character classes was not working for
384 ph10 385 characters (bytes) greater than 127 when not in UTF-8 mode.
385 ph10 388
386 ph10 392 12. Added the -M command line option to pcretest.
387    
388 ph10 388 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
389 ph10 389
390     15. Added the PCRE_NO_START_OPTIMIZE match-time option.
391 ph10 390
392 ph10 392 16. Added comments and documentation about mis-use of no_arg in the C++
393 ph10 390 wrapper.
394 ph10 392
395 ph10 391 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
396 ph10 392 from Martin Jerabek that uses macro names for all relevant character and
397 ph10 391 string constants.
398 ph10 405
399 ph10 393 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
400     SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
401 ph10 405 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
402     these, but not everybody uses configure.
403    
404     19. A conditional group that had only one branch was not being correctly
405     recognized as an item that could match an empty string. This meant that an
406     enclosing group might also not be so recognized, causing infinite looping
407     (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
408     with the subject "ab", where knowledge that the repeated group can match
409 ph10 396 nothing is needed in order to break the loop.
410 ph10 405
411 ph10 397 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
412 ph10 405 exec(), but without supplying a callout function, matching went wrong.
413    
414     21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
415     leak if the size of the offset vector was greater than 30. When the vector
416     is smaller, the saved offsets during recursion go onto a local stack
417     vector, but for larger vectors malloc() is used. It was failing to free
418     when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
419 ph10 400 error, in fact).
420 ph10 405
421 ph10 402 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
422 ph10 405 heapframe that is used only when UTF-8 support is enabled. This caused no
423     problem, but was untidy.
424 ph10 376
425 ph10 405 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
426     CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
427     included within another project.
428 ph10 411
429 ph10 407 24. Steven Van Ingelgem's patches to add more options to the CMake support,
430     slightly modified by me:
431 ph10 411
432     (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
433     not building pcregrep.
434    
435 ph10 407 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
436 ph10 411 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
437    
438 ph10 408 25. Forward references, both numeric and by name, in patterns that made use of
439     duplicate group numbers, could behave incorrectly or give incorrect errors,
440     because when scanning forward to find the reference group, PCRE was not
441     taking into account the duplicate group numbers. A pattern such as
442     ^X(?3)(a)(?|(b)|(q))(Y) is an example.
443 ph10 411
444     26. Changed a few more instances of "const unsigned char *" to USPTR, making
445     the feature of a custom pointer more persuasive (as requested by a user).
446 ph10 416
447 ph10 411 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
448 ph10 416 pcretest.c, inside #ifndefs, because it seems they are sometimes already
449     pre-defined.
450 ph10 392
451 ph10 416 28. Added support for (*UTF8) at the start of a pattern.
452    
453 ph10 413 29. Arrange for flags added by the "release type" setting in CMake to be shown
454     in the configuration summary.
455 ph10 405
456 ph10 411
457 ph10 374 Version 7.8 05-Sep-08
458 ph10 349 ---------------------
459    
460     1. Replaced UCP searching code with optimized version as implemented for Ad
461     Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
462 ph10 351 stage table and inline lookup instead of a function, giving speed ups of 2
463     to 5 times on some simple patterns that I tested. Permission was given to
464     distribute the MultiStage2.py script that generates the tables (it's not in
465     the tarball, but is in the Subversion repository).
466 ph10 349
467 ph10 351 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
468     scripts.
469 ph10 358
470 ph10 353 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
471     a group with a zero qualifier. The result of the study could be incorrect,
472 ph10 358 or the function might crash, depending on the pattern.
473    
474     4. Caseless matching was not working for non-ASCII characters in back
475 ph10 354 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
476 ph10 358 It now works when Unicode Property Support is available.
477    
478 ph10 355 5. In pcretest, an escape such as \x{de} in the data was always generating
479     a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
480     non-UTF-8 mode. If the value is greater than 255, it gives a warning about
481 ph10 358 truncation.
482 ph10 349
483 ph10 358 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
484    
485 ph10 357 7. Added two (int) casts to pcregrep when printing the difference of two
486     pointers, in case they are 64-bit values.
487 ph10 371
488     8. Added comments about Mac OS X stack usage to the pcrestack man page and to
489     test 2 if it fails.
490    
491 ph10 359 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
492 ph10 371 and a #define of that name to empty if it is not externally set. This is to
493     allow users of MSVC to set it if necessary.
494    
495     10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
496 ph10 359 the convenience functions in the pcre_get.c source file.
497 ph10 371
498 ph10 360 11. An option change at the start of a pattern that had top-level alternatives
499     could cause overwriting and/or a crash. This command provoked a crash in
500 ph10 371 some environments:
501    
502     printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
503    
504 ph10 360 This potential security problem was recorded as CVE-2008-2371.
505 ph10 371
506 ph10 361 12. For a pattern where the match had to start at the beginning or immediately
507     after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
508 ph10 371 pcre_dfa_exec() could read past the end of the passed subject if there was
509 ph10 361 no match. To help with detecting such bugs (e.g. with valgrind), I modified
510     pcretest so that it places the subject at the end of its malloc-ed buffer.
511 ph10 371
512 ph10 364 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
513 ph10 371 exec() might read past the end of the data buffer in UTF-8 mode.
514    
515 ph10 364 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
516 ph10 371 the data contained the byte 0x85 as part of a UTF-8 character within its
517     first line. This applied both to normal and DFA matching.
518    
519 ph10 366 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
520 ph10 371 /^[^d]*?$/8 failed to match "abc".
521 ph10 351
522 ph10 371 16. Added a missing copyright notice to pcrecpp_internal.h.
523    
524     17. Make it more clear in the documentation that values returned from
525 ph10 368 pcre_exec() in ovector are byte offsets, not character counts.
526 ph10 357
527 ph10 371 18. Tidied a few places to stop certain compilers from issuing warnings.
528 ph10 368
529 ph10 373 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
530     supplied by Stefan Weber. I made a further small update for 7.8 because
531 ph10 374 there is a change of source arrangements: the pcre_searchfuncs.c module is
532     replaced by pcre_ucd.c.
533 ph10 371
534 ph10 373
535 ph10 347 Version 7.7 07-May-08
536 ph10 321 ---------------------
537    
538     1. Applied Craig's patch to sort out a long long problem: "If we can't convert
539 ph10 345 a string to a long long, pretend we don't even have a long long." This is
540 ph10 321 done by checking for the strtoq, strtoll, and _strtoi64 functions.
541 ph10 345
542 ph10 322 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
543     pre-7.6 versions, which defined a global no_arg variable instead of putting
544 ph10 332 it in the RE class. (See also #8 below.)
545 ph10 345
546     3. Remove a line of dead code, identified by coverity and reported by Nuno
547     Lopes.
548    
549 ph10 324 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
550 ph10 321
551 ph10 324 (1) The include/exclude patterns were being applied to the whole pathnames
552 ph10 345 of files, instead of just to the final components.
553    
554 ph10 324 (2) If there was more than one level of directory, the subdirectories were
555     skipped unless they satisfied the include/exclude conditions. This is
556     inconsistent with GNU grep (and could even be seen as contrary to the
557     pcregrep specification - which I improved to make it absolutely clear).
558     The action now is always to scan all levels of directory, and just
559     apply the include/exclude patterns to regular files.
560 ph10 345
561 ph10 325 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
562 ph10 345 --exclude_dir in the tests to avoid scanning .svn directories.
563    
564 ph10 326 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
565 ph10 345 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
566 ph10 326 doesn't support NULs in patterns.
567 ph10 345
568     7. Added some missing "const"s to declarations of static tables in
569     pcre_compile.c and pcre_dfa_exec.c.
570    
571 ph10 328 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
572 ph10 329 caused by fix #2 above. (Subsequently also a second patch to fix the
573 ph10 345 first patch. And a third patch - this was a messy problem.)
574 ph10 321
575 ph10 345 9. Applied Craig's patch to remove the use of push_back().
576    
577 ph10 332 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
578     matching function regexec().
579 ph10 345
580     11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
581     which, however, unlike Perl's \g{...}, are subroutine calls, not back
582 ph10 333 references. PCRE supports relative numbers with this syntax (I don't think
583     Oniguruma does).
584 ph10 345
585     12. Previously, a group with a zero repeat such as (...){0} was completely
586 ph10 335 omitted from the compiled regex. However, this means that if the group
587     was called as a subroutine from elsewhere in the pattern, things went wrong
588 ph10 345 (an internal error was given). Such groups are now left in the compiled
589     pattern, with a new opcode that causes them to be skipped at execution
590 ph10 335 time.
591 ph10 345
592 ph10 341 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
593     to the way PCRE behaves:
594 ph10 345
595 ph10 336 (a) A lone ] character is dis-allowed (Perl treats it as data).
596 ph10 345
597     (b) A back reference to an unmatched subpattern matches an empty string
598 ph10 336 (Perl fails the current match path).
599 ph10 345
600 ph10 341 (c) A data ] in a character class must be notated as \] because if the
601 ph10 345 first data character in a class is ], it defines an empty class. (In
602 ph10 341 Perl it is not possible to have an empty class.) The empty class []
603     never matches; it forces failure and is equivalent to (*FAIL) or (?!).
604     The negative empty class [^] matches any one character, independently
605     of the DOTALL setting.
606 ph10 345
607     14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
608 ph10 340 non-existent subpattern following a character class starting with ']' and
609     containing () gave an internal compiling error instead of "reference to
610     non-existent subpattern". Fortunately, when the pattern did exist, the
611 ph10 345 compiled code was correct. (When scanning forwards to check for the
612 ph10 340 existencd of the subpattern, it was treating the data ']' as terminating
613 ph10 345 the class, so got the count wrong. When actually compiling, the reference
614 ph10 340 was subsequently set up correctly.)
615 ph10 345
616 ph10 341 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
617 ph10 345 it was being rejected as not supported by pcre_dfa_exec(), even though
618     other assertions are supported. I have made pcre_dfa_exec() support
619     (*FAIL).
620    
621 ph10 342 16. The implementation of 13c above involved the invention of a new opcode,
622 ph10 345 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
623     cannot be changed at match time, I realized I could make a small
624     improvement to matching performance by compiling OP_ALLANY instead of
625     OP_ANY for "." when DOTALL was set, and then removing the runtime tests
626     on the OP_ANY path.
627    
628     17. Compiling pcretest on Windows with readline support failed without the
629 ph10 343 following two fixes: (1) Make the unistd.h include conditional on
630 ph10 345 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
631    
632 ph10 344 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
633 ph10 345 ncurses library to be included for pcretest when ReadLine support is
634     requested, but also to allow for it to be overridden. This patch came from
635     Daniel Bergström.
636 ph10 347
637 ph10 346 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
638 ph10 347 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
639     any errors with the current Unicode tables. Thanks to Peter Kankowski for
640     spotting this.
641 ph10 324
642 ph10 332
643 ph10 319 Version 7.6 28-Jan-08
644 ph10 300 ---------------------
645    
646 ph10 302 1. A character class containing a very large number of characters with
647     codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
648     overflow.
649 ph10 309
650     2. Patch to cut out the "long long" test in pcrecpp_unittest when
651     HAVE_LONG_LONG is not defined.
652    
653 ph10 303 3. Applied Christian Ehrlicher's patch to update the CMake build files to
654 ph10 304 bring them up to date and include new features. This patch includes:
655 ph10 309
656 ph10 304 - Fixed PH's badly added libz and libbz2 support.
657     - Fixed a problem with static linking.
658 ph10 312 - Added pcredemo. [But later removed - see 7 below.]
659 ph10 304 - Fixed dftables problem and added an option.
660     - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
661     HAVE_LONG_LONG.
662     - Added readline support for pcretest.
663 ph10 309 - Added an listing of the option settings after cmake has run.
664    
665 ph10 314 4. A user submitted a patch to Makefile that makes it easy to create
666     "pcre.dll" under mingw when using Configure/Make. I added stuff to
667     Makefile.am that cause it to include this special target, without
668     affecting anything else. Note that the same mingw target plus all
669     the other distribution libraries and programs are now supported
670     when configuring with CMake (see 6 below) instead of with
671     Configure/Make.
672 ph10 309
673 ph10 308 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
674     This is an attempt to solve the reported problem "pcrecpp::no_arg is not
675 ph10 309 exported in the Windows port". It has not yet been confirmed that the patch
676     solves the problem, but it does no harm.
677 ph10 313
678 ph10 311 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
679 ph10 319 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
680 ph10 314 with CMake, and also correct the comment about stack recursion.
681 ph10 313
682 ph10 312 7. Remove the automatic building of pcredemo from the ./configure system and
683 ph10 313 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
684     of a program that users should build themselves after PCRE is installed, so
685     building it automatically is not really right. What is more, it gave
686 ph10 312 trouble in some build environments.
687 ph10 300
688 ph10 319 8. Further tidies to CMakeLists.txt from Sheri and Christian.
689 ph10 308
690 ph10 319
691 ph10 298 Version 7.5 10-Jan-08
692 ph10 263 ---------------------
693    
694     1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
695     values in parens when parsing an RE using the C++ wrapper."
696 ph10 286
697 ph10 264 2. Negative specials like \S did not work in character classes in UTF-8 mode.
698     Characters greater than 255 were excluded from the class instead of being
699     included.
700 ph10 286
701     3. The same bug as (2) above applied to negated POSIX classes such as
702 ph10 265 [:^space:].
703 ph10 286
704 ph10 267 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
705     defined or documented. It seems to have been a typo for PCRE_STATIC, so
706 ph10 286 I have changed it.
707    
708 ph10 268 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
709 ph10 286 first named subpattern) and a construct such as (?&a) would reference the
710     first named subpattern whose name started with "a" (in other words, the
711 ph10 272 length check was missing). Both these problems are fixed. "Subpattern name
712     expected" is now given for (?&) (a zero-length name), and this patch also
713     makes it give the same error for \k'' (previously it complained that that
714     was a reference to a non-existent subpattern).
715 ph10 286
716 ph10 269 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
717 ph10 286 this is right because (?- can be followed by option settings as well as by
718 ph10 269 digits. I have, however, made the messages clearer.
719 ph10 286
720 ph10 270 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
721     than the number used in the conditional) now cause a compile-time error.
722     This is actually not compatible with Perl, which accepts such patterns, but
723     treats the conditional as always being FALSE (as PCRE used to), but it
724     seems to me that giving a diagnostic is better.
725 ph10 286
726 ph10 274 8. Change "alphameric" to the more common word "alphanumeric" in comments
727 ph10 275 and messages.
728 ph10 286
729     9. Fix two occurrences of "backslash" in comments that should have been
730     "backspace".
731    
732     10. Remove two redundant lines of code that can never be obeyed (their function
733     was moved elsewhere).
734    
735     11. The program that makes PCRE's Unicode character property table had a bug
736     which caused it to generate incorrect table entries for sequences of
737 ph10 277 characters that have the same character type, but are in different scripts.
738 ph10 286 It amalgamated them into a single range, with the script of the first of
739 ph10 277 them. In other words, some characters were in the wrong script. There were
740     thirteen such cases, affecting characters in the following ranges:
741 ph10 286
742 ph10 277 U+002b0 - U+002c1
743     U+0060c - U+0060d
744 ph10 286 U+0061e - U+00612
745 ph10 277 U+0064b - U+0065e
746     U+0074d - U+0076d
747     U+01800 - U+01805
748     U+01d00 - U+01d77
749     U+01d9b - U+01dbf
750     U+0200b - U+0200f
751     U+030fc - U+030fe
752     U+03260 - U+0327f
753     U+0fb46 - U+0fbb1
754     U+10450 - U+1049d
755 ph10 286
756 ph10 279 12. The -o option (show only the matching part of a line) for pcregrep was not
757     compatible with GNU grep in that, if there was more than one match in a
758     line, it showed only the first of them. It now behaves in the same way as
759     GNU grep.
760 ph10 286
761     13. If the -o and -v options were combined for pcregrep, it printed a blank
762     line for every non-matching line. GNU grep prints nothing, and pcregrep now
763     does the same. The return code can be used to tell if there were any
764     non-matching lines.
765    
766 ph10 289 14. Added --file-offsets and --line-offsets to pcregrep.
767    
768     15. The pattern (?=something)(?R) was not being diagnosed as a potentially
769 ph10 282 infinitely looping recursion. The bug was that positive lookaheads were not
770 ph10 286 being skipped when checking for a possible empty match (negative lookaheads
771     and both kinds of lookbehind were skipped).
772    
773 ph10 289 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
774 ph10 284 inclusion of <windows.h> to before rather than after the definition of
775     INVALID_FILE_ATTRIBUTES (patch from David Byron).
776 ph10 263
777 ph10 289 17. Specifying a possessive quantifier with a specific limit for a Unicode
778 ph10 286 character property caused pcre_compile() to compile bad code, which led at
779     runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
780     are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
781     caused the error; without that there was no problem.
782 ph10 263
783 ph10 289 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
784 ph10 286
785 ph10 289 19. Added --enable-pcretest-libreadline.
786 ph10 286
787 ph10 289 20. In pcrecpp.cc, the variable 'count' was incremented twice in
788 ph10 288 RE::GlobalReplace(). As a result, the number of replacements returned was
789 ph10 298 double what it should be. I removed one of the increments, but Craig sent a
790     later patch that removed the other one (the right fix) and added unit tests
791     that check the return values (which was not done before).
792 ph10 292
793 ph10 291 21. Several CMake things:
794 ph10 287
795 ph10 291 (1) Arranged that, when cmake is used on Unix, the libraries end up with
796 ph10 292 the names libpcre and libpcreposix, not just pcre and pcreposix.
797 ph10 288
798 ph10 292 (2) The above change means that pcretest and pcregrep are now correctly
799     linked with the newly-built libraries, not previously installed ones.
800 ph10 291
801 ph10 292 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
802 ph10 298
803     22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
804     crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
805     UTF-8 newline character). The key issue is that the pattern starts .*;
806     this means that the match must be either at the beginning, or after a
807     newline. The bug was in the code for advancing after a failed match and
808     checking that the new position followed a newline. It was not taking
809 ph10 294 account of UTF-8 characters correctly.
810 ph10 298
811     23. PCRE was behaving differently from Perl in the way it recognized POSIX
812     character classes. PCRE was not treating the sequence [:...:] as a
813     character class unless the ... were all letters. Perl, however, seems to
814     allow any characters between [: and :], though of course it rejects as
815     unknown any "names" that contain non-letters, because all the known class
816     names consist only of letters. Thus, Perl gives an error for [[:1234:]],
817     for example, whereas PCRE did not - it did not recognize a POSIX character
818     class. This seemed a bit dangerous, so the code has been changed to be
819     closer to Perl. The behaviour is not identical to Perl, because PCRE will
820     diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
821     treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
822     Perl does, and where it didn't before.
823    
824 ph10 296 24. Rewrite so as to remove the single use of %n from pcregrep because in some
825     Windows environments %n is disabled by default.
826 ph10 292
827    
828 ph10 260 Version 7.4 21-Sep-07
829 ph10 230 ---------------------
830    
831     1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
832 ph10 231 means that a class such as [\s] counted as "explicit reference to CR or
833     LF". That isn't really right - the whole point of the change was to try to
834     help when there was an actual mention of one of the two characters. So now
835     the change happens only if \r or \n (or a literal CR or LF) character is
836 ph10 230 encountered.
837 ph10 231
838     2. The 32-bit options word was also used for 6 internal flags, but the numbers
839     of both had grown to the point where there were only 3 bits left.
840     Fortunately, there was spare space in the data structure, and so I have
841     moved the internal flags into a new 16-bit field to free up more option
842 ph10 230 bits.
843 ph10 231
844     3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
845     but did not set the internal JCHANGED flag - either of these is enough to
846     control the way the "get" function works - but the PCRE_INFO_JCHANGED
847     facility is supposed to tell if (?J) was ever used, so now (?J) at the
848 ph10 230 start sets both bits.
849    
850 ph10 231 4. Added options (at build time, compile time, exec time) to change \R from
851     matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
852 ph10 230
853 ph10 243 5. doc/pcresyntax.html was missing from the distribution.
854    
855     6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
856 ph10 233 compatibility, even though it is no longer used.
857 ph10 243
858     7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
859 ph10 254 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
860 ph10 259 windows.h file is present (where different names are used). [This was
861 ph10 257 reversed later after testing - see 16 below.]
862 ph10 243
863     8. Changed all #include <config.h> to #include "config.h". There were also
864 ph10 236 some further <pcre.h> cases that I changed to "pcre.h".
865 ph10 243
866 ph10 239 9. When pcregrep was used with the --colour option, it missed the line ending
867 ph10 243 sequence off the lines that it output.
868    
869     10. It was pointed out to me that arrays of string pointers cause lots of
870     relocations when a shared library is dynamically loaded. A technique of
871     using a single long string with a table of offsets can drastically reduce
872     these. I have refactored PCRE in four places to do this. The result is
873 ph10 240 dramatic:
874 ph10 243
875 ph10 240 Originally: 290
876     After changing UCP table: 187
877 ph10 243 After changing error message table: 43
878 ph10 240 After changing table of "verbs" 36
879     After changing table of Posix names 22
880 ph10 243
881 ph10 240 Thanks to the folks working on Gregex for glib for this insight.
882 ph10 247
883 ph10 244 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
884 ph10 247 unicode-properties was also set.
885    
886 ph10 248 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
887 ph10 231
888 ph10 253 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
889     checked only for CRLF.
890 ph10 233
891 ph10 259 14. Added casts to pcretest.c to avoid compiler warnings.
892    
893 ph10 256 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
894 ph10 253
895 ph10 257 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
896 ph10 259 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
897     entirely. This removes changes made in 7 above.
898 ph10 256
899 ph10 261 17. The CMake files have been updated, and there is now more information about
900     building with CMake in the NON-UNIX-USE document.
901 ph10 257
902 ph10 261
903 ph10 228 Version 7.3 28-Aug-07
904 ph10 157 ---------------------
905    
906 ph10 189 1. In the rejigging of the build system that eventually resulted in 7.1, the
907     line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
908     brackets there is not right, since it causes compilers to look for an
909     installed pcre.h, not the version that is in the source that is being
910     compiled (which of course may be different). I have changed it back to:
911 ph10 197
912 ph10 189 #include "pcre.h"
913 ph10 197
914     I have a vague recollection that the change was concerned with compiling in
915     different directories, but in the new build system, that is taken care of
916     by the VPATH setting the Makefile.
917    
918 ph10 190 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
919     when the subject happened to end in the byte 0x85 (e.g. if the last
920     character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
921     characters but of course it shouldn't be taken as a newline when it is part
922     of another character. The bug was that, for an unlimited repeat of . in
923     not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
924     characters when looking for a newline.
925 ph10 197
926     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
927    
928     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
929     in debug output.
930    
931 ph10 201 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
932     long printing in the pcrecpp unittest when running under MinGW.
933 ph10 189
934 ph10 197 6. ESC_K was left out of the EBCDIC table.
935 ph10 189
936 ph10 197 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
937     parentheses; I made it 1000, which seemed large enough. Unfortunately, the
938     limit also applies to "virtual nesting" when a pattern is recursive, and in
939     this case 1000 isn't so big. I have been able to remove this limit at the
940     expense of backing off one optimization in certain circumstances. Normally,
941     when pcre_exec() would call its internal match() function recursively and
942     immediately return the result unconditionally, it uses a "tail recursion"
943     feature to save stack. However, when a subpattern that can match an empty
944     string has an unlimited repetition quantifier, it no longer makes this
945     optimization. That gives it a stack frame in which to save the data for
946     checking that an empty string has been matched. Previously this was taken
947     from the 1000-entry workspace that had been reserved. So now there is no
948     explicit limit, but more stack is used.
949 ph10 189
950 ph10 200 8. Applied Daniel's patches to solve problems with the import/export magic
951     syntax that is required for Windows, and which was going wrong for the
952     pcreposix and pcrecpp parts of the library. These were overlooked when this
953     problem was solved for the main library.
954 ph10 197
955 ph10 202 9. There were some crude static tests to avoid integer overflow when computing
956     the size of patterns that contain repeated groups with explicit upper
957     limits. As the maximum quantifier is 65535, the maximum group length was
958     set at 30,000 so that the product of these two numbers did not overflow a
959     32-bit integer. However, it turns out that people want to use groups that
960     are longer than 30,000 bytes (though not repeat them that many times).
961     Change 7.0/17 (the refactoring of the way the pattern size is computed) has
962     made it possible to implement the integer overflow checks in a much more
963     dynamic way, which I have now done. The artificial limitation on group
964     length has been removed - we now have only the limit on the total length of
965     the compiled pattern, which depends on the LINK_SIZE setting.
966 ph10 208
967     10. Fixed a bug in the documentation for get/copy named substring when
968     duplicate names are permitted. If none of the named substrings are set, the
969     functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
970     empty string.
971    
972     11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
973     instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
974     because the ] is interpreted as the first data character and the
975     terminating ] is not found. PCRE has been made compatible with Perl in this
976     regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
977     cause memory overwriting.
978    
979 ph10 206 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
980     string has been matched (to stop an infinite loop). It was not recognizing
981 ph10 208 a conditional subpattern that could match an empty string if that
982 ph10 206 subpattern was within another subpattern. For example, it looped when
983 ph10 208 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
984 ph10 206 condition was not nested. This bug has been fixed.
985 ph10 208
986 ph10 207 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
987     past the start of the subject in the presence of bytes with the top bit
988     set, for example "\x8aBCD".
989 ph10 211
990 ph10 210 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
991     (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
992 ph10 200
993 ph10 211 14. Optimized (?!) to (*FAIL).
994 ph10 202
995 ph10 212 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
996     This restricts code points to be within the range 0 to 0x10FFFF, excluding
997     the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
998     full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
999     does: it's just the validity check that is more restrictive.
1000 ph10 220
1001     16. Inserted checks for integer overflows during escape sequence (backslash)
1002     processing, and also fixed erroneous offset values for syntax errors during
1003     backslash processing.
1004    
1005 ph10 214 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1006 ph10 220 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1007    
1008 ph10 215 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1009     caused an overrun.
1010 ph10 220
1011     19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1012     something other than just ASCII characters) inside a group that had an
1013     unlimited repeat caused a loop at compile time (while checking to see
1014     whether the group could match an empty string).
1015    
1016 ph10 217 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1017     [\P{Any}] did so. (Error in the code for printing property names.)
1018 ph10 210
1019 ph10 220 21. An orphan \E inside a character class could cause a crash.
1020    
1021     22. A repeated capturing bracket such as (A)? could cause a wild memory
1022 ph10 218 reference during compilation.
1023 ph10 220
1024     23. There are several functions in pcre_compile() that scan along a compiled
1025     expression for various reasons (e.g. to see if it's fixed length for look
1026 ph10 218 behind). There were bugs in these functions when a repeated \p or \P was
1027 ph10 220 present in the pattern. These operators have additional parameters compared
1028     with \d, etc, and these were not being taken into account when moving along
1029 ph10 218 the compiled data. Specifically:
1030 ph10 220
1031     (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1032     length.
1033    
1034     (b) An item such as \pL+ within a repeated group could cause crashes or
1035 ph10 218 loops.
1036 ph10 220
1037 ph10 218 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1038 ph10 220 "reference to non-existent subpattern" error.
1039    
1040 ph10 221 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1041    
1042 ph10 219 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1043 ph10 220 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1044 ph10 211
1045 ph10 222 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1046 ph10 218
1047 ph10 227 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1048 ph10 223 character were causing crashes (broken optimization).
1049 ph10 227
1050 ph10 224 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1051 ph10 227 \p or \P) caused a compile-time loop.
1052    
1053 ph10 226 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1054     break. For example, the unstudied pattern [\r\n]A does not match the string
1055     "\r\nA" because change 7.0/46 below moves the current point on by two
1056     characters after failing to match at the start. However, the pattern \nA
1057 ph10 227 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1058     the same is true. There doesn't seem any very clean way out of this, but
1059     what I have chosen to do makes the common cases work: PCRE now takes note
1060     of whether there can be an explicit match for \r or \n anywhere in the
1061     pattern, and if so, 7.0/46 no longer applies. As part of this change,
1062     there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1063     pattern has explicit CR or LF references.
1064 ph10 222
1065 ph10 227 29. Added (*CR) etc for changing newline setting at start of pattern.
1066 ph10 223
1067 ph10 227
1068 ph10 189 Version 7.2 19-Jun-07
1069     ---------------------
1070    
1071 ph10 157 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1072     which is apparently normally available under Windows.
1073 ph10 159
1074     2. Re-jig the pcregrep tests with different newline settings in an attempt
1075     to make them independent of the local environment's newline setting.
1076 ph10 157
1077 ph10 160 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1078 ph10 165
1079 ph10 161 4. Some of the "internals" tests were previously cut out when the link size
1080     was not 2, because the output contained actual offsets. The recent new
1081 ph10 165 "Z" feature of pcretest means that these can be cut out, making the tests
1082     usable with all link sizes.
1083    
1084 ph10 164 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1085     stack recursion. This gives a massive performance boost under BSD, but just
1086 ph10 165 a small improvement under Linux. However, it saves one field in the frame
1087 ph10 164 in all cases.
1088 ph10 172
1089 ph10 166 6. Added more features from the forthcoming Perl 5.10:
1090 ph10 172
1091 ph10 166 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1092     recursion call. It refers to the nth most recently opened parentheses.
1093 ph10 172
1094 ph10 166 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1095 ph10 172 to be opened parentheses.
1096    
1097     (c) Conditions that refer to capturing parentheses can be specified
1098 ph10 167 relatively, for example, (?(-2)... or (?(+3)...
1099 ph10 172
1100 ph10 168 (d) \K resets the start of the current match so that everything before
1101 ph10 172 is not part of it.
1102    
1103 ph10 171 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1104 ph10 172
1105 ph10 171 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1106 ph10 172 reference syntax.
1107 ph10 182
1108 ph10 175 (g) (?| introduces a group in which the numbering of parentheses in each
1109 ph10 182 alternative starts with the same number.
1110 ph10 172
1111 ph10 182 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1112    
1113 ph10 172 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1114     PCRE_INFO_JCHANGED.
1115    
1116     8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1117     terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1118 ph10 170 for detecting groups that can match an empty string.
1119 ph10 159
1120 ph10 172 9. A pattern with a very large number of alternatives (more than several
1121     hundred) was running out of internal workspace during the pre-compile
1122     phase, where pcre_compile() figures out how much memory will be needed. A
1123     bit of new cunning has reduced the workspace needed for groups with
1124     alternatives. The 1000-alternative test pattern now uses 12 bytes of
1125     workspace instead of running out of the 4096 that are available.
1126 ph10 182
1127 ph10 176 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1128 ph10 172
1129 ph10 179 11. Applied patch from Google to remove an optimization that didn't quite work.
1130     The report of the bug said:
1131 ph10 182
1132 ph10 179 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1133     pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1134 ph10 182 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1135 ph10 185
1136 ph10 184 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1137 ph10 185 it matched the wrong number of bytes.
1138 ph10 172
1139 ph10 179
1140 ph10 155 Version 7.1 24-Apr-07
1141 ph10 98 ---------------------
1142    
1143 ph10 111 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1144 ph10 122 that is more "standard", making use of automake and other Autotools. There
1145 ph10 99 is some re-arrangement of the files and adjustment of comments consequent
1146     on this.
1147 ph10 111
1148     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1149     for recursive directory scanning broke on some systems because the files
1150     are not scanned in any specific order and on different systems the order
1151     was different. A call to "sort" has been inserted into RunGrepTest for the
1152     approprate test as a short-term fix. In the longer term there may be an
1153 ph10 100 alternative.
1154 ph10 111
1155 ph10 100 3. I had an email from Eric Raymond about problems translating some of PCRE's
1156 ph10 111 man pages to HTML (despite the fact that I distribute HTML pages, some
1157     people do their own conversions for various reasons). The problems
1158     concerned the use of low-level troff macros .br and .in. I have therefore
1159     removed all such uses from the man pages (some were redundant, some could
1160 ph10 113 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1161     HTML has been updated to handle .nf/.fi and to complain if it encounters
1162     .br or .in.
1163 ph10 111
1164 ph10 100 4. Updated comments in configure.ac that get placed in config.h.in and also
1165 ph10 123 arranged for config.h to be included in the distribution, with the name
1166 ph10 111 config.h.generic, for the benefit of those who have to compile without
1167     Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1168    
1169     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1170     Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1171 ph10 127 makevp.bat for the current PCRE, using the additional files
1172 ph10 135 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1173 ph10 111
1174     6. A Windows user reported a minor discrepancy with test 2, which turned out
1175     to be caused by a trailing space on an input line that had got lost in his
1176 ph10 102 copy. The trailing space was an accident, so I've just removed it.
1177 ph10 111
1178 ph10 104 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1179 ph10 111 that is needed.
1180    
1181 ph10 105 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1182     as "const" (a) because they are and (b) because it helps the PHP
1183     maintainers who have recently made a script to detect big data structures
1184 ph10 111 in the php code that should be moved to the .rodata section. I remembered
1185     to update Builducptable as well, so it won't revert if ucptable.h is ever
1186 ph10 105 re-created.
1187 ph10 111
1188     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1189 ph10 107 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1190 ph10 111 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1191     support is not required. This saves 1.5-2K of code, which is important in
1192 ph10 107 some applications.
1193 ph10 111
1194 ph10 110 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1195 ph10 111 so as not to refer to the tables, even though these functions will never be
1196     called when UTF-8 support is disabled. Otherwise there are problems with a
1197     shared library.
1198    
1199 ph10 118 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1200    
1201     (a) It was defining its arguments as char * instead of void *.
1202    
1203 ph10 123 (b) It was assuming that all moves were upwards in memory; this was true
1204     a long time ago when I wrote it, but is no longer the case.
1205    
1206 ph10 118 The emulated memove() is provided for those environments that have neither
1207 ph10 123 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1208 ph10 118 is clearly not the case, as these two bugs were recently reported.
1209 ph10 123
1210 ph10 111 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1211 ph10 123 and Detrail to create the HTML documentation, the .txt form of the man
1212     pages, and it removes trailing spaces from listed files. It also creates
1213     pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1214     case, it wraps all the #defines with #ifndefs. This script should be run
1215 ph10 111 before "make dist".
1216 ph10 123
1217 ph10 115 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1218     with Unicode property support.
1219 ph10 123
1220     (a) For a maximizing quantifier, if the two different cases of the
1221     character were of different lengths in their UTF-8 codings (there are
1222     some cases like this - I found 11), and the matching function had to
1223 ph10 115 back up over a mixture of the two cases, it incorrectly assumed they
1224     were both the same length.
1225 ph10 123
1226     (b) When PCRE was configured to use the heap rather than the stack for
1227     recursion during matching, it was not correctly preserving the data for
1228     the other case of a UTF-8 character when checking ahead for a match
1229     while processing a minimizing repeat. If the check also involved
1230 ph10 115 matching a wide character, but failed, corruption could cause an
1231     erroneous result when trying to check for a repeat of the original
1232     character.
1233 ph10 123
1234 ph10 116 13. Some tidying changes to the testing mechanism:
1235 ph10 98
1236 ph10 116 (a) The RunTest script now detects the internal link size and whether there
1237     is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1238 ph10 123 values substituted by "configure". (The RunGrepTest script already did
1239     this for UTF-8.) The configure.ac script no longer substitutes the
1240     relevant variables.
1241    
1242 ph10 116 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1243     with length and offset values. This means that the output is different
1244     for different internal link sizes. Test 2 is skipped for link sizes
1245     other than 2 because of this, bypassing the problem. Unfortunately,
1246     there was also a test in test 3 (the locale tests) that used /B and
1247 ph10 123 failed for link sizes other than 2. Rather than cut the whole test out,
1248     I have added a new /Z option to pcretest that replaces the length and
1249     offset values with spaces. This is now used to make test 3 independent
1250 ph10 122 of link size. (Test 2 will be tidied up later.)
1251 ph10 123
1252     14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1253 ph10 122 segmentation fault instead of returning the appropriate error message.
1254 ph10 134
1255 ph10 131 15. In multiline mode when the newline sequence was set to "any", the pattern
1256 ph10 134 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1257     This doesn't seem right; it now treats the CRLF combination as the line
1258     ending, and so does not match in that case. It's only a pattern such as ^$
1259     that would hit this one: something like ^ABC$ would have failed after \r
1260     and then tried again after \r\n.
1261    
1262 ph10 131 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1263 ph10 134 in an attempt to make files that differ only in their line terminators
1264     compare equal. This works on Linux.
1265 ph10 142
1266 ph10 141 17. Under certain error circumstances pcregrep might try to free random memory
1267     as it exited. This is now fixed, thanks to valgrind.
1268 ph10 142
1269 ph10 141 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1270 ph10 142 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1271     was because its rules for how to advance for /g after matching an empty
1272 ph10 143 string at the end of a line did not allow for this case. They now check for
1273     it specially.
1274 ph10 150
1275     20. pcretest is supposed to handle patterns and data of any length, by
1276     extending its buffers when necessary. It was getting this wrong when the
1277 ph10 147 buffer for a data line had to be extended.
1278 ph10 150
1279 ph10 149 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1280 ph10 150 CRLF as a newline sequence.
1281 ph10 152
1282 ph10 151 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1283 ph10 152 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1284     I have nevertheless tidied it up.
1285 ph10 111
1286 ph10 152 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1287 ph10 123
1288 ph10 153 24. Added a man page for pcre-config.
1289 ph10 152
1290 ph10 153
1291 nigel 93 Version 7.0 19-Dec-06
1292     ---------------------
1293    
1294     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1295     moving to gcc 4.1.1.
1296    
1297     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1298     sys/time.h, which is documented as needed for this function. It doesn't
1299     seem to matter on Linux, but it showed up on some releases of OS X.
1300    
1301     3. It seems that there are systems where bytes whose values are greater than
1302     127 match isprint() in the "C" locale. The "C" locale should be the
1303     default when a C program starts up. In most systems, only ASCII printing
1304     characters match isprint(). This difference caused the output from pcretest
1305     to vary, making some of the tests fail. I have changed pcretest so that:
1306    
1307     (a) When it is outputting text in the compiled version of a pattern, bytes
1308     other than 32-126 are always shown as hex escapes.
1309    
1310     (b) When it is outputting text that is a matched part of a subject string,
1311     it does the same, unless a different locale has been set for the match
1312     (using the /L modifier). In this case, it uses isprint() to decide.
1313    
1314     4. Fixed a major bug that caused incorrect computation of the amount of memory
1315     required for a compiled pattern when options that changed within the
1316     pattern affected the logic of the preliminary scan that determines the
1317     length. The relevant options are -x, and -i in UTF-8 mode. The result was
1318     that the computed length was too small. The symptoms of this bug were
1319     either the PCRE error "internal error: code overflow" from pcre_compile(),
1320     or a glibc crash with a message such as "pcretest: free(): invalid next
1321     size (fast)". Examples of patterns that provoked this bug (shown in
1322     pcretest format) are:
1323    
1324     /(?-x: )/x
1325     /(?x)(?-x: \s*#\s*)/
1326     /((?i)[\x{c0}])/8
1327     /(?i:[\x{c0}])/8
1328    
1329     HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1330     is now done differently.
1331    
1332     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1333     wrapper classes; (b) implement a new function in the C++ scanner that is
1334     more efficient than the old way of doing things because it avoids levels of
1335     recursion in the regex matching; (c) add a paragraph to the documentation
1336     for the FullMatch() function.
1337    
1338     6. The escape sequence \n was being treated as whatever was defined as
1339     "newline". Not only was this contrary to the documentation, which states
1340     that \n is character 10 (hex 0A), but it also went horribly wrong when
1341     "newline" was defined as CRLF. This has been fixed.
1342    
1343     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1344     was being set to -1 for the "end of line" case (supposedly a value that no
1345     character can have). Though this value is never used (the check for end of
1346     line is "zero bytes in current character"), it caused compiler complaints.
1347     I've changed it to 0xffffffff.
1348    
1349     8. In pcre_version.c, the version string was being built by a sequence of
1350     C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1351     string (as it is for production releases) called a macro with an empty
1352     argument. The C standard says the result of this is undefined. The gcc
1353     compiler treats it as an empty string (which was what was wanted) but it is
1354     reported that Visual C gives an error. The source has been hacked around to
1355     avoid this problem.
1356    
1357     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1358     builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1359     instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1360     of them did).
1361    
1362     10. Originally, pcretest opened its input and output without "b"; then I was
1363     told that "b" was needed in some environments, so it was added for release
1364     5.0 to both the input and output. (It makes no difference on Unix-like
1365     systems.) Later I was told that it is wrong for the input on Windows. I've
1366     now abstracted the modes into two macros, to make it easier to fiddle with
1367     them, and removed "b" from the input mode under Windows.
1368    
1369     11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1370    
1371     12. Added -help and --help to pcretest as an official way of being reminded
1372     of the options.
1373    
1374     13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1375     and pcrecpp.cc because they annoy compilers at high warning levels.
1376    
1377     14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1378    
1379     15. Fixed an occurrence of == in configure.ac that should have been = (shell
1380     scripts are not C programs :-) and which was not noticed because it works
1381     on Linux.
1382    
1383     16. pcretest is supposed to handle any length of pattern and data line (as one
1384     line or as a continued sequence of lines) by extending its input buffer if
1385     necessary. This feature was broken for very long pattern lines, leading to
1386     a string of junk being passed to pcre_compile() if the pattern was longer
1387     than about 50K.
1388    
1389     17. I have done a major re-factoring of the way pcre_compile() computes the
1390     amount of memory needed for a compiled pattern. Previously, there was code
1391     that made a preliminary scan of the pattern in order to do this. That was
1392     OK when PCRE was new, but as the facilities have expanded, it has become
1393     harder and harder to keep it in step with the real compile phase, and there
1394     have been a number of bugs (see for example, 4 above). I have now found a
1395     cunning way of running the real compile function in a "fake" mode that
1396     enables it to compute how much memory it would need, while actually only
1397     ever using a few hundred bytes of working memory and without too many
1398     tests of the mode. This should make future maintenance and development
1399     easier. A side effect of this work is that the limit of 200 on the nesting
1400     depth of parentheses has been removed (though this was never a serious
1401     limitation, I suspect). However, there is a downside: pcre_compile() now
1402     runs more slowly than before (30% or more, depending on the pattern). I
1403     hope this isn't a big issue. There is no effect on runtime performance.
1404    
1405     18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1406     newline (only possible for the last line of a file) and it was a
1407     pattern that set a locale (followed by /Lsomething), pcretest crashed.
1408    
1409     19. Added additional timing features to pcretest. (1) The -tm option now times
1410     matching only, not compiling. (2) Both -t and -tm can be followed, as a
1411     separate command line item, by a number that specifies the number of
1412     repeats to use when timing. The default is 50000; this gives better
1413     precision, but takes uncomfortably long for very large patterns.
1414    
1415     20. Extended pcre_study() to be more clever in cases where a branch of a
1416     subpattern has no definite first character. For example, (a*|b*)[cd] would
1417     previously give no result from pcre_study(). Now it recognizes that the
1418     first character must be a, b, c, or d.
1419    
1420     21. There was an incorrect error "recursive call could loop indefinitely" if
1421     a subpattern (or the entire pattern) that was being tested for matching an
1422     empty string contained only one non-empty item after a nested subpattern.
1423     For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1424     incorrectly, because the \d was being skipped in the check.
1425    
1426     22. The pcretest program now has a new pattern option /B and a command line
1427     option -b, which is equivalent to adding /B to every pattern. This causes
1428     it to show the compiled bytecode, without the additional information that
1429     -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1430     is the same as /B/I).
1431    
1432     23. A new optimization is now able automatically to treat some sequences such
1433     as a*b as a*+b. More specifically, if something simple (such as a character
1434     or a simple class like \d) has an unlimited quantifier, and is followed by
1435     something that cannot possibly match the quantified thing, the quantifier
1436     is automatically "possessified".
1437    
1438     24. A recursive reference to a subpattern whose number was greater than 39
1439     went wrong under certain circumstances in UTF-8 mode. This bug could also
1440     have affected the operation of pcre_study().
1441    
1442     25. Realized that a little bit of performance could be had by replacing
1443     (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1444    
1445     26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1446    
1447     27. Possessive quantifiers such as a++ were previously implemented by turning
1448     them into atomic groups such as ($>a+). Now they have their own opcodes,
1449     which improves performance. This includes the automatically created ones
1450     from 23 above.
1451    
1452     28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1453     lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1454     the first matched character to be a colon. This applied both to named and
1455     numbered groups.
1456    
1457     29. The ucpinternal.h header file was missing its idempotency #ifdef.
1458    
1459     30. I was sent a "project" file called libpcre.a.dev which I understand makes
1460     building PCRE on Windows easier, so I have included it in the distribution.
1461    
1462     31. There is now a check in pcretest against a ridiculously large number being
1463     returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1464     loop, the loop is abandoned.
1465    
1466     32. Forward references to subpatterns in conditions such as (?(2)...) where
1467     subpattern 2 is defined later cause pcre_compile() to search forwards in
1468     the pattern for the relevant set of parentheses. This search went wrong
1469     when there were unescaped parentheses in a character class, parentheses
1470     escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1471    
1472     33. "Subroutine" calls and backreferences were previously restricted to
1473     referencing subpatterns earlier in the regex. This restriction has now
1474     been removed.
1475    
1476     34. Added a number of extra features that are going to be in Perl 5.10. On the
1477     whole, these are just syntactic alternatives for features that PCRE had
1478     previously implemented using the Python syntax or my own invention. The
1479     other formats are all retained for compatibility.
1480    
1481     (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1482     as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1483     also .NET compatible.
1484    
1485     (b) A recursion or subroutine call to a named group can now be defined as
1486     (?&name) as well as (?P>name).
1487    
1488     (c) A backreference to a named group can now be defined as \k<name> or
1489     \k'name' as well as (?P=name). The new forms, as well as being in Perl
1490     5.10, are also .NET compatible.
1491    
1492     (d) A conditional reference to a named group can now use the syntax
1493     (?(<name>) or (?('name') as well as (?(name).
1494    
1495     (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1496     groups (named and numbered) that are never evaluated inline, but can be
1497     called as "subroutines" from elsewhere. In effect, the DEFINE condition
1498     is always false. There may be only one alternative in such a group.
1499    
1500     (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1501     as the simple (?(R). The condition is true only if the most recent
1502     recursion is that of the given number or name. It does not search out
1503     through the entire recursion stack.
1504    
1505     (g) The escape \gN or \g{N} has been added, where N is a positive or
1506     negative number, specifying an absolute or relative reference.
1507    
1508     35. Tidied to get rid of some further signed/unsigned compiler warnings and
1509     some "unreachable code" warnings.
1510    
1511     36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1512     things, this adds five new scripts.
1513    
1514     37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1515     There were also incompatibilities regarding the handling of \Q..\E inside
1516     character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1517     hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1518    
1519     38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1520     matches an empty string, and forcibly breaks the loop. There were bugs in
1521     this code in non-simple cases. For a pattern such as ^(a()*)* matched
1522     against aaaa the result was just "a" rather than "aaaa", for example. Two
1523     separate and independent bugs (that affected different cases) have been
1524     fixed.
1525    
1526     39. Refactored the code to abolish the use of different opcodes for small
1527     capturing bracket numbers. This is a tidy that I avoided doing when I
1528     removed the limit on the number of capturing brackets for 3.5 back in 2001.
1529     The new approach is not only tidier, it makes it possible to reduce the
1530     memory needed to fix the previous bug (38).
1531    
1532     40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1533     sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1534     processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1535     mode.
1536    
1537     41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1538     report.
1539    
1540     42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1541     copy construction and assignment in the C++ wrapper.
1542    
1543     43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1544     couple of bugs that could have given wrong results in the "--newline=crlf"
1545     case.
1546    
1547     44. Added a number of casts and did some reorganization of signed/unsigned int
1548     variables following suggestions from Dair Grant. Also renamed the variable
1549     "this" as "item" because it is a C++ keyword.
1550    
1551     45. Arranged for dftables to add
1552    
1553     #include "pcre_internal.h"
1554    
1555     to pcre_chartables.c because without it, gcc 4.x may remove the array
1556     definition from the final binary if PCRE is built into a static library and
1557     dead code stripping is activated.
1558    
1559     46. For an unanchored pattern, if a match attempt fails at the start of a
1560     newline sequence, and the newline setting is CRLF or ANY, and the next two
1561     characters are CRLF, advance by two characters instead of one.
1562    
1563    
1564 nigel 91 Version 6.7 04-Jul-06
1565     ---------------------
1566    
1567     1. In order to handle tests when input lines are enormously long, pcretest has
1568     been re-factored so that it automatically extends its buffers when
1569     necessary. The code is crude, but this _is_ just a test program. The
1570     default size has been increased from 32K to 50K.
1571    
1572     2. The code in pcre_study() was using the value of the re argument before
1573     testing it for NULL. (Of course, in any sensible call of the function, it
1574     won't be NULL.)
1575    
1576     3. The memmove() emulation function in pcre_internal.h, which is used on
1577     systems that lack both memmove() and bcopy() - that is, hardly ever -
1578     was missing a "static" storage class specifier.
1579    
1580     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1581     containing an extended class (one that cannot be represented by a bitmap
1582     because it contains high-valued characters or Unicode property items, e.g.
1583     [\pZ]). Almost always one would set UTF-8 mode when processing such a
1584     pattern, but PCRE should not loop if you do not (it no longer does).
1585     [Detail: two cases were found: (a) a repeated subpattern containing an
1586     extended class; (b) a recursive reference to a subpattern that followed a
1587     previous extended class. It wasn't skipping over the extended class
1588     correctly when UTF-8 mode was not set.]
1589    
1590     5. A negated single-character class was not being recognized as fixed-length
1591     in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1592     compile error "lookbehind assertion is not fixed length".
1593    
1594     6. The RunPerlTest auxiliary script was showing an unexpected difference
1595     between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1596     write a Perl script that can interpret lines of an input file either as
1597     byte characters or as UTF-8, which is what "perltest" was being required to
1598     do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1599     can't do is switch easily at run time between having the "use utf8;" pragma
1600     or not. In the end, I fudged it by using the RunPerlTest script to insert
1601     "use utf8;" explicitly for the UTF-8 tests.
1602    
1603     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1604     the end of the subject string, contrary to the documentation and to what
1605     Perl does. This was true of both matching functions. Now it matches only at
1606     the start of the subject and immediately after *internal* newlines.
1607    
1608     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1609     a pointer to an int instead of a pointer to an unsigned long int. This
1610     caused problems on 64-bit systems.
1611    
1612     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1613     instance of the 'standard' template library not being so standard".
1614    
1615     10. There was no check on the number of named subpatterns nor the maximum
1616     length of a subpattern name. The product of these values is used to compute
1617     the size of the memory block for a compiled pattern. By supplying a very
1618     long subpattern name and a large number of named subpatterns, the size
1619     computation could be caused to overflow. This is now prevented by limiting
1620     the length of names to 32 characters, and the number of named subpatterns
1621     to 10,000.
1622    
1623     11. Subpatterns that are repeated with specific counts have to be replicated in
1624     the compiled pattern. The size of memory for this was computed from the
1625     length of the subpattern and the repeat count. The latter is limited to
1626     65535, but there was no limit on the former, meaning that integer overflow
1627     could in principle occur. The compiled length of a repeated subpattern is
1628     now limited to 30,000 bytes in order to prevent this.
1629    
1630     12. Added the optional facility to have named substrings with the same name.
1631    
1632     13. Added the ability to use a named substring as a condition, using the
1633     Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1634     are numbers (not recommended). Forward references are permitted.
1635    
1636     14. Added forward references in named backreferences (if you see what I mean).
1637    
1638     15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1639     pattern could run off the end of the subject. For example, the pattern
1640     "(?s)(.{1,5})"8 did this with the subject "ab".
1641    
1642     16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1643     PCRE_CASELESS was set when matching characters that were quantified with ?
1644     or *.
1645    
1646     17. A character class other than a single negated character that had a minimum
1647     but no maximum quantifier - for example [ab]{6,} - was not handled
1648     correctly by pce_dfa_exec(). It would match only one character.
1649    
1650     18. A valid (though odd) pattern that looked like a POSIX character
1651     class but used an invalid character after [ (for example [[,abc,]]) caused
1652     pcre_compile() to give the error "Failed: internal error: code overflow" or
1653     in some cases to crash with a glibc free() error. This could even happen if
1654     the pattern terminated after [[ but there just happened to be a sequence of
1655     letters, a binary zero, and a closing ] in the memory that followed.
1656    
1657     19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1658     over the years. Originally (before any Unicode support), just the bottom 8
1659     bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1660     output from "man perlunicode" includes this:
1661    
1662     The regular expression compiler produces polymorphic opcodes. That
1663     is, the pattern adapts to the data and automatically switches to
1664     the Unicode character scheme when presented with Unicode data--or
1665     instead uses a traditional byte scheme when presented with byte
1666     data.
1667    
1668     Sadly, a wide octal escape does not cause a switch, and in a string with
1669     no other multibyte characters, these octal escapes are treated as before.
1670     Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1671     /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1672     Unicode string.
1673    
1674     I have not perpetrated such confusion in PCRE. Up till now, it took just
1675     the bottom 8 bits, as in old Perl. I have now made octal escapes with
1676     values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1677     translate to the appropriate multibyte character.
1678    
1679     29. Applied some refactoring to reduce the number of warnings from Microsoft
1680     and Borland compilers. This has included removing the fudge introduced
1681     seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1682     a warning about an unused variable.
1683    
1684     21. PCRE has not included VT (character 0x0b) in the set of whitespace
1685     characters since release 4.0, because Perl (from release 5.004) does not.
1686     [Or at least, is documented not to: some releases seem to be in conflict
1687     with the documentation.] However, when a pattern was studied with
1688     pcre_study() and all its branches started with \s, PCRE still included VT
1689     as a possible starting character. Of course, this did no harm; it just
1690     caused an unnecessary match attempt.
1691    
1692     22. Removed a now-redundant internal flag bit that recorded the fact that case
1693     dependency changed within the pattern. This was once needed for "required
1694     byte" processing, but is no longer used. This recovers a now-scarce options
1695     bit. Also moved the least significant internal flag bit to the most-
1696     significant bit of the word, which was not previously used (hangover from
1697     the days when it was an int rather than a uint) to free up another bit for
1698     the future.
1699    
1700     23. Added support for CRLF line endings as well as CR and LF. As well as the
1701     default being selectable at build time, it can now be changed at runtime
1702     via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1703     specify that it is scanning data with non-default line endings.
1704    
1705     24. Changed the definition of CXXLINK to make it agree with the definition of
1706     LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1707    
1708     25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1709     recursions. This makes a big different to stack usage for some patterns.
1710    
1711     26. If a subpattern containing a named recursion or subroutine reference such
1712     as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1713     the space required for the compiled pattern went wrong and gave too small a
1714     value. Depending on the environment, this could lead to "Failed: internal
1715     error: code overflow at offset 49" or "glibc detected double free or
1716     corruption" errors.
1717    
1718     27. Applied patches from Google (a) to support the new newline modes and (b) to
1719     advance over multibyte UTF-8 characters in GlobalReplace.
1720    
1721     28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1722     difference for some implementation of PCRE in some Windows version.
1723    
1724     29. Added some extra testing facilities to pcretest:
1725    
1726     \q<number> in a data line sets the "match limit" value
1727     \Q<number> in a data line sets the "match recursion limt" value
1728     -S <number> sets the stack size, where <number> is in megabytes
1729    
1730     The -S option isn't available for Windows.
1731    
1732    
1733 nigel 89 Version 6.6 06-Feb-06
1734     ---------------------
1735    
1736     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1737     in pcreposix.h. I have copied the definition from pcre.h.
1738    
1739     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1740     because pcre.h is no longer a built file.
1741    
1742     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1743     not normally included in the compiled code.
1744    
1745    
1746 nigel 87 Version 6.5 01-Feb-06
1747     ---------------------
1748    
1749     1. When using the partial match feature with pcre_dfa_exec(), it was not
1750     anchoring the second and subsequent partial matches at the new starting
1751     point. This could lead to incorrect results. For example, with the pattern
1752     /1234/, partially matching against "123" and then "a4" gave a match.
1753    
1754     2. Changes to pcregrep:
1755    
1756     (a) All non-match returns from pcre_exec() were being treated as failures
1757     to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1758     error message is output. Some extra information is given for the
1759     PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
1760     probably the only errors that are likely to be caused by users (by
1761     specifying a regex that has nested indefinite repeats, for instance).
1762     If there are more than 20 of these errors, pcregrep is abandoned.
1763    
1764     (b) A binary zero was treated as data while matching, but terminated the
1765     output line if it was written out. This has been fixed: binary zeroes
1766     are now no different to any other data bytes.
1767    
1768     (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1769     used to set a locale for matching. The --locale=xxxx long option has
1770     been added (no short equivalent) to specify a locale explicitly on the
1771     pcregrep command, overriding the environment variables.
1772    
1773     (d) When -B was used with -n, some line numbers in the output were one less
1774     than they should have been.
1775    
1776     (e) Added the -o (--only-matching) option.
1777    
1778     (f) If -A or -C was used with -c (count only), some lines of context were
1779     accidentally printed for the final match.
1780    
1781     (g) Added the -H (--with-filename) option.
1782    
1783     (h) The combination of options -rh failed to suppress file names for files
1784     that were found from directory arguments.
1785    
1786     (i) Added the -D (--devices) and -d (--directories) options.
1787    
1788     (j) Added the -F (--fixed-strings) option.
1789    
1790     (k) Allow "-" to be used as a file name for -f as well as for a data file.
1791    
1792     (l) Added the --colo(u)r option.
1793    
1794     (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1795     is not present by default.
1796    
1797     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1798     items such as (?R) or (?1), when the recursion could match a number of
1799     alternatives. If it matched one of the alternatives, but subsequently,
1800     outside the recursion, there was a failure, the code tried to back up into
1801     the recursion. However, because of the way PCRE is implemented, this is not
1802     possible, and the result was an incorrect result from the match.
1803    
1804     In order to prevent this happening, the specification of recursion has
1805     been changed so that all such subpatterns are automatically treated as
1806     atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1807    
1808     4. I had overlooked the fact that, in some locales, there are characters for
1809     which isalpha() is true but neither isupper() nor islower() are true. In
1810     the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1811     and ordfeminine) are like this. This affected the treatment of \w and \W
1812     when they appeared in character classes, but not when they appeared outside
1813     a character class. The bit map for "word" characters is now created
1814     separately from the results of isalnum() instead of just taking it from the
1815     upper, lower, and digit maps. (Plus the underscore character, of course.)
1816    
1817     5. The above bug also affected the handling of POSIX character classes such as
1818     [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1819     permanent tables. Instead, the bit maps for such a class were previously
1820     created as the appropriate unions of the upper, lower, and digit bitmaps.
1821     Now they are created by subtraction from the [[:word:]] class, which has
1822     its own bitmap.
1823    
1824     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1825     It is created by subtracting the vertical space characters (\x09, \x0a,
1826     \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1827     subtraction was done in the overall bitmap for a character class, meaning
1828     that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1829     be recognized. This bug has been fixed.
1830    
1831     7. Patches from the folks at Google:
1832    
1833     (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1834     real life, but is still worth protecting against".
1835    
1836     (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1837     regular expressions".
1838    
1839     (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1840     have it.
1841    
1842     (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1843     "configure" and the latter not, in order to fix a problem somebody had
1844     with compiling the Arg class on HP-UX.
1845    
1846     (e) Improve the error-handling of the C++ wrapper a little bit.
1847    
1848     (f) New tests for checking recursion limiting.
1849    
1850     8. The pcre_memmove() function, which is used only if the environment does not
1851     have a standard memmove() function (and is therefore rarely compiled),
1852     contained two bugs: (a) use of int instead of size_t, and (b) it was not
1853     returning a result (though PCRE never actually uses the result).
1854    
1855     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1856     large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1857     returned instead of calling malloc() with an overflowing number that would
1858     most likely cause subsequent chaos.
1859    
1860     10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
1861    
1862     11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
1863     with this option is matched, the nmatch and pmatch options of regexec() are
1864     ignored.
1865    
1866     12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
1867     provided in case anyone wants to the the POSIX interface with UTF-8
1868     strings.
1869    
1870     13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
1871     C++ linking (needed for some HP-UX environments).
1872    
1873     14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
1874     (unused parameter) and in the pcre_printint() function (omitted "default"
1875     switch label when the default is to do nothing).
1876    
1877     15. Added some code to make it possible, when PCRE is compiled as a C++
1878     library, to replace subject pointers for pcre_exec() with a smart pointer
1879     class, thus making it possible to process discontinuous strings.
1880    
1881     16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
1882     much the same function. They were added by different people who were trying
1883     to make PCRE easy to compile on non-Unix systems. It has been suggested
1884     that PCRE_EXPORT be abolished now that there is more automatic apparatus
1885     for compiling on Windows systems. I have therefore replaced it with
1886     PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
1887     defaults to "extern" for C or "extern C" for C++, which works fine on
1888     Unix-like systems. It is now possible to override the value of PCRE_DATA_
1889     SCOPE with something explicit in config.h. In addition:
1890    
1891     (a) pcreposix.h still had just "extern" instead of either of these macros;
1892     I have replaced it with PCRE_DATA_SCOPE.
1893    
1894     (b) Functions such as _pcre_xclass(), which are internal to the library,
1895     but external in the C sense, all had PCRE_EXPORT in their definitions.
1896     This is apparently wrong for the Windows case, so I have removed it.
1897     (It makes no difference on Unix-like systems.)
1898    
1899     17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
1900     of recursive calls to match(). This is different to MATCH_LIMIT because
1901     that limits the total number of calls to match(), not all of which increase
1902     the depth of recursion. Limiting the recursion depth limits the amount of
1903     stack (or heap if NO_RECURSE is set) that is used. The default can be set
1904     when PCRE is compiled, and changed at run time. A patch from Google adds
1905     this functionality to the C++ interface.
1906    
1907     18. Changes to the handling of Unicode character properties:
1908    
1909     (a) Updated the table to Unicode 4.1.0.
1910    
1911     (b) Recognize characters that are not in the table as "Cn" (undefined).
1912    
1913     (c) I revised the way the table is implemented to a much improved format
1914     which includes recognition of ranges. It now supports the ranges that
1915     are defined in UnicodeData.txt, and it also amalgamates other
1916     characters into ranges. This has reduced the number of entries in the
1917     table from around 16,000 to around 3,000, thus reducing its size
1918     considerably. I realized I did not need to use a tree structure after
1919     all - a binary chop search is just as efficient. Having reduced the
1920     number of entries, I extended their size from 6 bytes to 8 bytes to
1921     allow for more data.
1922    
1923     (d) Added support for Unicode script names via properties such as \p{Han}.
1924    
1925     19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
1926     matching that character.
1927    
1928     20. When matching a repeated Unicode property with a minimum greater than zero,
1929     (for example \pL{2,}), PCRE could look past the end of the subject if it
1930     reached it while seeking the minimum number of characters. This could
1931     happen only if some of the characters were more than one byte long, because
1932     there is a check for at least the minimum number of bytes.
1933    
1934     21. Refactored the implementation of \p and \P so as to be more general, to
1935     allow for more different types of property in future. This has changed the
1936     compiled form incompatibly. Anybody with saved compiled patterns that use
1937     \p or \P will have to recompile them.
1938    
1939     22. Added "Any" and "L&" to the supported property types.
1940    
1941     23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
1942     but give a compile time error if the value is greater than 0xff.
1943    
1944     24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
1945     accidentally not being installed or uninstalled.
1946    
1947     25. The pcre.h file was built from pcre.h.in, but the only changes that were
1948     made were to insert the current release number. This seemed silly, because
1949     it made things harder for people building PCRE on systems that don't run
1950     "configure". I have turned pcre.h into a distributed file, no longer built
1951     by "configure", with the version identification directly included. There is
1952     no longer a pcre.h.in file.
1953    
1954     However, this change necessitated a change to the pcre-config script as
1955     well. It is built from pcre-config.in, and one of the substitutions was the
1956     release number. I have updated configure.ac so that ./configure now finds
1957     the release number by grepping pcre.h.
1958    
1959     26. Added the ability to run the tests under valgrind.
1960    
1961    
1962 nigel 85 Version 6.4 05-Sep-05
1963     ---------------------
1964    
1965     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
1966     "--" to be printed when multiple files were scanned, even when none of the
1967     -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
1968     consider it to be a bug, and have restored the previous behaviour.
1969    
1970     2. A couple of code tidies to get rid of compiler warnings.
1971    
1972     3. The pcretest program used to cheat by referring to symbols in the library
1973     whose names begin with _pcre_. These are internal symbols that are not
1974     really supposed to be visible externally, and in some environments it is
1975     possible to suppress them. The cheating is now confined to including
1976     certain files from the library's source, which is a bit cleaner.
1977    
1978     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
1979     file's purpose clearer.
1980    
1981     5. Reorganized pcre_ucp_findchar().
1982    
1983    
1984 nigel 83 Version 6.3 15-Aug-05
1985     ---------------------
1986    
1987     1. The file libpcre.pc.in did not have general read permission in the tarball.
1988    
1989     2. There were some problems when building without C++ support:
1990    
1991     (a) If C++ support was not built, "make install" and "make test" still
1992     tried to test it.
1993    
1994     (b) There were problems when the value of CXX was explicitly set. Some
1995     changes have been made to try to fix these, and ...
1996    
1997     (c) --disable-cpp can now be used to explicitly disable C++ support.
1998    
1999     (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2000     backslash in a target when C++ was disabled. This confuses some
2001     versions of "make", apparently. Using an intermediate variable solves
2002     this. (Same for CPP_LOBJ.)
2003    
2004     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2005     (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2006     necessary on certain architectures.
2007    
2008     4. Added a setting of -export-symbols-regex to the link command to remove
2009     those symbols that are exported in the C sense, but actually are local
2010     within the library, and not documented. Their names all begin with
2011     "_pcre_". This is not a perfect job, because (a) we have to except some
2012     symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2013     available (and never for static libraries). I have made a note to try to
2014     find a way round (a) in the future.
2015    
2016    
2017 nigel 81 Version 6.2 01-Aug-05
2018     ---------------------
2019    
2020     1. There was no test for integer overflow of quantifier values. A construction
2021     such as {1111111111111111} would give undefined results. What is worse, if
2022     a minimum quantifier for a parenthesized subpattern overflowed and became
2023     negative, the calculation of the memory size went wrong. This could have
2024     led to memory overwriting.
2025    
2026     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2027    
2028     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2029     operating environments where this matters.
2030    
2031     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2032     PCRE options from within the C++ wrapper.
2033    
2034     5. Named capturing subpatterns were not being correctly counted when a pattern
2035     was compiled. This caused two problems: (a) If there were more than 100
2036     such subpatterns, the calculation of the memory needed for the whole
2037     compiled pattern went wrong, leading to an overflow error. (b) Numerical
2038     back references of the form \12, where the number was greater than 9, were
2039     not recognized as back references, even though there were sufficient
2040     previous subpatterns.
2041    
2042     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2043     versions of gcc, e.g. 2.95.4.
2044    
2045    
2046 nigel 79 Version 6.1 21-Jun-05
2047     ---------------------
2048    
2049     1. There was one reference to the variable "posix" in pcretest.c that was not
2050     surrounded by "#if !defined NOPOSIX".
2051    
2052     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2053     the cross-check on the old pcre_info() function, for the benefit of the
2054     cut-down version of PCRE that is currently imported into Exim.
2055    
2056     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2057     allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2058     patterns that start (?i) but I don't think that matters. The use of (?i) is
2059     just an example; this all applies to the other options as well.
2060    
2061     4. Since libtool seems to echo the compile commands it is issuing, the output
2062     from "make" can be reduced a bit by putting "@" in front of each libtool
2063     compile command.
2064    
2065     5. Patch from the folks at Google for configure.in to be a bit more thorough
2066     in checking for a suitable C++ installation before trying to compile the
2067     C++ stuff. This should fix a reported problem when a compiler was present,
2068     but no suitable headers.
2069    
2070     6. The man pages all had just "PCRE" as their title. I have changed them to
2071     be the relevant file name. I have also arranged that these names are
2072     retained in the file doc/pcre.txt, which is a concatenation in text format
2073     of all the man pages except the little individual ones for each function.
2074    
2075     7. The NON-UNIX-USE file had not been updated for the different set of source
2076     files that come with release 6. I also added a few comments about the C++
2077     wrapper.
2078    
2079    
2080 nigel 77 Version 6.0 07-Jun-05
2081     ---------------------
2082    
2083     1. Some minor internal re-organization to help with my DFA experiments.
2084    
2085     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2086     didn't matter for the library itself when fully configured, but did matter
2087     when compiling without UCP support, or within Exim, where the ucp files are
2088     not imported.
2089    
2090     3. Refactoring of the library code to split up the various functions into
2091     different source modules. The addition of the new DFA matching code (see
2092     below) to a single monolithic source would have made it really too
2093     unwieldy, quite apart from causing all the code to be include in a
2094     statically linked application, when only some functions are used. This is
2095     relevant even without the DFA addition now that patterns can be compiled in
2096     one application and matched in another.
2097    
2098     The downside of splitting up is that there have to be some external
2099     functions and data tables that are used internally in different modules of
2100     the library but which are not part of the API. These have all had their
2101     names changed to start with "_pcre_" so that they are unlikely to clash
2102     with other external names.
2103    
2104     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2105     a different (DFA) algorithm. Although it is slower than the original
2106     function, it does have some advantages for certain types of matching
2107     problem.
2108    
2109     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2110     including restarting after a partial match.
2111    
2112     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2113     defined when compiling for Windows was sent to me. I have put it into the
2114     code, though I have no means of testing or verifying it.
2115    
2116     7. Added the pcre_refcount() auxiliary function.
2117    
2118     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2119     match before or at the first newline in the subject string. In pcretest,
2120     the /f option on a pattern can be used to set this.
2121    
2122     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2123     would behave wrongly. This has been present in PCRE since release 4.0.
2124    
2125     10. A number of changes to the pcregrep command:
2126    
2127     (a) Refactored how -x works; insert ^(...)$ instead of setting
2128     PCRE_ANCHORED and checking the length, in preparation for adding
2129     something similar for -w.
2130    
2131     (b) Added the -w (match as a word) option.
2132    
2133     (c) Refactored the way lines are read and buffered so as to have more
2134     than one at a time available.
2135    
2136     (d) Implemented a pcregrep test script.
2137    
2138     (e) Added the -M (multiline match) option. This allows patterns to match
2139     over several lines of the subject. The buffering ensures that at least
2140     8K, or the rest of the document (whichever is the shorter) is available
2141     for matching (and similarly the previous 8K for lookbehind assertions).
2142    
2143     (f) Changed the --help output so that it now says
2144    
2145     -w, --word-regex(p)
2146    
2147     instead of two lines, one with "regex" and the other with "regexp"
2148     because that confused at least one person since the short forms are the
2149     same. (This required a bit of code, as the output is generated
2150     automatically from a table. It wasn't just a text change.)
2151    
2152     (g) -- can be used to terminate pcregrep options if the next thing isn't an
2153     option but starts with a hyphen. Could be a pattern or a path name
2154     starting with a hyphen, for instance.
2155    
2156     (h) "-" can be given as a file name to represent stdin.
2157    
2158     (i) When file names are being printed, "(standard input)" is used for
2159     the standard input, for compatibility with GNU grep. Previously
2160     "<stdin>" was used.
2161    
2162     (j) The option --label=xxx can be used to supply a name to be used for
2163     stdin when file names are being printed. There is no short form.
2164    
2165     (k) Re-factored the options decoding logic because we are going to add
2166     two more options that take data. Such options can now be given in four
2167     different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2168    
2169     (l) Added the -A, -B, and -C options for requesting that lines of context
2170     around matches be printed.
2171    
2172     (m) Added the -L option to print the names of files that do not contain
2173     any matching lines, that is, the complement of -l.
2174    
2175     (n) The return code is 2 if any file cannot be opened, but pcregrep does
2176     continue to scan other files.
2177    
2178     (o) The -s option was incorrectly implemented. For compatibility with other
2179     greps, it now suppresses the error message for a non-existent or non-
2180     accessible file (but not the return code). There is a new option called
2181     -q that suppresses the output of matching lines, which was what -s was
2182     previously doing.
2183    
2184     (p) Added --include and --exclude options to specify files for inclusion
2185     and exclusion when recursing.
2186    
2187     11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2188     Hopefully, it now does.
2189    
2190     12. Missing cast in pcre_study().
2191    
2192     13. Added an "uninstall" target to the makefile.
2193    
2194     14. Replaced "extern" in the function prototypes in Makefile.in with
2195     "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2196     world, but is set differently for Windows.
2197    
2198     15. Added a second compiling function called pcre_compile2(). The only
2199     difference is that it has an extra argument, which is a pointer to an
2200     integer error code. When there is a compile-time failure, this is set
2201     non-zero, in addition to the error test pointer being set to point to an
2202     error message. The new argument may be NULL if no error number is required
2203     (but then you may as well call pcre_compile(), which is now just a
2204     wrapper). This facility is provided because some applications need a
2205     numeric error indication, but it has also enabled me to tidy up the way
2206     compile-time errors are handled in the POSIX wrapper.
2207    
2208     16. Added VPATH=.libs to the makefile; this should help when building with one
2209     prefix path and installing with another. (Or so I'm told by someone who
2210     knows more about this stuff than I do.)
2211    
2212     17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2213     passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2214     match everything, including newlines. This is not POSIX-compatible, but
2215     somebody wanted the feature. From pcretest it can be activated by using
2216     both the P and the s flags.
2217    
2218     18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2219    
2220     19. libpcre.pc was being incorrectly installed as executable.
2221    
2222     20. A couple of places in pcretest check for end-of-line by looking for '\n';
2223     it now also looks for '\r' so that it will work unmodified on Windows.
2224    
2225     21. Added Google's contributed C++ wrapper to the distribution.
2226    
2227     22. Added some untidy missing memory free() calls in pcretest, to keep
2228     Electric Fence happy when testing.
2229    
2230    
2231    
2232 nigel 75 Version 5.0 13-Sep-04
2233     ---------------------
2234    
2235     1. Internal change: literal characters are no longer packed up into items
2236     containing multiple characters in a single byte-string. Each character
2237     is now matched using a separate opcode. However, there may be more than one
2238     byte in the character in UTF-8 mode.
2239    
2240     2. The pcre_callout_block structure has two new fields: pattern_position and
2241     next_item_length. These contain the offset in the pattern to the next match
2242     item, and its length, respectively.
2243    
2244     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2245     insertion of callouts before each pattern item. Added the /C option to
2246     pcretest to make use of this.
2247    
2248     4. On the advice of a Windows user, the lines
2249    
2250     #if defined(_WIN32) || defined(WIN32)
2251     _setmode( _fileno( stdout ), 0x8000 );
2252     #endif /* defined(_WIN32) || defined(WIN32) */
2253    
2254     have been added to the source of pcretest. This apparently does useful
2255     magic in relation to line terminators.
2256    
2257     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2258     for the benefit of those environments where the "b" makes a difference.
2259    
2260     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2261     to know about it. I have put a hack into configure.in that adds in code
2262     to set GCC=yes if CC=icc. This seems to end up at a point in the
2263     generated configure script that is early enough to affect the setting of
2264     compiler options, which is what is needed, but I have no means of testing
2265     whether it really works. (The user who reported this had patched the
2266     generated configure script, which of course I cannot do.)
2267    
2268     LATER: After change 22 below (new libtool files), the configure script
2269     seems to know about icc (and also ecc). Therefore, I have commented out
2270     this hack in configure.in.
2271    
2272     7. Added support for pkg-config (2 patches were sent in).
2273    
2274     8. Negated POSIX character classes that used a combination of internal tables
2275     were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2276     [[:^ascii]]. Typically, they would match almost any characters. The other
2277     POSIX classes were not broken in this way.
2278    
2279     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2280     to find the match, as PCRE was deluded into thinking that the match had to
2281     start at the start point or following a newline. The same bug applied to
2282     patterns with negative forward assertions or any backward assertions
2283     preceding ".*" at the start, unless the pattern required a fixed first
2284     character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2285    
2286     10. In UTF-8 mode, when moving forwards in the subject after a failed match
2287     starting at the last subject character, bytes beyond the end of the subject
2288     string were read.
2289    
2290     11. Renamed the variable "class" as "classbits" to make life easier for C++
2291     users. (Previously there was a macro definition, but it apparently wasn't
2292     enough.)
2293    
2294     12. Added the new field "tables" to the extra data so that tables can be passed
2295     in at exec time, or the internal tables can be re-selected. This allows
2296     a compiled regex to be saved and re-used at a later time by a different
2297     program that might have everything at different addresses.
2298    
2299     13. Modified the pcre-config script so that, when run on Solaris, it shows a
2300     -R library as well as a -L library.
2301    
2302     14. The debugging options of pcretest (-d on the command line or D on a
2303     pattern) showed incorrect output for anything following an extended class
2304     that contained multibyte characters and which was followed by a quantifier.
2305    
2306     15. Added optional support for general category Unicode character properties
2307     via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2308     support. It adds about 90K to the size of the library. The meanings of the
2309     inbuilt class escapes such as \d and \s have NOT been changed.
2310    
2311     16. Updated pcredemo.c to include calls to free() to release the memory for the
2312     compiled pattern.
2313    
2314     17. The generated file chartables.c was being created in the source directory
2315     instead of in the building directory. This caused the build to fail if the
2316     source directory was different from the building directory, and was
2317     read-only.
2318    
2319     18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2320     file. No doubt somebody will tell me if they don't make sense... Also added
2321     Dan Mooney's comments about building on OpenVMS.
2322    
2323     19. Added support for partial matching via the PCRE_PARTIAL option for
2324     pcre_exec() and the \P data escape in pcretest.
2325    
2326     20. Extended pcretest with 3 new pattern features:
2327    
2328     (i) A pattern option of the form ">rest-of-line" causes pcretest to
2329     write the compiled pattern to the file whose name is "rest-of-line".
2330     This is a straight binary dump of the data, with the saved pointer to
2331     the character tables forced to be NULL. The study data, if any, is
2332     written too. After writing, pcretest reads a new pattern.
2333    
2334     (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2335     compiled pattern from the given file. There must not be any
2336     occurrences of "<" in the file name (pretty unlikely); if there are,
2337     pcretest will instead treat the initial "<" as a pattern delimiter.
2338     After reading in the pattern, pcretest goes on to read data lines as
2339     usual.
2340    
2341     (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2342     and 16-bit fields in a compiled pattern, to simulate a pattern that
2343     was compiled on a host of opposite endianness.
2344    
2345     21. The pcre-exec() function can now cope with patterns that were compiled on
2346     hosts of opposite endianness, with this restriction:
2347    
2348     As for any compiled expression that is saved and used later, the tables
2349     pointer field cannot be preserved; the extra_data field in the arguments
2350     to pcre_exec() should be used to pass in a tables address if a value
2351     other than the default internal tables were used at compile time.
2352    
2353     22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2354     now diagnosed as an error. Previously, most of the time, a negative number
2355     would have been treated as zero, but if in addition "ovector" was passed as
2356     NULL, a crash could occur.
2357    
2358     23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2359     new versions from the libtool 1.5 distribution (the last one is a copy of
2360     a file called libtool.m4). This seems to have fixed the need to patch
2361     "configure" to support Darwin 1.3 (which I used to do). However, I still
2362     had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2363     workstation).
2364    
2365     24. Changed the PCRE licence to be the more standard "BSD" licence.
2366    
2367    
2368 nigel 73 Version 4.5 01-Dec-03
2369     ---------------------
2370    
2371     1. There has been some re-arrangement of the code for the match() function so
2372     that it can be compiled in a version that does not call itself recursively.
2373     Instead, it keeps those local variables that need separate instances for
2374     each "recursion" in a frame on the heap, and gets/frees frames whenever it
2375     needs to "recurse". Keeping track of where control must go is done by means
2376     of setjmp/longjmp. The whole thing is implemented by a set of macros that
2377     hide most of the details from the main code, and operates only if
2378     NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2379     "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2380     operating.
2381    
2382     To make it easier for callers to provide specially tailored get/free
2383     functions for this usage, two new functions, pcre_stack_malloc, and
2384     pcre_stack_free, are used. They are always called in strict stacking order,
2385     and the size of block requested is always the same.
2386    
2387     The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2388     PCRE has been compiled to use the stack or the heap for recursion. The
2389     -C option of pcretest uses this to show which version is compiled.
2390    
2391     A new data escape \S, is added to pcretest; it causes the amounts of store
2392     obtained and freed by both kinds of malloc/free at match time to be added
2393     to the output.
2394    
2395     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2396     what's available on my current Linux desktop machine.
2397    
2398     3. When matching a UTF-8 string, the test for a valid string at the start has
2399     been extended. If start_offset is not zero, PCRE now checks that it points
2400     to a byte that is the start of a UTF-8 character. If not, it returns
2401     PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2402     this is necessary because there may be backward assertions in the pattern.
2403     When matching the same subject several times, it may save resources to use
2404     PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2405    
2406     4. The code for checking the validity of UTF-8 strings has been tightened so
2407     that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2408     containing "overlong sequences".
2409    
2410     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2411     I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2412     should have been "&", but it just so happened that all the cases this let
2413     through by mistake were picked up later in the function.
2414    
2415     6. I had used a variable called "isblank" - this is a C99 function, causing
2416     some compilers to warn. To avoid this, I renamed it (as "blankclass").
2417    
2418     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2419     prompting; (b) run "./pcretest /dev/null" at the start of the test script
2420     so the version is shown; (c) stop "make test" echoing "./RunTest".
2421    
2422     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2423    
2424     9. The prototype for memmove() for systems that don't have it was using
2425     size_t, but the inclusion of the header that defines size_t was later. I've
2426     moved the #includes for the C headers earlier to avoid this.
2427    
2428     10. Added some adjustments to the code to make it easier to compiler on certain
2429     special systems:
2430    
2431     (a) Some "const" qualifiers were missing.
2432     (b) Added the macro EXPORT before all exported functions; by default this
2433     is defined to be empty.
2434     (c) Changed the dftables auxiliary program (that builds chartables.c) so
2435     that it reads its output file name as an argument instead of writing
2436     to the standard output and assuming this can be redirected.
2437    
2438     11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2439     class containing characters with values greater than 255, PCRE compilation
2440     went into a loop.
2441    
2442     12. A recursive reference to a subpattern that was within another subpattern
2443     that had a minimum quantifier of zero caused PCRE to crash. For example,
2444     (x(y(?2))z)? provoked this bug with a subject that got as far as the
2445     recursion. If the recursively-called subpattern itself had a zero repeat,
2446     that was OK.
2447    
2448     13. In pcretest, the buffer for reading a data line was set at 30K, but the
2449     buffer into which it was copied (for escape processing) was still set at
2450     1024, so long lines caused crashes.
2451    
2452     14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2453     "internal error: code overflow...". This applied to any character class
2454     that was followed by a possessive quantifier.
2455    
2456     15. Modified the Makefile to add libpcre.la as a prerequisite for
2457     libpcreposix.la because I was told this is needed for a parallel build to
2458     work.
2459    
2460     16. If a pattern that contained .* following optional items at the start was
2461     studied, the wrong optimizing data was generated, leading to matching
2462     errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2463     matching string must start with a or b or c. The correct conclusion for
2464     this pattern is that a match can start with any character.
2465    
2466    
2467 nigel 71 Version 4.4 13-Aug-03
2468     ---------------------
2469    
2470     1. In UTF-8 mode, a character class containing characters with values between
2471     127 and 255 was not handled correctly if the compiled pattern was studied.
2472     In fixing this, I have also improved the studying algorithm for such
2473     classes (slightly).
2474    
2475     2. Three internal functions had redundant arguments passed to them. Removal
2476     might give a very teeny performance improvement.
2477    
2478     3. Documentation bug: the value of the capture_top field in a callout is *one
2479     more than* the number of the hightest numbered captured substring.
2480    
2481     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2482     in incorrectly linking with a previously installed version. They now link
2483     explicitly with libpcre.la.
2484    
2485     5. configure.in no longer needs to recognize Cygwin specially.
2486    
2487     6. A problem in pcre.in for Windows platforms is fixed.
2488    
2489     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2490     pcretest, it used to include the size of the study block as part of its
2491     output. Unfortunately, the structure contains a field that has a different
2492     size on different hardware architectures. This meant that the tests that
2493     showed this size failed. As the block is currently always of a fixed size,
2494     this information isn't actually particularly useful in pcretest output, so
2495     I have just removed it.
2496    
2497     8. Three pre-processor statements accidentally did not start in column 1.
2498     Sadly, there are *still* compilers around that complain, even though
2499     standard C has not required this for well over a decade. Sigh.
2500    
2501     9. In pcretest, the code for checking callouts passed small integers in the
2502     callout_data field, which is a void * field. However, some picky compilers
2503     complained about the casts involved for this on 64-bit systems. Now
2504     pcretest passes the address of the small integer instead, which should get
2505     rid of the warnings.
2506    
2507     10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2508     both compile and run time, and gives an error if an invalid UTF-8 sequence
2509     is found. There is a option for disabling this check in cases where the
2510     string is known to be correct and/or the maximum performance is wanted.
2511    
2512     11. In response to a bug report, I changed one line in Makefile.in from
2513    
2514     -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2515     to
2516     -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2517    
2518     to look similar to other lines, but I have no way of telling whether this
2519     is the right thing to do, as I do not use Windows. No doubt I'll get told
2520     if it's wrong...
2521    
2522    
2523 nigel 69 Version 4.3 21-May-03
2524     ---------------------
2525    
2526     1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2527     Makefile.
2528    
2529     2. Some refactoring to improve the quality of the code:
2530    
2531     (i) The utf8_table... variables are now declared "const".
2532    
2533     (ii) The code for \cx, which used the "case flipping" table to upper case
2534     lower case letters, now just substracts 32. This is ASCII-specific,
2535     but the whole concept of \cx is ASCII-specific, so it seems
2536     reasonable.
2537    
2538     (iii) PCRE was using its character types table to recognize decimal and
2539     hexadecimal digits in the pattern. This is silly, because it handles
2540     only 0-9, a-f, and A-F, but the character types table is locale-
2541     specific, which means strange things might happen. A private
2542     table is now used for this - though it costs 256 bytes, a table is
2543     much faster than multiple explicit tests. Of course, the standard
2544     character types table is still used for matching digits in subject
2545     strings against \d.
2546    
2547     (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2548     ending in _t are). So I've renamed it as ESC_tee.
2549    
2550     3. The first argument for regexec() in the POSIX wrapper should have been
2551     defined as "const".
2552    
2553     4. Changed pcretest to use malloc() for its buffers so that they can be
2554     Electric Fenced for debugging.
2555    
2556     5. There were several places in the code where, in UTF-8 mode, PCRE would try
2557     to read one or more bytes before the start of the subject string. Often this
2558     had no effect on PCRE's behaviour, but in some circumstances it could
2559     provoke a segmentation fault.
2560    
2561     6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2562     to try to read one or more bytes before the start of the subject string.
2563    
2564     7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2565     UTF-8 support could misbehave in various ways if the subject string
2566     contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2567     area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2568     back over UTF-8 characters.)
2569    
2570    
2571 nigel 67 Version 4.2 14-Apr-03
2572     ---------------------
2573    
2574     1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2575    
2576     2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2577     [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2578     [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2579     [WIN_PREFIX]: new variable, "cyg" for Cygwin
2580     * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2581     and BUILD_EXEEXT
2582     Note: automatic setting of the BUILD variables is not yet working
2583     set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2584     compile-time but not at link-time
2585     [LINK]: use for linking executables only
2586     make different versions for Windows and non-Windows
2587     [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2588     libraries
2589     [LINK_FOR_BUILD]: new variable
2590     [OBJEXT]: use throughout
2591     [EXEEXT]: use throughout
2592     <winshared>: new target
2593     <wininstall>: new target
2594     <dftables.o>: use native compiler
2595     <dftables>: use native linker
2596     <install>: handle Windows platform correctly
2597     <clean>: ditto
2598     <check>: ditto
2599     copy DLL to top builddir before testing
2600    
2601     As part of these changes, -no-undefined was removed again. This was reported
2602     to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2603     in any case.
2604    
2605     3. Some tidies to get rid of compiler warnings:
2606    
2607     . In the match_data structure, match_limit was an unsigned long int, whereas
2608     match_call_count was an int. I've made them both unsigned long ints.
2609    
2610     . In pcretest the fact that a const uschar * doesn't automatically cast to
2611     a void * provoked a warning.
2612    
2613     . Turning on some more compiler warnings threw up some "shadow" variables
2614     and a few more missing casts.
2615    
2616     4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2617     option, a class that contained a single character with a value between 128
2618     and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2619    
2620     5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2621     option, a class that contained several characters, but with at least one
2622     whose value was between 128 and 255 caused PCRE to crash.
2623    
2624    
2625 nigel 65 Version 4.1 12-Mar-03
2626     ---------------------
2627 nigel 63
2628 nigel 65 1. Compiling with gcc -pedantic found a couple of places where casts were
2629     needed, and a string in dftables.c that was longer than standard compilers are
2630     required to support.
2631    
2632     2. Compiling with Sun's compiler found a few more places where the code could
2633     be tidied up in order to avoid warnings.
2634    
2635     3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2636     first of these names is deprecated in the latest Autoconf in favour of the name
2637     CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2638     compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2639     analogy I have changed it to CFLAGS_FOR_BUILD.
2640    
2641     4. Added -no-undefined to the linking command in the Makefile, because this is
2642     apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2643     linking step for the pcreposix library.
2644    
2645     5. PCRE was failing to diagnose the case of two named groups with the same
2646     name.
2647    
2648     6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2649     literal character that is needed in the subject for a match, and scans along to
2650     ensure that it is present before embarking on the full matching process. This
2651     saves time in cases of nested unlimited repeats that are never going to match.
2652     Problem: the scan can take a lot of time if the subject is very long (e.g.
2653     megabytes), thus penalizing straightforward matches. It is now done only if the
2654     amount of subject to be scanned is less than 1000 bytes.
2655    
2656     7. A lesser problem with the same optimization is that it was recording the
2657     first character of an anchored pattern as "needed", thus provoking a search
2658     right along the subject, even when the first match of the pattern was going to
2659     fail. The "needed" character is now not set for anchored patterns, unless it
2660     follows something in the pattern that is of non-fixed length. Thus, it still
2661     fulfils its original purpose of finding quick non-matches in cases of nested
2662     unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2663    
2664    
2665     Version 4.0 17-Feb-03
2666     ---------------------
2667    
2668 nigel 63 1. If a comment in an extended regex that started immediately after a meta-item
2669     extended to the end of string, PCRE compiled incorrect data. This could lead to
2670     all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2671    
2672     2. Moved to autoconf 2.53 and libtool 1.4.2.
2673    
2674     3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2675     the special perltest8 script is no longer needed - all the tests can be run
2676     from a single perltest script.
2677    
2678     4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2679     by \s. It has now been removed in PCRE. This means it isn't recognized as
2680     whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2681     class [:space:] *does* include VT, thereby creating a mess.
2682    
2683     5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2684     space and tab.
2685    
2686     6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2687     its new features into the main test script, reducing the number of scripts.
2688    
2689     7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2690     were backward compatible, and made the (?i) apply to the whole pattern, as if
2691     /i were given. Now it behaves more logically, and applies the option setting
2692     only to what follows. PCRE has been changed to follow suit. However, if it
2693     finds options settings right at the start of the pattern, it extracts them into
2694     the global options, as before. Thus, they show up in the info data.
2695    
2696     8. Added support for the \Q...\E escape sequence. Characters in between are
2697     treated as literals. This is slightly different from Perl in that $ and @ are
2698     also handled as literals inside the quotes. In Perl, they will cause variable
2699     interpolation. Note the following examples:
2700    
2701     Pattern PCRE matches Perl matches
2702    
2703     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2704     \Qabc\$xyz\E abc\$xyz abc\$xyz
2705     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2706    
2707     For compatibility with Perl, \Q...\E sequences are recognized inside character
2708     classes as well as outside them.
2709    
2710     9. Re-organized 3 code statements in pcretest to avoid "overflow in
2711     floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2712     (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2713     signed/unsigned warnings.
2714    
2715     10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2716     option for pcretest, so I've replaced it by a simple function that does just
2717     that job.
2718    
2719     11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2720     "pcregrep -".
2721    
2722     12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2723     Java package. This provides some syntactic sugar for simple cases of what my
2724     documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2725     as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2726     item, you can use this simplified notation. Note that only makes sense with
2727     greedy quantifiers. Consequently, the use of the possessive quantifier forces
2728     greediness, whatever the setting of the PCRE_UNGREEDY option.
2729    
2730     13. A change of greediness default within a pattern was not taking effect at
2731     the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2732     subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2733     was abstracted outside.
2734    
2735     14. PCRE now supports the \G assertion. It is true when the current matching
2736     position is at the start point of the match. This differs from \A when the
2737     starting offset is non-zero. Used with the /g option of pcretest (or similar
2738     code), it works in the same way as it does for Perl's /g option. If all
2739     alternatives of a regex begin with \G, the expression is anchored to the start
2740     match position, and the "anchored" flag is set in the compiled expression.
2741    
2742     15. Some bugs concerning the handling of certain option changes within patterns
2743     have been fixed. These applied to options other than (?ims). For example,
2744     "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2745     been the other way round. Some of this was related to change 7 above.
2746    
2747     16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2748     features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2749     and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2750     POSIX classes only within a class (e.g. /[[:alpha:]]/).
2751    
2752     17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2753     mode. Unlike ".", it always matches newline, whatever the setting of
2754     PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2755     assertions. Perl allows it, but it doesn't (in general) work because it can't
2756     calculate the length of the lookbehind. At least, that's the case for Perl
2757     5.8.0 - I've been told they are going to document that it doesn't work in
2758     future.
2759    
2760     18. Added an error diagnosis for escapes that PCRE does not support: these are
2761     \L, \l, \N, \P, \p, \U, \u, and \X.
2762    
2763     19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2764     reading past the end of the pattern in cases such as /[abcd/.
2765    
2766     20. PCRE was getting more memory than necessary for patterns with classes that
2767     contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2768    
2769     21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2770     compiling PCRE for use with Virtual Pascal.
2771    
2772     22. Small fix to the Makefile to make it work properly if the build is done
2773     outside the source tree.
2774    
2775     23. Added a new extension: a condition to go with recursion. If a conditional
2776     subpattern starts with (?(R) the "true" branch is used if recursion has
2777     happened, whereas the "false" branch is used only at the top level.
2778    
2779     24. When there was a very long string of literal characters (over 255 bytes
2780     without UTF support, over 250 bytes with UTF support), the computation of how
2781     much memory was required could be incorrect, leading to segfaults or other
2782     strange effects.
2783    
2784     25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2785     start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2786     there was a subsequent back reference to those brackets. This meant that, for
2787     example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2788     possible to check for precisely this case. All we can do is abandon the
2789     optimization if .* occurs inside capturing brackets when there are any back
2790     references whatsoever. (See below for a better fix that came later.)
2791    
2792     26. The handling of the optimization for finding the first character of a
2793     non-anchored pattern, and for finding a character that is required later in the
2794     match were failing in some cases. This didn't break the matching; it just
2795     failed to optimize when it could. The way this is done has been re-implemented.
2796    
2797     27. Fixed typo in error message for invalid (?R item (it said "(?p").
2798    
2799     28. Added a new feature that provides some of the functionality that Perl
2800     provides with (?{...}). The facility is termed a "callout". The way it is done
2801     in PCRE is for the caller to provide an optional function, by setting
2802     pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2803     global variable. By default it is unset, which disables all calling out. To get
2804     the function called, the regex must include (?C) at appropriate points. This
2805     is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2806     This provides a means of identifying different callout points. When PCRE
2807     reaches such a point in the regex, if pcre_callout has been set, the external
2808     function is called. It is provided with data in a structure called
2809     pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2810     matching continues; if it returns a non-zero value, the match at the current
2811     point fails. However, backtracking will occur if possible. [This was changed
2812     later and other features added - see item 49 below.]
2813    
2814     29. pcretest is upgraded to test the callout functionality. It provides a
2815     callout function that displays information. By default, it shows the start of
2816     the match and the current position in the text. There are some new data escapes
2817     to vary what happens:
2818    
2819     \C+ in addition, show current contents of captured substrings
2820     \C- do not supply a callout function
2821     \C!n return 1 when callout number n is reached
2822     \C!n!m return 1 when callout number n is reached for the mth time
2823    
2824     30. If pcregrep was called with the -l option and just a single file name, it
2825     output "<stdin>" if a match was found, instead of the file name.
2826    
2827     31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2828     slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2829     pcre_exec(). This saves a malloc/free per call. The default value of
2830     POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2831     when configuring.
2832    
2833     32. The default maximum size of a compiled pattern is 64K. There have been a
2834     few cases of people hitting this limit. The code now uses macros to handle the
2835     storing of links as offsets within the compiled pattern. It defaults to 2-byte
2836     links, but this can be changed to 3 or 4 bytes by --with-link-size when
2837     configuring. Tests 2 and 5 work only with 2-byte links because they output
2838     debugging information about compiled patterns.
2839    
2840     33. Internal code re-arrangements:
2841    
2842     (a) Moved the debugging function for printing out a compiled regex into
2843     its own source file (printint.c) and used #include to pull it into
2844     pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2845     separate copies.
2846    
2847     (b) Defined the list of op-code names for debugging as a macro in
2848     internal.h so that it is next to the definition of the opcodes.
2849    
2850     (c) Defined a table of op-code lengths for simpler skipping along compiled
2851     code. This is again a macro in internal.h so that it is next to the
2852     definition of the opcodes.
2853    
2854     34. Added support for recursive calls to individual subpatterns, along the
2855     lines of Robin Houston's patch (but implemented somewhat differently).
2856    
2857     35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2858     allow it to read and process whole directories in Win32. This code was
2859     contributed by Lionel Fourquaux; it has not been tested by me.
2860    
2861     36. Added support for named subpatterns. The Python syntax (?P<name>...) is
2862     used to name a group. Names consist of alphanumerics and underscores, and must
2863     be unique. Back references use the syntax (?P=name) and recursive calls use
2864     (?P>name) which is a PCRE extension to the Python extension. Groups still have
2865     numbers. The function pcre_fullinfo() can be used after compilation to extract
2866     a name/number map. There are three relevant calls:
2867    
2868     PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
2869     PCRE_INFO_NAMECOUNT yields the number of entries
2870     PCRE_INFO_NAMETABLE yields a pointer to the map.
2871    
2872     The map is a vector of fixed-size entries. The size of each entry depends on
2873     the length of the longest name used. The first two bytes of each entry are the
2874     group number, most significant byte first. There follows the corresponding
2875     name, zero terminated. The names are in alphabetical order.
2876    
2877     37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
2878     case instead of 255. Making it the same both with and without UTF-8 support
2879     means that the same test output works with both.
2880    
2881     38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
2882     calling malloc() with a zero argument.
2883    
2884     39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
2885     optimization. I've improved things by keeping a bitmap of backreferences with
2886     numbers 1-31 so that if .* occurs inside capturing brackets that are not in
2887     fact referenced, the optimization can be applied. It is unlikely that a
2888     relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
2889     the match to follow \n) will appear inside brackets with a number greater than
2890     31, but if it does, any back reference > 31 suppresses the optimization.
2891    
2892     40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
2893     of disabling numbered capturing parentheses. Any opening parenthesis that is
2894     not followed by ? behaves as if it were followed by ?: but named parentheses
2895     can still be used for capturing (and they will acquire numbers in the usual
2896     way).
2897    
2898     41. Redesigned the return codes from the match() function into yes/no/error so
2899     that errors can be passed back from deep inside the nested calls. A malloc
2900     failure while inside a recursive subpattern call now causes the
2901     PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
2902    
2903     42. It is now possible to set a limit on the number of times the match()
2904     function is called in a call to pcre_exec(). This facility makes it possible to
2905     limit the amount of recursion and backtracking, though not in a directly
2906     obvious way, because the match() function is used in a number of different
2907     circumstances. The count starts from zero for each position in the subject
2908     string (for non-anchored patterns). The default limit is, for compatibility, a
2909     large number, namely 10 000 000. You can change this in two ways:
2910    
2911     (a) When configuring PCRE before making, you can use --with-match-limit=n
2912     to set a default value for the compiled library.
2913    
2914     (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
2915     a different value is set. See 45 below.
2916    
2917     If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
2918    
2919     43. Added a new function pcre_config(int, void *) to enable run-time extraction
2920     of things that can be changed at compile time. The first argument specifies
2921     what is wanted and the second points to where the information is to be placed.
2922     The current list of available information is:
2923    
2924     PCRE_CONFIG_UTF8
2925    
2926     The output is an integer that is set to one if UTF-8 support is available;
2927     otherwise it is set to zero.
2928    
2929     PCRE_CONFIG_NEWLINE
2930    
2931     The output is an integer that it set to the value of the code that is used for
2932     newline. It is either LF (10) or CR (13).
2933    
2934     PCRE_CONFIG_LINK_SIZE
2935    
2936     The output is an integer that contains the number of bytes used for internal
2937     linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
2938    
2939     PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
2940    
2941     The output is an integer that contains the threshold above which the POSIX
2942     interface uses malloc() for output vectors. See item 31 above.
2943    
2944     PCRE_CONFIG_MATCH_LIMIT
2945    
2946     The output is an unsigned integer that contains the default limit of the number
2947     of match() calls in a pcre_exec() execution. See 42 above.
2948    
2949     44. pcretest has been upgraded by the addition of the -C option. This causes it
2950     to extract all the available output from the new pcre_config() function, and to
2951     output it. The program then exits immediately.
2952    
2953     45. A need has arisen to pass over additional data with calls to pcre_exec() in
2954     order to support additional features. One way would have been to define
2955     pcre_exec2() (for example) with extra arguments, but this would not have been
2956     extensible, and would also have required all calls to the original function to
2957     be mapped to the new one. Instead, I have chosen to extend the mechanism that
2958     is used for passing in "extra" data from pcre_study().
2959    
2960     The pcre_extra structure is now exposed and defined in pcre.h. It currently
2961     contains the following fields:
2962    
2963     flags a bitmap indicating which of the following fields are set
2964     study_data opaque data from pcre_study()
2965     match_limit a way of specifying a limit on match() calls for a specific
2966     call to pcre_exec()
2967     callout_data data for callouts (see 49 below)
2968    
2969     The flag bits are also defined in pcre.h, and are
2970    
2971     PCRE_EXTRA_STUDY_DATA
2972     PCRE_EXTRA_MATCH_LIMIT
2973     PCRE_EXTRA_CALLOUT_DATA
2974    
2975     The pcre_study() function now returns one of these new pcre_extra blocks, with
2976     the actual study data pointed to by the study_data field, and the
2977     PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
2978     before. That is, this change is entirely upwards-compatible and requires no
2979     change to existing code.
2980    
2981     If you want to pass in additional data to pcre_exec(), you can either place it
2982     in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
2983     block.
2984    
2985     46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
2986     data string contains the escape sequence \M, pcretest calls pcre_exec() several
2987     times with different match limits, until it finds the minimum value needed for
2988     pcre_exec() to complete. The value is then output. This can be instructive; for
2989     most simple matches the number is quite small, but for pathological cases it
2990     gets very large very quickly.
2991    
2992     47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
2993     returns the size of the data block pointed to by the study_data field in a
2994     pcre_extra block, that is, the value that was passed as the argument to
2995     pcre_malloc() when PCRE was getting memory in which to place the information
2996     created by pcre_study(). The fourth argument should point to a size_t variable.
2997     pcretest has been extended so that this information is shown after a successful
2998     pcre_study() call when information about the compiled regex is being displayed.
2999    
3000     48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3001     because what follows is always an absolute path. (Later: it turns out that this
3002     is more than cosmetic for MinGW, because it doesn't like empty path
3003     components.)
3004    
3005     49. Some changes have been made to the callout feature (see 28 above):
3006    
3007     (i) A callout function now has three choices for what it returns:
3008    
3009     0 => success, carry on matching
3010     > 0 => failure at this point, but backtrack if possible
3011     < 0 => serious error, return this value from pcre_exec()
3012    
3013     Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3014     values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3015     "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3016     use by callout functions. It will never be used by PCRE itself.
3017    
3018     (ii) The pcre_extra structure (see 45 above) has a void * field called
3019     callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3020     pcre_callout_block structure has a field of the same name. The contents of
3021     the field passed in the pcre_extra structure are passed to the callout
3022     function in the corresponding field in the callout block. This makes it
3023     easier to use the same callout-containing regex from multiple threads. For
3024     testing, the pcretest program has a new data escape
3025    
3026     \C*n pass the number n (may be negative) as callout_data
3027    
3028     If the callout function in pcretest receives a non-zero value as
3029     callout_data, it returns that value.
3030    
3031     50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3032     there were some redundant $(CFLAGS) in commands that are now specified as
3033     $(LINK), which already includes $(CFLAGS).
3034    
3035     51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3036     has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3037     with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3038     one-byte characters throughout. Note that case-insensitive matching applies
3039     only to characters whose values are less than 256. PCRE doesn't support the
3040     notion of cases for higher-valued characters.
3041    
3042     (i) A character class whose characters are all within 0-255 is handled as
3043     a bit map, and the map is inverted for negative classes. Previously, a
3044     character > 255 always failed to match such a class; however it should
3045     match if the class was a negative one (e.g. [^ab]). This has been fixed.
3046    
3047     (ii) A negated character class with a single character < 255 is coded as
3048     "not this character" (OP_NOT). This wasn't working properly when the test
3049     character was multibyte, either singly or repeated.
3050    
3051     (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3052     mode, for example: \x{100}{2,3}.
3053    
3054     (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3055     singly or repeated) now correctly test multibyte characters. However,
3056     PCRE doesn't recognize any characters with values greater than 255 as
3057     digits, spaces, or word characters. Such characters always match \D, \S,
3058     and \W, and never match \d, \s, or \w.
3059    
3060     (v) Classes may now contain characters and character ranges with values
3061     greater than 255. For example: [ab\x{100}-\x{400}].
3062    
3063     (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3064     PCRE in UTF-8 mode.
3065    
3066     52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3067     PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3068     retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3069     value.)
3070    
3071     53. The single man page has become too large. I have therefore split it up into
3072     a number of separate man pages. These also give rise to individual HTML pages;
3073     these are now put in a separate directory, and there is an index.html page that
3074     lists them all. Some hyperlinking between the pages has been installed.
3075    
3076     54. Added convenience functions for handling named capturing parentheses.
3077    
3078     55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3079     aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3080     true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3081     are faulted.
3082    
3083     56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3084     calling configure. These values are used when compiling the dftables.c program
3085     which is run to generate the source of the default character tables. They
3086     default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3087     you will need to set these values.
3088    
3089     57. Updated the building process for Windows DLL, as provided by Fred Cox.
3090    
3091    
3092     Version 3.9 02-Jan-02
3093 nigel 61 ---------------------
3094    
3095     1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3096    
3097     2. If --disable-static was given, the building process failed when trying to
3098     build pcretest and pcregrep. (For some reason it was using libtool to compile
3099     them, which is not right, as they aren't part of the library.)
3100    
3101    
3102 nigel 59 Version 3.8 18-Dec-01
3103     ---------------------
3104    
3105     1. The experimental UTF-8 code was completely screwed up. It was packing the
3106     bytes in the wrong order. How dumb can you get?
3107    
3108    
3109 nigel 57 Version 3.7 29-Oct-01
3110     ---------------------
3111    
3112     1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3113     This caused pcretest, when used on the test data, to segfault. Unfortunately,
3114     this didn't happen under Solaris 8, where I normally test things.
3115    
3116 nigel 59 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3117     doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3118     isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3119     this fix an hour or so after the initial 3.7 release.)
3120 nigel 57
3121 nigel 59
3122 nigel 55 Version 3.6 23-Oct-01
3123     ---------------------
3124    
3125     1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3126     offsets passed as NULL with zero offset count.
3127    
3128     2. The config.guess and config.sub files had not been updated when I moved to
3129     the latest autoconf.
3130    
3131    
3132 nigel 53 Version 3.5 15-Aug-01
3133     ---------------------
3134 nigel 5
3135 nigel 53 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3136     had been forgotten.
3137    
3138     2. By using declared but undefined structures, we can avoid using "void"
3139     definitions in pcre.h while keeping the internal definitions of the structures
3140     private.
3141    
3142     3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3143     user point of view, this means that both static and shared libraries are built
3144     by default, but this can be individually controlled. More of the work of
3145     handling this static/shared cases is now inside libtool instead of PCRE's make
3146     file.
3147    
3148     4. The pcretest utility is now installed along with pcregrep because it is
3149     useful for users (to test regexs) and by doing this, it automatically gets
3150     relinked by libtool. The documentation has been turned into a man page, so
3151     there are now .1, .txt, and .html versions in /doc.
3152    
3153     5. Upgrades to pcregrep:
3154     (i) Added long-form option names like gnu grep.
3155     (ii) Added --help to list all options with an explanatory phrase.
3156     (iii) Added -r, --recursive to recurse into sub-directories.
3157     (iv) Added -f, --file to read patterns from a file.
3158    
3159     6. pcre_exec() was referring to its "code" argument before testing that
3160     argument for NULL (and giving an error if it was NULL).
3161    
3162     7. Upgraded Makefile.in to allow for compiling in a different directory from
3163     the source directory.
3164    
3165     8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3166     options bits, the pointer it was passed was to an int instead of to an unsigned
3167     long int. This mattered only on 64-bit systems.
3168    
3169     9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3170     generated) instead of pcre.in, which it its source. Also made the same change
3171     in several of the .c files.
3172    
3173     10. A new release of gcc defines printf() as a macro, which broke pcretest
3174     because it had an ifdef in the middle of a string argument for printf(). Fixed
3175     by using separate calls to printf().
3176    
3177     11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3178     script, to force use of CR or LF instead of \n in the source. On non-Unix
3179     systems, the value can be set in config.h.
3180    
3181     12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3182     absolute limit. Changed the text of the error message to make this clear, and
3183     likewise updated the man page.
3184    
3185     13. The limit of 99 on the number of capturing subpatterns has been removed.
3186     The new limit is 65535, which I hope will not be a "real" limit.
3187    
3188    
3189 nigel 51 Version 3.4 22-Aug-00
3190     ---------------------
3191    
3192     1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3193    
3194     2. Diagnose condition (?(0) as an error instead of crashing on matching.
3195    
3196    
3197 nigel 49 Version 3.3 01-Aug-00
3198     ---------------------
3199    
3200     1. If an octal character was given, but the value was greater than \377, it
3201     was not getting masked to the least significant bits, as documented. This could
3202     lead to crashes in some systems.
3203    
3204     2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3205     the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3206    
3207     3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3208     These just pass their arguments on to (pcre_free)(), but they are provided
3209     because some uses of PCRE bind it to non-C systems that can call its functions,
3210     but cannot call free() or pcre_free() directly.
3211    
3212     4. Add "make test" as a synonym for "make check". Corrected some comments in
3213     the Makefile.
3214    
3215     5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3216     Makefile.
3217    
3218     6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3219     command called pgrep for grepping around the active processes.
3220    
3221     7. Added the beginnings of support for UTF-8 character strings.
3222    
3223     8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3224     RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3225     the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3226     out for the ar command.)
3227    
3228    
3229 nigel 47 Version 3.2 12-May-00
3230     ---------------------
3231    
3232     This is purely a bug fixing release.
3233    
3234     1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3235     of ZA. This was just one example of several cases that could provoke this bug,
3236     which was introduced by change 9 of version 2.00. The code for breaking
3237     infinite loops after an iteration that matches an empty string was't working
3238     correctly.
3239    
3240     2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3241     when matched against abbab (for example). After matching an empty string, it
3242     wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3243     caused it to match further down the string than it should.
3244    
3245     3. The code contained an inclusion of sys/types.h. It isn't clear why this
3246     was there because it doesn't seem to be needed, and it causes trouble on some
3247     systems, as it is not a Standard C header. It has been removed.
3248    
3249     4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3250     were reported on the Macintosh. The changes were from
3251    
3252     while ((c = *(++ptr)) != 0 && c != '\n');
3253     to
3254     while ((c = *(++ptr)) != 0 && c != '\n') ;
3255    
3256     Totally extraordinary, but if that's what it takes...
3257    
3258     5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3259     available. Added HAVE_BCOPY and an autoconf test for it; if neither
3260     HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3261     assumes the way PCRE uses memmove() (always moving upwards).
3262    
3263     6. PCRE is being used in one environment where strchr() is not available. There
3264     was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3265     faster code anyway.
3266    
3267    
3268 nigel 45 Version 3.1 09-Feb-00
3269     ---------------------
3270    
3271     The only change in this release is the fixing of some bugs in Makefile.in for
3272     the "install" target:
3273    
3274     (1) It was failing to install pcreposix.h.
3275    
3276     (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3277    
3278    
3279 nigel 43 Version 3.0 01-Feb-00
3280     ---------------------
3281 nigel 41
3282     1. Add support for the /+ modifier to perltest (to output $` like it does in
3283     pcretest).
3284    
3285     2. Add support for the /g modifier to perltest.
3286    
3287     3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3288     matches null strings.
3289    
3290     4. Fix perltest so that it doesn't do unwanted things when fed an empty
3291     pattern. Perl treats empty patterns specially - it reuses the most recent
3292     pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3293     effect.
3294    
3295     5. The POSIX interface was broken in that it was just handing over the POSIX
3296     captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3297     required a bigger vector, with some working space on the end. This means that
3298     the POSIX wrapper now has to get and free some memory, and copy the results.
3299    
3300 nigel 43 6. Added some simple autoconf support, placing the test data and the
3301     documentation in separate directories, re-organizing some of the
3302     information files, and making it build pcre-config (a GNU standard). Also added
3303     libtool support for building PCRE as a shared library, which is now the
3304     default.
3305 nigel 41
3306 nigel 43 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3307     09 are not valid octal constants. Single digits will be used for minor values
3308     less than 10.
3309    
3310     8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3311     existing programs that set these in the POSIX interface can use PCRE without
3312     modification.
3313    
3314     9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3315     return all that pcre_info() returns, plus additional data. The pcre_info()
3316     function is retained for compatibility, but is considered to be obsolete.
3317    
3318     10. Added experimental recursion feature (?R) to handle one common case that
3319     Perl 5.6 will be able to do with (?p{...}).
3320    
3321     11. Added support for POSIX character classes like [:alpha:], which Perl is
3322     adopting.
3323    
3324    
3325 nigel 39 Version 2.08 31-Aug-99
3326     ----------------------
3327    
3328     1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3329     trying to match at the startoffset position, but instead was moving forward to
3330     the next newline as if a previous match had failed.
3331    
3332     2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3333     and could get into a loop if a null string was matched other than at the start
3334     of the subject.
3335    
3336     3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3337     be distinguished at compile time, and for completeness also added PCRE_DATE.
3338    
3339     5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3340     in GnuWin32 environments.
3341    
3342    
3343 nigel 37 Version 2.07 29-Jul-99
3344     ----------------------
3345    
3346     1. The documentation is now supplied in plain text form and HTML as well as in
3347     the form of man page sources.
3348    
3349     2. C++ compilers don't like assigning (void *) values to other pointer types.
3350     In particular this affects malloc(). Although there is no problem in Standard
3351     C, I've put in casts to keep C++ compilers happy.
3352    
3353     3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3354     should be (const char *).
3355    
3356     4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3357     be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3358     However, I haven't made this a standard facility. The documentation doesn't
3359     mention it, and the Makefile doesn't support it.
3360    
3361     5. The Makefile now contains an "install" target, with editable destinations at
3362     the top of the file. The pcretest program is not installed.
3363    
3364     6. pgrep -V now gives the PCRE version number and date.
3365    
3366     7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3367     causing the entire string to be ignored, instead of just the last character.
3368    
3369     8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3370     non-matching string, it can take a very, very long time, even for strings of
3371     quite modest length, because of the nested recursion. PCRE now does better in
3372     some of these cases. It does this by remembering the last required literal
3373     character in the pattern, and pre-searching the subject to ensure it is present
3374     before running the real match. In other words, it applies a heuristic to detect
3375     some types of certain failure quickly, and in the above example, if presented
3376     with a string that has no trailing " it gives "no match" very quickly.
3377    
3378     9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3379     other alternatives are tried instead.
3380    
3381    
3382 nigel 35 Version 2.06 09-Jun-99
3383     ----------------------
3384    
3385     1. Change pcretest's output for amount of store used to show just the code
3386     space, because the remainder (the data block) varies in size between 32-bit and
3387     64-bit systems.
3388    
3389     2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3390     start matching at. This allows lookbehinds to work when searching for multiple
3391     occurrences in a string.
3392    
3393     3. Added additional options to pcretest for testing multiple occurrences:
3394    
3395     /+ outputs the rest of the string that follows a match
3396     /g loops for multiple occurrences, using the new startoffset argument
3397     /G loops for multiple occurrences by passing an incremented pointer
3398    
3399     4. PCRE wasn't doing the "first character" optimization for patterns starting
3400     with \b or \B, though it was doing it for other lookbehind assertions. That is,
3401     it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3402     the letter 'x'. On long subject strings, this gives a significant speed-up.
3403    
3404    
3405 nigel 33 Version 2.05 21-Apr-99
3406     ----------------------
3407    
3408     1. Changed the type of magic_number from int to long int so that it works
3409     properly on 16-bit systems.
3410    
3411     2. Fixed a bug which caused patterns starting with .* not to work correctly
3412     when the subject string contained newline characters. PCRE was assuming
3413     anchoring for such patterns in all cases, which is not correct because .* will
3414     not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3415     DOTALL is set at top level; otherwise it knows that patterns starting with .*
3416     must be retried after every newline in the subject.
3417    
3418    
3419 nigel 31 Version 2.04 18-Feb-99
3420     ----------------------
3421    
3422     1. For parenthesized subpatterns with repeats whose minimum was zero, the
3423     computation of the store needed to hold the pattern was incorrect (too large).
3424     If such patterns were nested a few deep, this could multiply and become a real
3425     problem.
3426    
3427     2. Added /M option to pcretest to show the memory requirement of a specific
3428     pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3429    
3430     3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3431     compiled in such a way that the backtracking after subsequent failure was
3432     pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3433     ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3434    
3435    
3436 nigel 29 Version 2.03 02-Feb-99
3437     ----------------------
3438    
3439     1. Fixed typo and small mistake in man page.
3440    
3441 nigel 31 2. Added 4th condition (GPL supersedes if conflict) and created separate
3442     LICENCE file containing the conditions.
3443 nigel 29
3444     3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3445     Perl, that is the internal \ allows the delimiter to be included in the
3446     pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3447     the final delimiter, add \ to the end of the pattern (to test the error).
3448    
3449     4. Added the convenience functions for extracting substrings after a successful
3450     match. Updated pcretest to make it able to test these functions.
3451    
3452    
3453 nigel 27 Version 2.02 14-Jan-99
3454     ----------------------
3455    
3456     1. Initialized the working variables associated with each extraction so that
3457     their saving and restoring doesn't refer to uninitialized store.
3458    
3459     2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3460     compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3461     fix the problem.
3462    
3463     3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3464     calls, and wasn't printing the correct value for compiling calls. Increased the
3465     default value of LOOPREPEAT, and the number of significant figures in the
3466     times.
3467    
3468     4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3469    
3470     5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3471     a building problem on Windows NT with a FAT file system.
3472    
3473    
3474 nigel 25 Version 2.01 21-Oct-98
3475     ----------------------
3476    
3477     1. Changed the API for pcre_compile() to allow for the provision of a pointer
3478     to character tables built by pcre_maketables() in the current locale. If NULL
3479     is passed, the default tables are used.
3480    
3481    
3482 nigel 23 Version 2.00 24-Sep-98
3483 nigel 21 ----------------------
3484    
3485 nigel 23 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3486     it any more.
3487    
3488     2. Allow quantification of (?>) groups, and make it work correctly.
3489    
3490     3. The first character computation wasn't working for (?>) groups.
3491    
3492     4. Correct the implementation of \Z (it is permitted to match on the \n at the
3493     end of the subject) and add 5.005's \z, which really does match only at the
3494     very end of the subject.
3495    
3496     5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3497    
3498     6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3499     DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3500     localized options. All options to pcre_study() were also removed.
3501    
3502     7. Add other new features from 5.005:
3503    
3504     $(?<= positive lookbehind
3505     $(?<! negative lookbehind
3506     (?imsx-imsx) added the unsetting capability
3507     such a setting is global if at outer level; local otherwise
3508     (?imsx-imsx:) non-capturing groups with option setting
3509     (?(cond)re|re) conditional pattern matching
3510    
3511     A backreference to itself in a repeated group matches the previous
3512     captured string.
3513    
3514     8. General tidying up of studying (both automatic and via "study")
3515     consequential on the addition of new assertions.
3516    
3517     9. As in 5.005, unlimited repeated groups that could match an empty substring
3518     are no longer faulted at compile time. Instead, the loop is forcibly broken at
3519     runtime if any iteration does actually match an empty substring.
3520    
3521     10. Include the RunTest script in the distribution.
3522    
3523     11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3524     discrepancies, some of which were old and were also with respect to 5.004. They
3525     have now been fixed.
3526    
3527    
3528     Version 1.09 28-Apr-98
3529     ----------------------
3530    
3531 nigel 21 1. A negated single character class followed by a quantifier with a minimum
3532     value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3533     program crashes, or just wrong answers. This did not apply to negated classes
3534     containing more than one character, or to minima other than one.
3535    
3536    
3537 nigel 19 Version 1.08 27-Mar-98
3538     ----------------------
3539    
3540     1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3541    
3542     2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3543     latter must appear before anything that relies on it in the pattern.
3544    
3545    
3546 nigel 17 Version 1.07 16-Feb-98
3547     ----------------------
3548    
3549     1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3550     repeat of a potentially empty string).
3551    
3552    
3553 nigel 15 Version 1.06 23-Jan-98
3554     ----------------------
3555    
3556     1. Added Markus Oberhumer's little patches for C++.
3557    
3558     2. Literal strings longer than 255 characters were broken.
3559    
3560    
3561 nigel 13 Version 1.05 23-Dec-97
3562     ----------------------
3563    
3564     1. Negated character classes containing more than one character were failing if
3565     PCRE_CASELESS was set at run time.
3566    
3567    
3568 nigel 11 Version 1.04 19-Dec-97
3569     ----------------------
3570    
3571     1. Corrected the man page, where some "const" qualifiers had been omitted.
3572    
3573     2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
3574     input syntax.
3575    
3576     3. Fixed memory leak which occurred when a regex with back references was
3577     matched with an offsets vector that wasn't big enough. The temporary memory
3578     that is used in this case wasn't being freed if the match failed.
3579    
3580     4. Tidied pcretest to ensure it frees memory that it gets.
3581    
3582     5. Temporary memory was being obtained in the case where the passed offsets
3583     vector was exactly big enough.
3584    
3585     6. Corrected definition of offsetof() from change 5 below.
3586    
3587     7. I had screwed up change 6 below and broken the rules for the use of
3588     setjmp(). Now fixed.
3589    
3590    
3591 nigel 9 Version 1.03 18-Dec-97
3592     ----------------------
3593    
3594     1. A erroneous regex with a missing opening parenthesis was correctly
3595     diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
3596     on some systems.
3597    
3598     2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
3599     it was reported that one broken compiler failed on the former because "code" is
3600     also an independent variable.
3601    
3602     3. The erroneous regex a[]b caused an array overrun reference.
3603    
3604     4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
3605     fail on data ending with that character. (It was going on too far, and checking
3606     the next character, typically a binary zero.) This was specific to the
3607     optimized code for single-character negative classes.
3608    
3609     5. Added a contributed patch from the TIN world which does the following:
3610    
3611     + Add an undef for memmove, in case the the system defines a macro for it.
3612    
3613     + Add a definition of offsetof(), in case there isn't one. (I don't know
3614     the reason behind this - offsetof() is part of the ANSI standard - but
3615     it does no harm).
3616    
3617     + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
3618     most of the places where whitespace preceded '#'. I have given up and
3619     allowed the remaining 2 cases to be at the margin.
3620    
3621     + Rename some variables in pcre to eliminate shadowing. This seems very
3622     pedantic, but does no harm, of course.
3623    
3624     6. Moved the call to setjmp() into its own function, to get rid of warnings
3625     from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
3626    
3627     7. Constructs such as \d{8,} were compiling into the equivalent of
3628 nigel 11 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
3629 nigel 9 outcome, but in this particular case used more store than had been allocated,
3630     which caused the bug to be discovered because it threw up an internal error.
3631    
3632     8. The debugging code in both pcre and pcretest for outputting the compiled
3633     form of a regex was going wrong in the case of back references followed by
3634     curly-bracketed repeats.
3635    
3636    
3637 nigel 7 Version 1.02 12-Dec-97
3638     ----------------------
3639    
3640     1. Typos in pcre.3 and comments in the source fixed.
3641    
3642     2. Applied a contributed patch to get rid of places where it used to remove
3643     'const' from variables, and fixed some signed/unsigned and uninitialized
3644     variable warnings.
3645    
3646     3. Added the "runtest" target to Makefile.
3647    
3648     4. Set default compiler flag to -O2 rather than just -O.
3649    
3650    
3651 nigel 5 Version 1.01 19-Nov-97
3652     ----------------------
3653    
3654     1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
3655     like /([ab]*)*/, that is, for classes with more than one character in them.
3656    
3657     2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
3658     as /((?>a*))*/ (a PCRE_EXTRA facility).
3659    
3660    
3661     Version 1.00 18-Nov-97
3662     ----------------------
3663    
3664     1. Added compile-time macros to support systems such as SunOS4 which don't have
3665     memmove() or strerror() but have other things that can be used instead.
3666    
3667     2. Arranged that "make clean" removes the executables.
3668    
3669    
3670 nigel 3 Version 0.99 27-Oct-97
3671     ----------------------
3672    
3673     1. Fixed bug in code for optimizing classes with only one character. It was
3674     initializing a 32-byte map regardless, which could cause it to run off the end
3675     of the memory it had got.
3676    
3677     2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
3678    
3679    
3680     Version 0.98 22-Oct-97
3681     ----------------------
3682    
3683     1. Fixed bug in code for handling temporary memory usage when there are more
3684     back references than supplied space in the ovector. This could cause segfaults.
3685    
3686    
3687     Version 0.97 21-Oct-97
3688     ----------------------
3689    
3690     1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
3691    
3692     2. Optimized negated single characters not to use a bit map.
3693    
3694     3. Brought error texts together as macro definitions; clarified some of them;
3695     fixed one that was wrong - it said "range out of order" when it meant "invalid
3696     escape sequence".
3697    
3698     4. Changed some char * arguments to const char *.
3699    
3700     5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
3701    
3702     6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
3703     pcretest.
3704    
3705    
3706     Version 0.96 16-Oct-97
3707     ----------------------
3708    
3709     1. Added a simple "pgrep" utility to the distribution.
3710    
3711     2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
3712     unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
3713     where "ddd" means "one or more decimal digits".
3714    
3715     3. Fixed serious bug. If a pattern had a back reference, but the call to
3716     pcre_exec() didn't supply a large enough ovector to record the related
3717     identifying subpattern, the match always failed. PCRE now remembers the number
3718     of the largest back reference, and gets some temporary memory in which to save
3719     the offsets during matching if necessary, in order to ensure that
3720     backreferences always work.
3721    
3722     4. Increased the compatibility with Perl in a number of ways:
3723    
3724     (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
3725     to request this handling. The option can be set at compile or exec time.
3726    
3727     (b) $ matches before a terminating newline by default; an option
3728     PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
3729     mode). The option can be set at compile or exec time.
3730    
3731     (c) The handling of \ followed by a digit other than 0 is now supposed to be
3732     the same as Perl's. If the decimal number it represents is less than 10
3733     or there aren't that many previous left capturing parentheses, an octal
3734     escape is read. Inside a character class, it's always an octal escape,
3735     even if it is a single digit.
3736    
3737     (d) An escaped but undefined alphabetic character is taken as a literal,
3738     unless PCRE_EXTRA is set. Currently this just reserves the remaining
3739     escapes.
3740    
3741     (e) {0} is now permitted. (The previous item is removed from the compiled
3742     pattern).
3743    
3744     5. Changed all the names of code files so that the basic parts are no longer
3745     than 10 characters, and abolished the teeny "globals.c" file.
3746    
3747     6. Changed the handling of character classes; they are now done with a 32-byte
3748     bit map always.
3749    
3750     7. Added the -d and /D options to pcretest to make it possible to look at the
3751     internals of compilation without having to recompile pcre.
3752    
3753    
3754     Version 0.95 23-Sep-97
3755     ----------------------
3756    
3757     1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
3758     \x20 at the start of a run of normal characters. These were being treated as
3759     real characters, instead of the source characters being re-checked.
3760    
3761    
3762     Version 0.94 18-Sep-97
3763     ----------------------
3764    
3765     1. The functions are now thread-safe, with the caveat that the global variables
3766     containing pointers to malloc() and free() or alternative functions are the
3767     same for all threads.
3768    
3769     2. Get pcre_study() to generate a bitmap of initial characters for non-
3770     anchored patterns when this is possible, and use it if passed to pcre_exec().
3771    
3772    
3773     Version 0.93 15-Sep-97
3774     ----------------------
3775    
3776     1. /(b)|(:+)/ was computing an incorrect first character.
3777    
3778     2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
3779     but not actually doing anything yet.
3780    
3781     3. Treat "-" characters in classes that cannot be part of ranges as literals,
3782     as Perl does (e.g. [-az] or [az-]).
3783    
3784     4. Set the anchored flag if a branch starts with .* or .*? because that tests
3785     all possible positions.
3786    
3787     5. Split up into different modules to avoid including unneeded functions in a
3788     compiled binary. However, compile and exec are still in one module. The "study"
3789     function is split off.
3790    
3791     6. The character tables are now in a separate module whose source is generated
3792     by an auxiliary program - but can then be edited by hand if required. There are
3793     now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
3794     toupper() in the code.
3795    
3796     7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
3797     make them global. Abolish the function for setting them, as the caller can now
3798     set them directly.
3799    
3800    
3801     Version 0.92 11-Sep-97
3802     ----------------------
3803    
3804     1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
3805     (e.g. /a{1,3}/) was broken (I mis-optimized it).
3806    
3807     2. Caseless matching was not working in character classes if the characters in
3808     the pattern were in upper case.
3809    
3810     3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
3811    
3812     4. Make PCRE_ANCHORED public and accept as a compile option.
3813    
3814     5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
3815     PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
3816     pass them.
3817    
3818     6. Give an error if bad option bits passed at compile or run time.
3819