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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 496 - (hide annotations) (download)
Tue Mar 2 19:11:17 2010 UTC (4 years, 6 months ago) by ph10
File size: 182334 byte(s)
Fix DEFINE bug for forward reference with a possessive quantifier.

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