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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 505 - (hide annotations) (download)
Tue Mar 9 16:50:47 2010 UTC (4 years, 5 months ago) by ph10
File size: 184737 byte(s)
Improve compile-time overrun checking.

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