ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 551 - (hide annotations) (download)
Sun Oct 10 17:33:07 2010 UTC (4 years, 6 months ago) by ph10
File size: 190344 byte(s)
Make (*COMMIT) override (*THEN) and similar.

1 nigel 3 ChangeLog for PCRE
2     ------------------
4 ph10 550 Version 8.11 10-Oct-2010
5     ------------------------
7     1. (*THEN) was not working properly if there were untried alternatives prior
8     to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
9     backtracked to try for "b" instead of moving to the next alternative branch
10     at the same level (in this case, to look for "c"). The Perl documentation
11     is clear that when (*THEN) is backtracked onto, it goes to the "next
12     alternative in the innermost enclosing group".
13 ph10 551
14     2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
15     such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
16     result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
17     (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
18     (*THEN).
19 ph10 550
21 ph10 548 Version 8.10 25-Jun-2010
22 ph10 510 ------------------------
24 ph10 512 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
25 ph10 510 THEN.
27 ph10 512 2. (*ACCEPT) was not working when inside an atomic group.
28 ph10 510
29 ph10 535 3. Inside a character class, \B is treated as a literal by default, but
30     faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
31 ph10 513 causes the error). The code is unchanged, but I tidied the documentation.
32 ph10 535
33     4. Inside a character class, PCRE always treated \R and \X as literals,
34 ph10 513 whereas Perl faults them if its -w option is set. I have changed PCRE so
35     that it faults them when PCRE_EXTRA is set.
36 ph10 535
37 ph10 514 5. Added support for \N, which always matches any character other than
38     newline. (It is the same as "." when PCRE_DOTALL is not set.)
39 ph10 535
40 ph10 515 6. When compiling pcregrep with newer versions of gcc which may have
41 ph10 535 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
42 ph10 515 declared with attribute warn_unused_result" were given. Just casting the
43 ph10 535 result to (void) does not stop the warnings; a more elaborate fudge is
44     needed. I've used a macro to implement this.
45 ph10 512
46 ph10 535 7. Minor change to pcretest.c to avoid a compiler warning.
48 ph10 517 8. Added four artifical Unicode properties to help with an option to make
49 ph10 518 \s etc use properties (see next item). The new properties are: Xan
50     (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
51 ph10 535
52 ph10 518 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
53 ph10 535 use Unicode properties. (*UCP) at the start of a pattern can be used to set
54 ph10 518 this option. Modified pcretest to add /W to test this facility. Added
55     REG_UCP to make it available via the POSIX interface.
56 ph10 513
57 ph10 535 10. Added --line-buffered to pcregrep.
59     11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
60     studied, and the match started with a letter with a code point greater than
61 ph10 520 127 whose first byte was different to the first byte of the other case of
62 ph10 545 the letter, the other case of this starting letter was not recognized
63 ph10 536 (#976).
64 ph10 535
65 ph10 536 12. If a pattern that was studied started with a repeated Unicode property
66 ph10 523 test, for example, \p{Nd}+, there was the theoretical possibility of
67 ph10 535 setting up an incorrect bitmap of starting bytes, but fortunately it could
68     not have actually happened in practice until change 8 above was made (it
69 ph10 523 added property types that matched character-matching opcodes).
70 ph10 535
71 ph10 536 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
72 ph10 535 possible starting bytes for non-anchored patterns.
74 ph10 536 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
75 ph10 535 \R, and also a number of cases that involve Unicode properties, both
76 ph10 532 explicit and implicit when PCRE_UCP is set.
77 ph10 529
78 ph10 536 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
79 ph10 535 input, it could crash or give wrong results if characters with values
80     greater than 0xc0 were present in the subject string. (Detail: it assumed
81 ph10 529 UTF-8 input when processing these items.)
82 ph10 535
83 ph10 536 16. Added a lot of (int) casts to avoid compiler warnings in systems where
84     size_t is 64-bit (#991).
85 ph10 535
86 ph10 536 17. Added a check for running out of memory when PCRE is compiled with
87     --disable-stack-for-recursion (#990).
88 ph10 545
89 ph10 537 18. If the last data line in a file for pcretest does not have a newline on
90 ph10 545 the end, a newline was missing in the output.
92     19. The default pcre_chartables.c file recognizes only ASCII characters (values
93     less than 128) in its various bitmaps. However, there is a facility for
94     generating tables according to the current locale when PCRE is compiled. It
95     turns out that in some environments, 0x85 and 0xa0, which are Unicode space
96     characters, are recognized by isspace() and therefore were getting set in
97 ph10 541 these tables, and indeed these tables seem to approximate to ISO 8859. This
98     caused a problem in UTF-8 mode when pcre_study() was used to create a list
99     of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
100     which of course cannot start UTF-8 characters. I have changed the code so
101     that only real ASCII characters (less than 128) and the correct starting
102 ph10 545 bytes for UTF-8 encodings are set for characters greater than 127 when in
103 ph10 541 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
104     altogether.)
105 ph10 545
106 ph10 541 20. Added the /T option to pcretest so as to be able to run tests with non-
107     standard character tables, thus making it possible to include the tests
108 ph10 545 used for 19 above in the standard set of tests.
110     21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
111     reference to a subpattern the other side of a comment that contains an
112     opening parenthesis caused either an internal compiling error, or a
113 ph10 544 reference to the wrong subpattern.
114 ph10 517
115 ph10 520
116 ph10 508 Version 8.02 19-Mar-2010
117 ph10 491 ------------------------
119     1. The Unicode data tables have been updated to Unicode 5.2.0.
121 ph10 492 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
122     configured.
123 ph10 507
124 ph10 493 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
125 ph10 507 original author of that file, following a query about its status.
127     4. On systems that do not have stdint.h (e.g. Solaris), check for and include
128 ph10 494 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
129 ph10 507
130 ph10 496 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
131     quantifier applied to a forward-referencing subroutine call, could compile
132     incorrect code or give the error "internal error: previously-checked
133     referenced subpattern not found".
134 ph10 507
135     6. Both MS Visual Studio and Symbian OS have problems with initializing
136 ph10 497 variables to point to external functions. For these systems, therefore,
137     pcre_malloc etc. are now initialized to local functions that call the
138     relevant global functions.
139 ph10 507
140 ph10 498 7. There were two entries missing in the vectors called coptable and poptable
141     in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
142 ph10 507 I've fixed the data, and added a kludgy way of testing at compile time that
143     the lengths are correct (equal to the number of opcodes).
145     8. Following on from 7, I added a similar kludge to check the length of the
146     eint vector in pcreposix.c.
148     9. Error texts for pcre_compile() are held as one long string to avoid too
149     much relocation at load time. To find a text, the string is searched,
150 ph10 499 counting zeros. There was no check for running off the end of the string,
151     which could happen if a new error number was added without updating the
152 ph10 507 string.
154     10. \K gave a compile-time error if it appeared in a lookbehind assersion.
156 ph10 500 11. \K was not working if it appeared in an atomic group or in a group that
157 ph10 507 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
158     \K is "not well defined" if used in an assertion. PCRE now accepts it if
159     the assertion is positive, but not if it is negative.
161 ph10 501 12. Change 11 fortuitously reduced the size of the stack frame used in the
162 ph10 507 "match()" function of pcre_exec.c by one pointer. Forthcoming
163     implementation of support for (*MARK) will need an extra pointer on the
164 ph10 501 stack; I have reserved it now, so that the stack frame size does not
165     decrease.
166 ph10 507
167     13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
168     item in branch that calls a recursion is a subroutine call - as in the
169 ph10 503 second branch in the above example - was incorrectly given the compile-
170     time error "recursive call could loop indefinitely" because pcre_compile()
171 ph10 507 was not correctly checking the subroutine for matching a non-empty string.
173 ph10 505 14. The checks for overrunning compiling workspace could trigger after an
174 ph10 507 overrun had occurred. This is a "should never occur" error, but it can be
175 ph10 505 triggered by pathological patterns such as hundreds of nested parentheses.
176 ph10 507 The checks now trigger 100 bytes before the end of the workspace.
177 ph10 491
178 ph10 507 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
179 ph10 492
180 ph10 507
181 ph10 489 Version 8.01 19-Jan-2010
182     ------------------------
183 ph10 471
184     1. If a pattern contained a conditional subpattern with only one branch (in
185 ph10 489 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
186 ph10 473 computed the wrong minimum data length (which is of course zero for such
187 ph10 489 subpatterns). This could cause incorrect "no match" results.
188 ph10 473
189     2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
190 ph10 472 the pattern is reset in the first branch, pcre_compile() failed with
191     "internal error: code overflow at offset...". This happened only when
192 ph10 473 the reset was to the original external option setting. (An optimization
193     abstracts leading options settings into an external setting, which was the
194 ph10 472 cause of this.)
195 ph10 471
196 ph10 473 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
197     of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
198     assertion pattern did not match (meaning that the assertion was true), it
199     was incorrectly treated as false if the SKIP had been reached during the
200     matching. This also applied to assertions used as conditions.
201 ph10 471
202 ph10 473 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
203     assertion subpattern, including such a pattern used as a condition,
204     unpredictable results occurred, instead of the error return
206 ph10 478
207 ph10 474 5. The C++ GlobalReplace function was not working like Perl for the special
208     situation when an empty string is matched. It now does the fancy magic
209 ph10 478 stuff that is necessary.
211     6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
212 ph10 475 removed. (These were left over from very, very early versions of PCRE.)
213 ph10 478
214 ph10 475 7. Some cosmetic changes to the code to make life easier when compiling it
215     as part of something else:
216 ph10 478
217     (a) Change DEBUG to PCRE_DEBUG.
219     (b) In pcre_compile(), rename the member of the "branch_chain" structure
220     called "current" as "current_branch", to prevent a collision with the
221 ph10 475 Linux macro when compiled as a kernel module.
222 ph10 478
223     (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
224     prevent a collision with the Linux macro when compiled as a kernel
225 ph10 475 module.
226 ph10 478
227 ph10 475 8. In pcre_compile() there are some checks for integer overflows that used to
228     cast potentially large values to (double). This has been changed to that
229 ph10 478 when building, a check for int64_t is made, and if it is found, it is used
230     instead, thus avoiding the use of floating point arithmetic. (There is no
231     other use of FP in PCRE.) If int64_t is not found, the fallback is to
232     double.
234     9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
235 ph10 476 2005 (difference between two addresses compared to an unsigned value).
236 ph10 473
237 ph10 478 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
238     custom one, because of the following reported problem in Windows:
239 ph10 473
240 ph10 478 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
241 ph10 489 under Win32.
242 ph10 478 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
243 ph10 489 therefore missing the function definition.
244 ph10 478 - The compiler thus generates a "C" signature for the test function.
245     - The linker fails to find the "C" function.
246     - PCRE fails to configure if asked to do so against libbz2.
247 ph10 486
248 ph10 479 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
249     messages were output:
250 ph10 478
251 ph10 479 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
252     rerunning libtoolize, to keep the correct libtool macros in-tree.
253     Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
254 ph10 486
255 ph10 479 I have done both of these things.
256 ph10 486
257 ph10 480 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
258 ph10 486 most of the time, it *can* run out if it is given a pattern that contains a
259     runaway infinite recursion. I updated the discussion in the pcrestack man
260     page.
262     13. Now that we have gone to the x.xx style of version numbers, the minor
263     version may start with zero. Using 08 or 09 is a bad idea because users
264     might check the value of PCRE_MINOR in their code, and 08 or 09 may be
265     interpreted as invalid octal numbers. I've updated the previous comment in
266     configure.ac, and also added a check that gives an error if 08 or 09 are
267 ph10 481 used.
268 ph10 486
269 ph10 482 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
270 ph10 483 causing partial matching to fail when the end of the subject matched \W
271     in a UTF-8 pattern where \W was quantified with a minimum of 3.
272 ph10 486
273 ph10 483 15. There were some discrepancies between the declarations in pcre_internal.h
274 ph10 486 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
275 ph10 483 their definitions. The declarations used "const uschar *" and the
276     definitions used USPTR. Even though USPTR is normally defined as "const
277     unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
278     reported that: "This difference in casting confuses some C++ compilers, for
279     example, SunCC recognizes above declarations as different functions and
280 ph10 486 generates broken code for hbpcre." I have changed the declarations to use
281 ph10 483 USPTR.
282 ph10 486
283     16. GNU libtool is named differently on some systems. The autogen.sh script now
284 ph10 484 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
285     (FreeBSD).
286 ph10 478
287 ph10 486 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
288     (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
289     comment: "Figure out how to create a longlong from a string: strtoll and
290     equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
291     instance, but it only takes 2 args instead of 3!"
292 ph10 479
293 ph10 488 18. A subtle bug concerned with back references has been fixed by a change of
294     specification, with a corresponding code fix. A pattern such as
295     ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
296     refers, was giving matches when it shouldn't. For example, xa=xaaa would
297     match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
298     same bug. Such groups have to be quantified to be useful, or contained
299     inside another quantified group. (If there's no repetition, the reference
300     can never match.) The problem arises because, having left the group and
301     moved on to the rest of the pattern, a later failure that backtracks into
302     the group uses the captured value from the final iteration of the group
303     rather than the correct earlier one. I have fixed this in PCRE by forcing
304     any group that contains a reference to itself to be an atomic group; that
305     is, there cannot be any backtracking into it once it has completed. This is
306     similar to recursive and subroutine calls.
307 ph10 486
308 ph10 488
309 ph10 469 Version 8.00 19-Oct-09
310 ph10 418 ----------------------
312     1. The table for translating pcre_compile() error codes into POSIX error codes
313 ph10 461 was out-of-date, and there was no check on the pcre_compile() error code
314     being within the table. This could lead to an OK return being given in
315 ph10 418 error.
316 ph10 461
317     2. Changed the call to open a subject file in pcregrep from fopen(pathname,
318     "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
319     in a Windows environment.
321 ph10 420 3. The pcregrep --count option prints the count for each file even when it is
322     zero, as does GNU grep. However, pcregrep was also printing all files when
323     --files-with-matches was added. Now, when both options are given, it prints
324     counts only for those files that have at least one match. (GNU grep just
325 ph10 461 prints the file name in this circumstance, but including the count seems
326     more useful - otherwise, why use --count?) Also ensured that the
327 ph10 420 combination -clh just lists non-zero counts, with no names.
328 ph10 461
329     4. The long form of the pcregrep -F option was incorrectly implemented as
330     --fixed_strings instead of --fixed-strings. This is an incompatible change,
331     but it seems right to fix it, and I didn't think it was worth preserving
332     the old behaviour.
334     5. The command line items --regex=pattern and --regexp=pattern were not
335 ph10 422 recognized by pcregrep, which required --regex pattern or --regexp pattern
336 ph10 461 (with a space rather than an '='). The man page documented the '=' forms,
337 ph10 422 which are compatible with GNU grep; these now work.
338 ph10 461
339     6. No libpcreposix.pc file was created for pkg-config; there was just
340 ph10 423 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
341 ph10 461
342 ph10 425 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
343 ph10 461 when UCP support is not needed, by modifying the Python script that
344 ph10 425 generates it from Unicode data files. This should not matter if the module
345     is correctly used as a library, but I received one complaint about 50K of
346     unwanted data. My guess is that the person linked everything into his
347     program rather than using a library. Anyway, it does no harm.
348 ph10 461
349 ph10 426 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
350 ph10 461 was a minimum greater than 1 for a wide character in a possessive
351 ph10 431 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
352     which had an unlimited repeat of a nested, fixed maximum repeat of a wide
353     character. Chaos in the form of incorrect output or a compiling loop could
354     result.
355 ph10 461
356 ph10 426 9. The restrictions on what a pattern can contain when partial matching is
357 ph10 461 requested for pcre_exec() have been removed. All patterns can now be
358 ph10 426 partially matched by this function. In addition, if there are at least two
359 ph10 435 slots in the offset vector, the offset of the earliest inspected character
360     for the match and the offset of the end of the subject are set in them when
361 ph10 461 PCRE_ERROR_PARTIAL is returned.
363 ph10 428 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
364     synonymous with PCRE_PARTIAL, for backwards compatibility, and
365     PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
366 ph10 462 and may be more useful for multi-segment matching.
367 ph10 461
368     11. Partial matching with pcre_exec() is now more intuitive. A partial match
369     used to be given if ever the end of the subject was reached; now it is
370     given only if matching could not proceed because another character was
371     needed. This makes a difference in some odd cases such as Z(*FAIL) with the
372     string "Z", which now yields "no match" instead of "partial match". In the
373     case of pcre_dfa_exec(), "no match" is given if every matching path for the
374     final character ended with (*FAIL).
376 ph10 428 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
377 ph10 461 if the pattern had a "must contain" character that was already found in the
378 ph10 428 earlier partial match, unless partial matching was again requested. For
379     example, with the pattern /dog.(body)?/, the "must contain" character is
380     "g". If the first part-match was for the string "dog", restarting with
381 ph10 435 "sbody" failed. This bug has been fixed.
382 ph10 461
383     13. The string returned by pcre_dfa_exec() after a partial match has been
384     changed so that it starts at the first inspected character rather than the
385     first character of the match. This makes a difference only if the pattern
386     starts with a lookbehind assertion or \b or \B (\K is not supported by
387     pcre_dfa_exec()). It's an incompatible change, but it makes the two
388 ph10 435 matching functions compatible, and I think it's the right thing to do.
389 ph10 461
390 ph10 435 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
391 ph10 461 so that the demonstration program is easily available in environments where
392     PCRE has not been installed from source.
394 ph10 435 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
395 ph10 430 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
396     library.
397 ph10 461
398 ph10 435 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
399 ph10 432 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
400 ph10 461 is not the first non-POSIX option to be added. Clearly some people find
401 ph10 433 these options useful.
402 ph10 461
403     17. If a caller to the POSIX matching function regexec() passes a non-zero
404 ph10 438 value for nmatch with a NULL value for pmatch, the value of
405 ph10 461 nmatch is forced to zero.
407 ph10 437 18. RunGrepTest did not have a test for the availability of the -u option of
408 ph10 461 the diff command, as RunTest does. It now checks in the same way as
409 ph10 437 RunTest, and also checks for the -b option.
410 ph10 461
411 ph10 438 19. If an odd number of negated classes containing just a single character
412     interposed, within parentheses, between a forward reference to a named
413 ph10 461 subpattern and the definition of the subpattern, compilation crashed with
414     an internal error, complaining that it could not find the referenced
415 ph10 438 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
416 ph10 461 [The bug was that it was starting one character too far in when skipping
417     over the character class, thus treating the ] as data rather than
418     terminating the class. This meant it could skip too much.]
420 ph10 442 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
421 ph10 461 /g option in pcretest when the pattern contains \K, which makes it possible
422 ph10 442 to have an empty string match not at the start, even when the pattern is
423 ph10 461 anchored. Updated pcretest and pcredemo to use this option.
425 ph10 446 21. If the maximum number of capturing subpatterns in a recursion was greater
426 ph10 461 than the maximum at the outer level, the higher number was returned, but
427     with unset values at the outer level. The correct (outer level) value is
428 ph10 446 now given.
429 ph10 461
430 ph10 447 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
431     PCRE did not set those parentheses (unlike Perl). I have now found a way to
432     make it do so. The string so far is captured, making this feature
433     compatible with Perl.
434 ph10 461
435     23. The tests have been re-organized, adding tests 11 and 12, to make it
436 ph10 448 possible to check the Perl 5.10 features against Perl 5.10.
437 ph10 461
438 ph10 454 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
439 ph10 461 pattern matches a fixed length string. PCRE did not allow this; now it
440     does. Neither allows recursion.
442     25. I finally figured out how to implement a request to provide the minimum
443     length of subject string that was needed in order to match a given pattern.
444     (It was back references and recursion that I had previously got hung up
445     on.) This code has now been added to pcre_study(); it finds a lower bound
446 ph10 455 to the length of subject needed. It is not necessarily the greatest lower
447     bound, but using it to avoid searching strings that are too short does give
448     some useful speed-ups. The value is available to calling programs via
449     pcre_fullinfo().
450 ph10 461
451 ph10 455 26. While implementing 25, I discovered to my embarrassment that pcretest had
452     not been passing the result of pcre_study() to pcre_dfa_exec(), so the
453     study optimizations had never been tested with that matching function.
454     Oops. What is worse, even when it was passed study data, there was a bug in
455     pcre_dfa_exec() that meant it never actually used it. Double oops. There
456     were also very few tests of studied patterns with pcre_dfa_exec().
457 ph10 461
458 ph10 457 27. If (?| is used to create subpatterns with duplicate numbers, they are now
459     allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
460     on the other side of the coin, they are no longer allowed to have different
461     names, because these cannot be distinguished in PCRE, and this has caused
462     confusion. (This is a difference from Perl.)
463 ph10 461
464     28. When duplicate subpattern names are present (necessarily with different
465     numbers, as required by 27 above), and a test is made by name in a
466     conditional pattern, either for a subpattern having been matched, or for
467     recursion in such a pattern, all the associated numbered subpatterns are
468 ph10 459 tested, and the overall condition is true if the condition is true for any
469     one of them. This is the way Perl works, and is also more like the way
470     testing by number works.
471 ph10 418
472 ph10 461
473 ph10 412 Version 7.9 11-Apr-09
474 ph10 376 ---------------------
476 ph10 392 1. When building with support for bzlib/zlib (pcregrep) and/or readline
477 ph10 376 (pcretest), all targets were linked against these libraries. This included
478     libpcre, libpcreposix, and libpcrecpp, even though they do not use these
479     libraries. This caused unwanted dependencies to be created. This problem
480 ph10 392 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
481 ph10 376 pcretest is linked with readline.
482 ph10 392
483 ph10 376 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
484     "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
485     moved outside it again, because FALSE and TRUE are already defined in AIX,
486     but BOOL is not.
487 ph10 392
488     3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
489 ph10 409 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
490 ph10 392
491 ph10 376 4. The pcregrep documentation said spaces were inserted as well as colons (or
492     hyphens) following file names and line numbers when outputting matching
493 ph10 392 lines. This is not true; no spaces are inserted. I have also clarified the
494 ph10 376 wording for the --colour (or --color) option.
495 ph10 392
496 ph10 378 5. In pcregrep, when --colour was used with -o, the list of matching strings
497     was not coloured; this is different to GNU grep, so I have changed it to be
498     the same.
499 ph10 392
500 ph10 378 6. When --colo(u)r was used in pcregrep, only the first matching substring in
501 ph10 392 each matching line was coloured. Now it goes on to look for further matches
502     of any of the test patterns, which is the same behaviour as GNU grep.
504     7. A pattern that could match an empty string could cause pcregrep to loop; it
505     doesn't make sense to accept an empty string match in pcregrep, so I have
506 ph10 379 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
507     seems to be how GNU grep behaves.
508 ph10 392
509 ph10 380 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
510 ph10 392 start or after a newline", because the conditional assertion was not being
511     correctly handled. The rule now is that both the assertion and what follows
512     in the first alternative must satisfy the test.
514 ph10 405 9. If auto-callout was enabled in a pattern with a conditional group whose
515 ph10 399 condition was an assertion, PCRE could crash during matching, both with
516     pcre_exec() and pcre_dfa_exec().
517 ph10 392
518     10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
519     used for matching.
521     11. Unicode property support in character classes was not working for
522 ph10 385 characters (bytes) greater than 127 when not in UTF-8 mode.
523 ph10 388
524 ph10 392 12. Added the -M command line option to pcretest.
526 ph10 388 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
527 ph10 389
528     15. Added the PCRE_NO_START_OPTIMIZE match-time option.
529 ph10 390
530 ph10 392 16. Added comments and documentation about mis-use of no_arg in the C++
531 ph10 390 wrapper.
532 ph10 392
533 ph10 391 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
534 ph10 392 from Martin Jerabek that uses macro names for all relevant character and
535 ph10 391 string constants.
536 ph10 405
537 ph10 393 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
538     SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
539 ph10 405 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
540     these, but not everybody uses configure.
542     19. A conditional group that had only one branch was not being correctly
543     recognized as an item that could match an empty string. This meant that an
544     enclosing group might also not be so recognized, causing infinite looping
545     (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
546     with the subject "ab", where knowledge that the repeated group can match
547 ph10 396 nothing is needed in order to break the loop.
548 ph10 405
549 ph10 397 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
550 ph10 405 exec(), but without supplying a callout function, matching went wrong.
552     21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
553     leak if the size of the offset vector was greater than 30. When the vector
554     is smaller, the saved offsets during recursion go onto a local stack
555     vector, but for larger vectors malloc() is used. It was failing to free
556     when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
557 ph10 400 error, in fact).
558 ph10 405
559 ph10 402 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
560 ph10 405 heapframe that is used only when UTF-8 support is enabled. This caused no
561     problem, but was untidy.
562 ph10 376
563 ph10 405 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
564     CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
565     included within another project.
566 ph10 411
567 ph10 407 24. Steven Van Ingelgem's patches to add more options to the CMake support,
568     slightly modified by me:
569 ph10 411
570     (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
571     not building pcregrep.
573 ph10 407 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
574 ph10 411 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
576 ph10 408 25. Forward references, both numeric and by name, in patterns that made use of
577     duplicate group numbers, could behave incorrectly or give incorrect errors,
578     because when scanning forward to find the reference group, PCRE was not
579     taking into account the duplicate group numbers. A pattern such as
580     ^X(?3)(a)(?|(b)|(q))(Y) is an example.
581 ph10 411
582     26. Changed a few more instances of "const unsigned char *" to USPTR, making
583     the feature of a custom pointer more persuasive (as requested by a user).
584 ph10 416
585 ph10 411 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
586 ph10 416 pcretest.c, inside #ifndefs, because it seems they are sometimes already
587     pre-defined.
588 ph10 392
589 ph10 416 28. Added support for (*UTF8) at the start of a pattern.
591 ph10 413 29. Arrange for flags added by the "release type" setting in CMake to be shown
592     in the configuration summary.
593 ph10 405
594 ph10 411
595 ph10 374 Version 7.8 05-Sep-08
596 ph10 349 ---------------------
598     1. Replaced UCP searching code with optimized version as implemented for Ad
599     Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
600 ph10 351 stage table and inline lookup instead of a function, giving speed ups of 2
601     to 5 times on some simple patterns that I tested. Permission was given to
602     distribute the MultiStage2.py script that generates the tables (it's not in
603     the tarball, but is in the Subversion repository).
604 ph10 349
605 ph10 351 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
606     scripts.
607 ph10 358
608 ph10 353 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
609     a group with a zero qualifier. The result of the study could be incorrect,
610 ph10 358 or the function might crash, depending on the pattern.
612     4. Caseless matching was not working for non-ASCII characters in back
613 ph10 354 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
614 ph10 358 It now works when Unicode Property Support is available.
616 ph10 355 5. In pcretest, an escape such as \x{de} in the data was always generating
617     a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
618     non-UTF-8 mode. If the value is greater than 255, it gives a warning about
619 ph10 358 truncation.
620 ph10 349
621 ph10 358 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
623 ph10 357 7. Added two (int) casts to pcregrep when printing the difference of two
624     pointers, in case they are 64-bit values.
625 ph10 371
626     8. Added comments about Mac OS X stack usage to the pcrestack man page and to
627     test 2 if it fails.
629 ph10 359 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
630 ph10 371 and a #define of that name to empty if it is not externally set. This is to
631     allow users of MSVC to set it if necessary.
633     10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
634 ph10 359 the convenience functions in the pcre_get.c source file.
635 ph10 371
636 ph10 360 11. An option change at the start of a pattern that had top-level alternatives
637     could cause overwriting and/or a crash. This command provoked a crash in
638 ph10 371 some environments:
640     printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
642 ph10 360 This potential security problem was recorded as CVE-2008-2371.
643 ph10 371
644 ph10 361 12. For a pattern where the match had to start at the beginning or immediately
645     after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
646 ph10 371 pcre_dfa_exec() could read past the end of the passed subject if there was
647 ph10 361 no match. To help with detecting such bugs (e.g. with valgrind), I modified
648     pcretest so that it places the subject at the end of its malloc-ed buffer.
649 ph10 371
650 ph10 364 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
651 ph10 371 exec() might read past the end of the data buffer in UTF-8 mode.
653 ph10 364 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
654 ph10 371 the data contained the byte 0x85 as part of a UTF-8 character within its
655     first line. This applied both to normal and DFA matching.
657 ph10 366 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
658 ph10 371 /^[^d]*?$/8 failed to match "abc".
659 ph10 351
660 ph10 371 16. Added a missing copyright notice to pcrecpp_internal.h.
662     17. Make it more clear in the documentation that values returned from
663 ph10 368 pcre_exec() in ovector are byte offsets, not character counts.
664 ph10 357
665 ph10 371 18. Tidied a few places to stop certain compilers from issuing warnings.
666 ph10 368
667 ph10 373 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
668     supplied by Stefan Weber. I made a further small update for 7.8 because
669 ph10 374 there is a change of source arrangements: the pcre_searchfuncs.c module is
670     replaced by pcre_ucd.c.
671 ph10 371
672 ph10 373
673 ph10 347 Version 7.7 07-May-08
674 ph10 321 ---------------------
676     1. Applied Craig's patch to sort out a long long problem: "If we can't convert
677 ph10 345 a string to a long long, pretend we don't even have a long long." This is
678 ph10 321 done by checking for the strtoq, strtoll, and _strtoi64 functions.
679 ph10 345
680 ph10 322 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
681     pre-7.6 versions, which defined a global no_arg variable instead of putting
682 ph10 332 it in the RE class. (See also #8 below.)
683 ph10 345
684     3. Remove a line of dead code, identified by coverity and reported by Nuno
685     Lopes.
687 ph10 324 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
688 ph10 321
689 ph10 324 (1) The include/exclude patterns were being applied to the whole pathnames
690 ph10 345 of files, instead of just to the final components.
692 ph10 324 (2) If there was more than one level of directory, the subdirectories were
693     skipped unless they satisfied the include/exclude conditions. This is
694     inconsistent with GNU grep (and could even be seen as contrary to the
695     pcregrep specification - which I improved to make it absolutely clear).
696     The action now is always to scan all levels of directory, and just
697     apply the include/exclude patterns to regular files.
698 ph10 345
699 ph10 325 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
700 ph10 345 --exclude_dir in the tests to avoid scanning .svn directories.
702 ph10 326 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
703 ph10 345 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
704 ph10 326 doesn't support NULs in patterns.
705 ph10 345
706     7. Added some missing "const"s to declarations of static tables in
707     pcre_compile.c and pcre_dfa_exec.c.
709 ph10 328 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
710 ph10 329 caused by fix #2 above. (Subsequently also a second patch to fix the
711 ph10 345 first patch. And a third patch - this was a messy problem.)
712 ph10 321
713 ph10 345 9. Applied Craig's patch to remove the use of push_back().
715 ph10 332 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
716     matching function regexec().
717 ph10 345
718     11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
719     which, however, unlike Perl's \g{...}, are subroutine calls, not back
720 ph10 333 references. PCRE supports relative numbers with this syntax (I don't think
721     Oniguruma does).
722 ph10 345
723     12. Previously, a group with a zero repeat such as (...){0} was completely
724 ph10 335 omitted from the compiled regex. However, this means that if the group
725     was called as a subroutine from elsewhere in the pattern, things went wrong
726 ph10 345 (an internal error was given). Such groups are now left in the compiled
727     pattern, with a new opcode that causes them to be skipped at execution
728 ph10 335 time.
729 ph10 345
730 ph10 341 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
731     to the way PCRE behaves:
732 ph10 345
733 ph10 336 (a) A lone ] character is dis-allowed (Perl treats it as data).
734 ph10 345
735     (b) A back reference to an unmatched subpattern matches an empty string
736 ph10 336 (Perl fails the current match path).
737 ph10 345
738 ph10 341 (c) A data ] in a character class must be notated as \] because if the
739 ph10 345 first data character in a class is ], it defines an empty class. (In
740 ph10 341 Perl it is not possible to have an empty class.) The empty class []
741     never matches; it forces failure and is equivalent to (*FAIL) or (?!).
742     The negative empty class [^] matches any one character, independently
743     of the DOTALL setting.
744 ph10 345
745     14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
746 ph10 340 non-existent subpattern following a character class starting with ']' and
747     containing () gave an internal compiling error instead of "reference to
748     non-existent subpattern". Fortunately, when the pattern did exist, the
749 ph10 345 compiled code was correct. (When scanning forwards to check for the
750 ph10 340 existencd of the subpattern, it was treating the data ']' as terminating
751 ph10 345 the class, so got the count wrong. When actually compiling, the reference
752 ph10 340 was subsequently set up correctly.)
753 ph10 345
754 ph10 341 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
755 ph10 345 it was being rejected as not supported by pcre_dfa_exec(), even though
756     other assertions are supported. I have made pcre_dfa_exec() support
757     (*FAIL).
759 ph10 342 16. The implementation of 13c above involved the invention of a new opcode,
760 ph10 345 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
761     cannot be changed at match time, I realized I could make a small
762     improvement to matching performance by compiling OP_ALLANY instead of
763     OP_ANY for "." when DOTALL was set, and then removing the runtime tests
764     on the OP_ANY path.
766     17. Compiling pcretest on Windows with readline support failed without the
767 ph10 343 following two fixes: (1) Make the unistd.h include conditional on
768 ph10 345 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
770 ph10 344 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
771 ph10 345 ncurses library to be included for pcretest when ReadLine support is
772     requested, but also to allow for it to be overridden. This patch came from
773     Daniel Bergström.
774 ph10 347
775 ph10 346 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
776 ph10 347 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
777     any errors with the current Unicode tables. Thanks to Peter Kankowski for
778     spotting this.
779 ph10 324
780 ph10 332
781 ph10 319 Version 7.6 28-Jan-08
782 ph10 300 ---------------------
784 ph10 302 1. A character class containing a very large number of characters with
785     codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
786     overflow.
787 ph10 309
788     2. Patch to cut out the "long long" test in pcrecpp_unittest when
789     HAVE_LONG_LONG is not defined.
791 ph10 303 3. Applied Christian Ehrlicher's patch to update the CMake build files to
792 ph10 304 bring them up to date and include new features. This patch includes:
793 ph10 309
794 ph10 304 - Fixed PH's badly added libz and libbz2 support.
795     - Fixed a problem with static linking.
796 ph10 312 - Added pcredemo. [But later removed - see 7 below.]
797 ph10 304 - Fixed dftables problem and added an option.
798     - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
800     - Added readline support for pcretest.
801 ph10 309 - Added an listing of the option settings after cmake has run.
803 ph10 314 4. A user submitted a patch to Makefile that makes it easy to create
804     "pcre.dll" under mingw when using Configure/Make. I added stuff to
805     Makefile.am that cause it to include this special target, without
806     affecting anything else. Note that the same mingw target plus all
807     the other distribution libraries and programs are now supported
808     when configuring with CMake (see 6 below) instead of with
809     Configure/Make.
810 ph10 309
811 ph10 308 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
812     This is an attempt to solve the reported problem "pcrecpp::no_arg is not
813 ph10 309 exported in the Windows port". It has not yet been confirmed that the patch
814     solves the problem, but it does no harm.
815 ph10 313
816 ph10 311 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
817 ph10 319 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
818 ph10 314 with CMake, and also correct the comment about stack recursion.
819 ph10 313
820 ph10 312 7. Remove the automatic building of pcredemo from the ./configure system and
821 ph10 313 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
822     of a program that users should build themselves after PCRE is installed, so
823     building it automatically is not really right. What is more, it gave
824 ph10 312 trouble in some build environments.
825 ph10 300
826 ph10 319 8. Further tidies to CMakeLists.txt from Sheri and Christian.
827 ph10 308
828 ph10 319
829 ph10 298 Version 7.5 10-Jan-08
830 ph10 263 ---------------------
832     1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
833     values in parens when parsing an RE using the C++ wrapper."
834 ph10 286
835 ph10 264 2. Negative specials like \S did not work in character classes in UTF-8 mode.
836     Characters greater than 255 were excluded from the class instead of being
837     included.
838 ph10 286
839     3. The same bug as (2) above applied to negated POSIX classes such as
840 ph10 265 [:^space:].
841 ph10 286
842 ph10 267 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
843     defined or documented. It seems to have been a typo for PCRE_STATIC, so
844 ph10 286 I have changed it.
846 ph10 268 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
847 ph10 286 first named subpattern) and a construct such as (?&a) would reference the
848     first named subpattern whose name started with "a" (in other words, the
849 ph10 272 length check was missing). Both these problems are fixed. "Subpattern name
850     expected" is now given for (?&) (a zero-length name), and this patch also
851     makes it give the same error for \k'' (previously it complained that that
852     was a reference to a non-existent subpattern).
853 ph10 286
854 ph10 269 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
855 ph10 286 this is right because (?- can be followed by option settings as well as by
856 ph10 269 digits. I have, however, made the messages clearer.
857 ph10 286
858 ph10 270 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
859     than the number used in the conditional) now cause a compile-time error.
860     This is actually not compatible with Perl, which accepts such patterns, but
861     treats the conditional as always being FALSE (as PCRE used to), but it
862     seems to me that giving a diagnostic is better.
863 ph10 286
864 ph10 274 8. Change "alphameric" to the more common word "alphanumeric" in comments
865 ph10 275 and messages.
866 ph10 286
867     9. Fix two occurrences of "backslash" in comments that should have been
868     "backspace".
870     10. Remove two redundant lines of code that can never be obeyed (their function
871     was moved elsewhere).
873     11. The program that makes PCRE's Unicode character property table had a bug
874     which caused it to generate incorrect table entries for sequences of
875 ph10 277 characters that have the same character type, but are in different scripts.
876 ph10 286 It amalgamated them into a single range, with the script of the first of
877 ph10 277 them. In other words, some characters were in the wrong script. There were
878     thirteen such cases, affecting characters in the following ranges:
879 ph10 286
880 ph10 277 U+002b0 - U+002c1
881     U+0060c - U+0060d
882 ph10 286 U+0061e - U+00612
883 ph10 277 U+0064b - U+0065e
884     U+0074d - U+0076d
885     U+01800 - U+01805
886     U+01d00 - U+01d77
887     U+01d9b - U+01dbf
888     U+0200b - U+0200f
889     U+030fc - U+030fe
890     U+03260 - U+0327f
891     U+0fb46 - U+0fbb1
892     U+10450 - U+1049d
893 ph10 286
894 ph10 279 12. The -o option (show only the matching part of a line) for pcregrep was not
895     compatible with GNU grep in that, if there was more than one match in a
896     line, it showed only the first of them. It now behaves in the same way as
897     GNU grep.
898 ph10 286
899     13. If the -o and -v options were combined for pcregrep, it printed a blank
900     line for every non-matching line. GNU grep prints nothing, and pcregrep now
901     does the same. The return code can be used to tell if there were any
902     non-matching lines.
904 ph10 289 14. Added --file-offsets and --line-offsets to pcregrep.
906     15. The pattern (?=something)(?R) was not being diagnosed as a potentially
907 ph10 282 infinitely looping recursion. The bug was that positive lookaheads were not
908 ph10 286 being skipped when checking for a possible empty match (negative lookaheads
909     and both kinds of lookbehind were skipped).
911 ph10 289 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
912 ph10 284 inclusion of <windows.h> to before rather than after the definition of
913     INVALID_FILE_ATTRIBUTES (patch from David Byron).
914 ph10 263
915 ph10 289 17. Specifying a possessive quantifier with a specific limit for a Unicode
916 ph10 286 character property caused pcre_compile() to compile bad code, which led at
917     runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
918     are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
919     caused the error; without that there was no problem.
920 ph10 263
921 ph10 289 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
922 ph10 286
923 ph10 289 19. Added --enable-pcretest-libreadline.
924 ph10 286
925 ph10 289 20. In pcrecpp.cc, the variable 'count' was incremented twice in
926 ph10 288 RE::GlobalReplace(). As a result, the number of replacements returned was
927 ph10 298 double what it should be. I removed one of the increments, but Craig sent a
928     later patch that removed the other one (the right fix) and added unit tests
929     that check the return values (which was not done before).
930 ph10 292
931 ph10 291 21. Several CMake things:
932 ph10 287
933 ph10 291 (1) Arranged that, when cmake is used on Unix, the libraries end up with
934 ph10 292 the names libpcre and libpcreposix, not just pcre and pcreposix.
935 ph10 288
936 ph10 292 (2) The above change means that pcretest and pcregrep are now correctly
937     linked with the newly-built libraries, not previously installed ones.
938 ph10 291
940 ph10 298
941     22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
942     crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
943     UTF-8 newline character). The key issue is that the pattern starts .*;
944     this means that the match must be either at the beginning, or after a
945     newline. The bug was in the code for advancing after a failed match and
946     checking that the new position followed a newline. It was not taking
947 ph10 294 account of UTF-8 characters correctly.
948 ph10 298
949     23. PCRE was behaving differently from Perl in the way it recognized POSIX
950     character classes. PCRE was not treating the sequence [:...:] as a
951     character class unless the ... were all letters. Perl, however, seems to
952     allow any characters between [: and :], though of course it rejects as
953     unknown any "names" that contain non-letters, because all the known class
954     names consist only of letters. Thus, Perl gives an error for [[:1234:]],
955     for example, whereas PCRE did not - it did not recognize a POSIX character
956     class. This seemed a bit dangerous, so the code has been changed to be
957     closer to Perl. The behaviour is not identical to Perl, because PCRE will
958     diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
959     treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
960     Perl does, and where it didn't before.
962 ph10 296 24. Rewrite so as to remove the single use of %n from pcregrep because in some
963     Windows environments %n is disabled by default.
964 ph10 292
966 ph10 260 Version 7.4 21-Sep-07
967 ph10 230 ---------------------
969     1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
970 ph10 231 means that a class such as [\s] counted as "explicit reference to CR or
971     LF". That isn't really right - the whole point of the change was to try to
972     help when there was an actual mention of one of the two characters. So now
973     the change happens only if \r or \n (or a literal CR or LF) character is
974 ph10 230 encountered.
975 ph10 231
976     2. The 32-bit options word was also used for 6 internal flags, but the numbers
977     of both had grown to the point where there were only 3 bits left.
978     Fortunately, there was spare space in the data structure, and so I have
979     moved the internal flags into a new 16-bit field to free up more option
980 ph10 230 bits.
981 ph10 231
982     3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
983     but did not set the internal JCHANGED flag - either of these is enough to
984     control the way the "get" function works - but the PCRE_INFO_JCHANGED
985     facility is supposed to tell if (?J) was ever used, so now (?J) at the
986 ph10 230 start sets both bits.
988 ph10 231 4. Added options (at build time, compile time, exec time) to change \R from
989     matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
990 ph10 230
991 ph10 243 5. doc/pcresyntax.html was missing from the distribution.
993     6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
994 ph10 233 compatibility, even though it is no longer used.
995 ph10 243
996     7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
997 ph10 254 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
998 ph10 259 windows.h file is present (where different names are used). [This was
999 ph10 257 reversed later after testing - see 16 below.]
1000 ph10 243
1001     8. Changed all #include <config.h> to #include "config.h". There were also
1002 ph10 236 some further <pcre.h> cases that I changed to "pcre.h".
1003 ph10 243
1004 ph10 239 9. When pcregrep was used with the --colour option, it missed the line ending
1005 ph10 243 sequence off the lines that it output.
1007     10. It was pointed out to me that arrays of string pointers cause lots of
1008     relocations when a shared library is dynamically loaded. A technique of
1009     using a single long string with a table of offsets can drastically reduce
1010     these. I have refactored PCRE in four places to do this. The result is
1011 ph10 240 dramatic:
1012 ph10 243
1013 ph10 240 Originally: 290
1014     After changing UCP table: 187
1015 ph10 243 After changing error message table: 43
1016 ph10 240 After changing table of "verbs" 36
1017     After changing table of Posix names 22
1018 ph10 243
1019 ph10 240 Thanks to the folks working on Gregex for glib for this insight.
1020 ph10 247
1021 ph10 244 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1022 ph10 247 unicode-properties was also set.
1024 ph10 248 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1025 ph10 231
1026 ph10 253 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1027     checked only for CRLF.
1028 ph10 233
1029 ph10 259 14. Added casts to pcretest.c to avoid compiler warnings.
1031 ph10 256 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1032 ph10 253
1033 ph10 257 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1034 ph10 259 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1035     entirely. This removes changes made in 7 above.
1036 ph10 256
1037 ph10 261 17. The CMake files have been updated, and there is now more information about
1038     building with CMake in the NON-UNIX-USE document.
1039 ph10 257
1040 ph10 261
1041 ph10 228 Version 7.3 28-Aug-07
1042 ph10 157 ---------------------
1044 ph10 189 1. In the rejigging of the build system that eventually resulted in 7.1, the
1045     line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1046     brackets there is not right, since it causes compilers to look for an
1047     installed pcre.h, not the version that is in the source that is being
1048     compiled (which of course may be different). I have changed it back to:
1049 ph10 197
1050 ph10 189 #include "pcre.h"
1051 ph10 197
1052     I have a vague recollection that the change was concerned with compiling in
1053     different directories, but in the new build system, that is taken care of
1054     by the VPATH setting the Makefile.
1056 ph10 190 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1057     when the subject happened to end in the byte 0x85 (e.g. if the last
1058     character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1059     characters but of course it shouldn't be taken as a newline when it is part
1060     of another character. The bug was that, for an unlimited repeat of . in
1061     not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1062     characters when looking for a newline.
1063 ph10 197
1064     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1066     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1067     in debug output.
1069 ph10 201 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1070     long printing in the pcrecpp unittest when running under MinGW.
1071 ph10 189
1072 ph10 197 6. ESC_K was left out of the EBCDIC table.
1073 ph10 189
1074 ph10 197 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1075     parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1076     limit also applies to "virtual nesting" when a pattern is recursive, and in
1077     this case 1000 isn't so big. I have been able to remove this limit at the
1078     expense of backing off one optimization in certain circumstances. Normally,
1079     when pcre_exec() would call its internal match() function recursively and
1080     immediately return the result unconditionally, it uses a "tail recursion"
1081     feature to save stack. However, when a subpattern that can match an empty
1082     string has an unlimited repetition quantifier, it no longer makes this
1083     optimization. That gives it a stack frame in which to save the data for
1084     checking that an empty string has been matched. Previously this was taken
1085     from the 1000-entry workspace that had been reserved. So now there is no
1086     explicit limit, but more stack is used.
1087 ph10 189
1088 ph10 200 8. Applied Daniel's patches to solve problems with the import/export magic
1089     syntax that is required for Windows, and which was going wrong for the
1090     pcreposix and pcrecpp parts of the library. These were overlooked when this
1091     problem was solved for the main library.
1092 ph10 197
1093 ph10 202 9. There were some crude static tests to avoid integer overflow when computing
1094     the size of patterns that contain repeated groups with explicit upper
1095     limits. As the maximum quantifier is 65535, the maximum group length was
1096     set at 30,000 so that the product of these two numbers did not overflow a
1097     32-bit integer. However, it turns out that people want to use groups that
1098     are longer than 30,000 bytes (though not repeat them that many times).
1099     Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1100     made it possible to implement the integer overflow checks in a much more
1101     dynamic way, which I have now done. The artificial limitation on group
1102     length has been removed - we now have only the limit on the total length of
1103     the compiled pattern, which depends on the LINK_SIZE setting.
1104 ph10 208
1105     10. Fixed a bug in the documentation for get/copy named substring when
1106     duplicate names are permitted. If none of the named substrings are set, the
1107     functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1108     empty string.
1110     11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1111     instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1112     because the ] is interpreted as the first data character and the
1113     terminating ] is not found. PCRE has been made compatible with Perl in this
1114     regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1115     cause memory overwriting.
1117 ph10 206 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1118     string has been matched (to stop an infinite loop). It was not recognizing
1119 ph10 208 a conditional subpattern that could match an empty string if that
1120 ph10 206 subpattern was within another subpattern. For example, it looped when
1121 ph10 208 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1122 ph10 206 condition was not nested. This bug has been fixed.
1123 ph10 208
1124 ph10 207 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1125     past the start of the subject in the presence of bytes with the top bit
1126     set, for example "\x8aBCD".
1127 ph10 211
1128 ph10 210 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1129     (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1130 ph10 200
1131 ph10 211 14. Optimized (?!) to (*FAIL).
1132 ph10 202
1133 ph10 212 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1134     This restricts code points to be within the range 0 to 0x10FFFF, excluding
1135     the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1136     full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1137     does: it's just the validity check that is more restrictive.
1138 ph10 220
1139     16. Inserted checks for integer overflows during escape sequence (backslash)
1140     processing, and also fixed erroneous offset values for syntax errors during
1141     backslash processing.
1143 ph10 214 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1144 ph10 220 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1146 ph10 215 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1147     caused an overrun.
1148 ph10 220
1149     19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1150     something other than just ASCII characters) inside a group that had an
1151     unlimited repeat caused a loop at compile time (while checking to see
1152     whether the group could match an empty string).
1154 ph10 217 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1155     [\P{Any}] did so. (Error in the code for printing property names.)
1156 ph10 210
1157 ph10 220 21. An orphan \E inside a character class could cause a crash.
1159     22. A repeated capturing bracket such as (A)? could cause a wild memory
1160 ph10 218 reference during compilation.
1161 ph10 220
1162     23. There are several functions in pcre_compile() that scan along a compiled
1163     expression for various reasons (e.g. to see if it's fixed length for look
1164 ph10 218 behind). There were bugs in these functions when a repeated \p or \P was
1165 ph10 220 present in the pattern. These operators have additional parameters compared
1166     with \d, etc, and these were not being taken into account when moving along
1167 ph10 218 the compiled data. Specifically:
1168 ph10 220
1169     (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1170     length.
1172     (b) An item such as \pL+ within a repeated group could cause crashes or
1173 ph10 218 loops.
1174 ph10 220
1175 ph10 218 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1176 ph10 220 "reference to non-existent subpattern" error.
1178 ph10 221 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1180 ph10 219 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1181 ph10 220 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1182 ph10 211
1183 ph10 222 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1184 ph10 218
1185 ph10 227 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1186 ph10 223 character were causing crashes (broken optimization).
1187 ph10 227
1188 ph10 224 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1189 ph10 227 \p or \P) caused a compile-time loop.
1191 ph10 226 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1192     break. For example, the unstudied pattern [\r\n]A does not match the string
1193     "\r\nA" because change 7.0/46 below moves the current point on by two
1194     characters after failing to match at the start. However, the pattern \nA
1195 ph10 227 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1196     the same is true. There doesn't seem any very clean way out of this, but
1197     what I have chosen to do makes the common cases work: PCRE now takes note
1198     of whether there can be an explicit match for \r or \n anywhere in the
1199     pattern, and if so, 7.0/46 no longer applies. As part of this change,
1200     there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1201     pattern has explicit CR or LF references.
1202 ph10 222
1203 ph10 227 29. Added (*CR) etc for changing newline setting at start of pattern.
1204 ph10 223
1205 ph10 227
1206 ph10 189 Version 7.2 19-Jun-07
1207     ---------------------
1209 ph10 157 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1210     which is apparently normally available under Windows.
1211 ph10 159
1212     2. Re-jig the pcregrep tests with different newline settings in an attempt
1213     to make them independent of the local environment's newline setting.
1214 ph10 157
1215 ph10 160 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1216 ph10 165
1217 ph10 161 4. Some of the "internals" tests were previously cut out when the link size
1218     was not 2, because the output contained actual offsets. The recent new
1219 ph10 165 "Z" feature of pcretest means that these can be cut out, making the tests
1220     usable with all link sizes.
1222 ph10 164 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1223     stack recursion. This gives a massive performance boost under BSD, but just
1224 ph10 165 a small improvement under Linux. However, it saves one field in the frame
1225 ph10 164 in all cases.
1226 ph10 172
1227 ph10 166 6. Added more features from the forthcoming Perl 5.10:
1228 ph10 172
1229 ph10 166 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1230     recursion call. It refers to the nth most recently opened parentheses.
1231 ph10 172
1232 ph10 166 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1233 ph10 172 to be opened parentheses.
1235     (c) Conditions that refer to capturing parentheses can be specified
1236 ph10 167 relatively, for example, (?(-2)... or (?(+3)...
1237 ph10 172
1238 ph10 168 (d) \K resets the start of the current match so that everything before
1239 ph10 172 is not part of it.
1241 ph10 171 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1242 ph10 172
1243 ph10 171 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1244 ph10 172 reference syntax.
1245 ph10 182
1246 ph10 175 (g) (?| introduces a group in which the numbering of parentheses in each
1247 ph10 182 alternative starts with the same number.
1248 ph10 172
1249 ph10 182 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1251 ph10 172 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1254     8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1255     terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1256 ph10 170 for detecting groups that can match an empty string.
1257 ph10 159
1258 ph10 172 9. A pattern with a very large number of alternatives (more than several
1259     hundred) was running out of internal workspace during the pre-compile
1260     phase, where pcre_compile() figures out how much memory will be needed. A
1261     bit of new cunning has reduced the workspace needed for groups with
1262     alternatives. The 1000-alternative test pattern now uses 12 bytes of
1263     workspace instead of running out of the 4096 that are available.
1264 ph10 182
1265 ph10 176 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1266 ph10 172
1267 ph10 179 11. Applied patch from Google to remove an optimization that didn't quite work.
1268     The report of the bug said:
1269 ph10 182
1270 ph10 179 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1271     pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1272 ph10 182 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1273 ph10 185
1274 ph10 184 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1275 ph10 185 it matched the wrong number of bytes.
1276 ph10 172
1277 ph10 179
1278 ph10 155 Version 7.1 24-Apr-07
1279 ph10 98 ---------------------
1281 ph10 111 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1282 ph10 122 that is more "standard", making use of automake and other Autotools. There
1283 ph10 99 is some re-arrangement of the files and adjustment of comments consequent
1284     on this.
1285 ph10 111
1286     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1287     for recursive directory scanning broke on some systems because the files
1288     are not scanned in any specific order and on different systems the order
1289     was different. A call to "sort" has been inserted into RunGrepTest for the
1290     approprate test as a short-term fix. In the longer term there may be an
1291 ph10 100 alternative.
1292 ph10 111
1293 ph10 100 3. I had an email from Eric Raymond about problems translating some of PCRE's
1294 ph10 111 man pages to HTML (despite the fact that I distribute HTML pages, some
1295     people do their own conversions for various reasons). The problems
1296     concerned the use of low-level troff macros .br and .in. I have therefore
1297     removed all such uses from the man pages (some were redundant, some could
1298 ph10 113 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1299     HTML has been updated to handle .nf/.fi and to complain if it encounters
1300     .br or .in.
1301 ph10 111
1302 ph10 100 4. Updated comments in configure.ac that get placed in config.h.in and also
1303 ph10 123 arranged for config.h to be included in the distribution, with the name
1304 ph10 111 config.h.generic, for the benefit of those who have to compile without
1305     Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1307     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1308     Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1309 ph10 127 makevp.bat for the current PCRE, using the additional files
1310 ph10 135 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1311 ph10 111
1312     6. A Windows user reported a minor discrepancy with test 2, which turned out
1313     to be caused by a trailing space on an input line that had got lost in his
1314 ph10 102 copy. The trailing space was an accident, so I've just removed it.
1315 ph10 111
1316 ph10 104 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1317 ph10 111 that is needed.
1319 ph10 105 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1320     as "const" (a) because they are and (b) because it helps the PHP
1321     maintainers who have recently made a script to detect big data structures
1322 ph10 111 in the php code that should be moved to the .rodata section. I remembered
1323     to update Builducptable as well, so it won't revert if ucptable.h is ever
1324 ph10 105 re-created.
1325 ph10 111
1326     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1327 ph10 107 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1328 ph10 111 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1329     support is not required. This saves 1.5-2K of code, which is important in
1330 ph10 107 some applications.
1331 ph10 111
1332 ph10 110 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1333 ph10 111 so as not to refer to the tables, even though these functions will never be
1334     called when UTF-8 support is disabled. Otherwise there are problems with a
1335     shared library.
1337 ph10 118 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1339     (a) It was defining its arguments as char * instead of void *.
1341 ph10 123 (b) It was assuming that all moves were upwards in memory; this was true
1342     a long time ago when I wrote it, but is no longer the case.
1344 ph10 118 The emulated memove() is provided for those environments that have neither
1345 ph10 123 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1346 ph10 118 is clearly not the case, as these two bugs were recently reported.
1347 ph10 123
1348 ph10 111 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1349 ph10 123 and Detrail to create the HTML documentation, the .txt form of the man
1350     pages, and it removes trailing spaces from listed files. It also creates
1351     pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1352     case, it wraps all the #defines with #ifndefs. This script should be run
1353 ph10 111 before "make dist".
1354 ph10 123
1355 ph10 115 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1356     with Unicode property support.
1357 ph10 123
1358     (a) For a maximizing quantifier, if the two different cases of the
1359     character were of different lengths in their UTF-8 codings (there are
1360     some cases like this - I found 11), and the matching function had to
1361 ph10 115 back up over a mixture of the two cases, it incorrectly assumed they
1362     were both the same length.
1363 ph10 123
1364     (b) When PCRE was configured to use the heap rather than the stack for
1365     recursion during matching, it was not correctly preserving the data for
1366     the other case of a UTF-8 character when checking ahead for a match
1367     while processing a minimizing repeat. If the check also involved
1368 ph10 115 matching a wide character, but failed, corruption could cause an
1369     erroneous result when trying to check for a repeat of the original
1370     character.
1371 ph10 123
1372 ph10 116 13. Some tidying changes to the testing mechanism:
1373 ph10 98
1374 ph10 116 (a) The RunTest script now detects the internal link size and whether there
1375     is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1376 ph10 123 values substituted by "configure". (The RunGrepTest script already did
1377     this for UTF-8.) The configure.ac script no longer substitutes the
1378     relevant variables.
1380 ph10 116 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1381     with length and offset values. This means that the output is different
1382     for different internal link sizes. Test 2 is skipped for link sizes
1383     other than 2 because of this, bypassing the problem. Unfortunately,
1384     there was also a test in test 3 (the locale tests) that used /B and
1385 ph10 123 failed for link sizes other than 2. Rather than cut the whole test out,
1386     I have added a new /Z option to pcretest that replaces the length and
1387     offset values with spaces. This is now used to make test 3 independent
1388 ph10 122 of link size. (Test 2 will be tidied up later.)
1389 ph10 123
1390     14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1391 ph10 122 segmentation fault instead of returning the appropriate error message.
1392 ph10 134
1393 ph10 131 15. In multiline mode when the newline sequence was set to "any", the pattern
1394 ph10 134 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1395     This doesn't seem right; it now treats the CRLF combination as the line
1396     ending, and so does not match in that case. It's only a pattern such as ^$
1397     that would hit this one: something like ^ABC$ would have failed after \r
1398     and then tried again after \r\n.
1400 ph10 131 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1401 ph10 134 in an attempt to make files that differ only in their line terminators
1402     compare equal. This works on Linux.
1403 ph10 142
1404 ph10 141 17. Under certain error circumstances pcregrep might try to free random memory
1405     as it exited. This is now fixed, thanks to valgrind.
1406 ph10 142
1407 ph10 141 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1408 ph10 142 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1409     was because its rules for how to advance for /g after matching an empty
1410 ph10 143 string at the end of a line did not allow for this case. They now check for
1411     it specially.
1412 ph10 150
1413     20. pcretest is supposed to handle patterns and data of any length, by
1414     extending its buffers when necessary. It was getting this wrong when the
1415 ph10 147 buffer for a data line had to be extended.
1416 ph10 150
1417 ph10 149 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1418 ph10 150 CRLF as a newline sequence.
1419 ph10 152
1420 ph10 151 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1421 ph10 152 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1422     I have nevertheless tidied it up.
1423 ph10 111
1424 ph10 152 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1425 ph10 123
1426 ph10 153 24. Added a man page for pcre-config.
1427 ph10 152
1428 ph10 153
1429 nigel 93 Version 7.0 19-Dec-06
1430     ---------------------
1432     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1433     moving to gcc 4.1.1.
1435     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1436     sys/time.h, which is documented as needed for this function. It doesn't
1437     seem to matter on Linux, but it showed up on some releases of OS X.
1439     3. It seems that there are systems where bytes whose values are greater than
1440     127 match isprint() in the "C" locale. The "C" locale should be the
1441     default when a C program starts up. In most systems, only ASCII printing
1442     characters match isprint(). This difference caused the output from pcretest
1443     to vary, making some of the tests fail. I have changed pcretest so that:
1445     (a) When it is outputting text in the compiled version of a pattern, bytes
1446     other than 32-126 are always shown as hex escapes.
1448     (b) When it is outputting text that is a matched part of a subject string,
1449     it does the same, unless a different locale has been set for the match
1450     (using the /L modifier). In this case, it uses isprint() to decide.
1452     4. Fixed a major bug that caused incorrect computation of the amount of memory
1453     required for a compiled pattern when options that changed within the
1454     pattern affected the logic of the preliminary scan that determines the
1455     length. The relevant options are -x, and -i in UTF-8 mode. The result was
1456     that the computed length was too small. The symptoms of this bug were
1457     either the PCRE error "internal error: code overflow" from pcre_compile(),
1458     or a glibc crash with a message such as "pcretest: free(): invalid next
1459     size (fast)". Examples of patterns that provoked this bug (shown in
1460     pcretest format) are:
1462     /(?-x: )/x
1463     /(?x)(?-x: \s*#\s*)/
1464     /((?i)[\x{c0}])/8
1465     /(?i:[\x{c0}])/8
1467     HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1468     is now done differently.
1470     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1471     wrapper classes; (b) implement a new function in the C++ scanner that is
1472     more efficient than the old way of doing things because it avoids levels of
1473     recursion in the regex matching; (c) add a paragraph to the documentation
1474     for the FullMatch() function.
1476     6. The escape sequence \n was being treated as whatever was defined as
1477     "newline". Not only was this contrary to the documentation, which states
1478     that \n is character 10 (hex 0A), but it also went horribly wrong when
1479     "newline" was defined as CRLF. This has been fixed.
1481     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1482     was being set to -1 for the "end of line" case (supposedly a value that no
1483     character can have). Though this value is never used (the check for end of
1484     line is "zero bytes in current character"), it caused compiler complaints.
1485     I've changed it to 0xffffffff.
1487     8. In pcre_version.c, the version string was being built by a sequence of
1488     C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1489     string (as it is for production releases) called a macro with an empty
1490     argument. The C standard says the result of this is undefined. The gcc
1491     compiler treats it as an empty string (which was what was wanted) but it is
1492     reported that Visual C gives an error. The source has been hacked around to
1493     avoid this problem.
1495     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1496     builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1497     instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1498     of them did).
1500     10. Originally, pcretest opened its input and output without "b"; then I was
1501     told that "b" was needed in some environments, so it was added for release
1502     5.0 to both the input and output. (It makes no difference on Unix-like
1503     systems.) Later I was told that it is wrong for the input on Windows. I've
1504     now abstracted the modes into two macros, to make it easier to fiddle with
1505     them, and removed "b" from the input mode under Windows.
1507     11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1509     12. Added -help and --help to pcretest as an official way of being reminded
1510     of the options.
1512     13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1513     and pcrecpp.cc because they annoy compilers at high warning levels.
1515     14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1517     15. Fixed an occurrence of == in configure.ac that should have been = (shell
1518     scripts are not C programs :-) and which was not noticed because it works
1519     on Linux.
1521     16. pcretest is supposed to handle any length of pattern and data line (as one
1522     line or as a continued sequence of lines) by extending its input buffer if
1523     necessary. This feature was broken for very long pattern lines, leading to
1524     a string of junk being passed to pcre_compile() if the pattern was longer
1525     than about 50K.
1527     17. I have done a major re-factoring of the way pcre_compile() computes the
1528     amount of memory needed for a compiled pattern. Previously, there was code
1529     that made a preliminary scan of the pattern in order to do this. That was
1530     OK when PCRE was new, but as the facilities have expanded, it has become
1531     harder and harder to keep it in step with the real compile phase, and there
1532     have been a number of bugs (see for example, 4 above). I have now found a
1533     cunning way of running the real compile function in a "fake" mode that
1534     enables it to compute how much memory it would need, while actually only
1535     ever using a few hundred bytes of working memory and without too many
1536     tests of the mode. This should make future maintenance and development
1537     easier. A side effect of this work is that the limit of 200 on the nesting
1538     depth of parentheses has been removed (though this was never a serious
1539     limitation, I suspect). However, there is a downside: pcre_compile() now
1540     runs more slowly than before (30% or more, depending on the pattern). I
1541     hope this isn't a big issue. There is no effect on runtime performance.
1543     18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1544     newline (only possible for the last line of a file) and it was a
1545     pattern that set a locale (followed by /Lsomething), pcretest crashed.
1547     19. Added additional timing features to pcretest. (1) The -tm option now times
1548     matching only, not compiling. (2) Both -t and -tm can be followed, as a
1549     separate command line item, by a number that specifies the number of
1550     repeats to use when timing. The default is 50000; this gives better
1551     precision, but takes uncomfortably long for very large patterns.
1553     20. Extended pcre_study() to be more clever in cases where a branch of a
1554     subpattern has no definite first character. For example, (a*|b*)[cd] would
1555     previously give no result from pcre_study(). Now it recognizes that the
1556     first character must be a, b, c, or d.
1558     21. There was an incorrect error "recursive call could loop indefinitely" if
1559     a subpattern (or the entire pattern) that was being tested for matching an
1560     empty string contained only one non-empty item after a nested subpattern.
1561     For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1562     incorrectly, because the \d was being skipped in the check.
1564     22. The pcretest program now has a new pattern option /B and a command line
1565     option -b, which is equivalent to adding /B to every pattern. This causes
1566     it to show the compiled bytecode, without the additional information that
1567     -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1568     is the same as /B/I).
1570     23. A new optimization is now able automatically to treat some sequences such
1571     as a*b as a*+b. More specifically, if something simple (such as a character
1572     or a simple class like \d) has an unlimited quantifier, and is followed by
1573     something that cannot possibly match the quantified thing, the quantifier
1574     is automatically "possessified".
1576     24. A recursive reference to a subpattern whose number was greater than 39
1577     went wrong under certain circumstances in UTF-8 mode. This bug could also
1578     have affected the operation of pcre_study().
1580     25. Realized that a little bit of performance could be had by replacing
1581     (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1583     26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1585     27. Possessive quantifiers such as a++ were previously implemented by turning
1586     them into atomic groups such as ($>a+). Now they have their own opcodes,
1587     which improves performance. This includes the automatically created ones
1588     from 23 above.
1590     28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1591     lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1592     the first matched character to be a colon. This applied both to named and
1593     numbered groups.
1595     29. The ucpinternal.h header file was missing its idempotency #ifdef.
1597     30. I was sent a "project" file called libpcre.a.dev which I understand makes
1598     building PCRE on Windows easier, so I have included it in the distribution.
1600     31. There is now a check in pcretest against a ridiculously large number being
1601     returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1602     loop, the loop is abandoned.
1604     32. Forward references to subpatterns in conditions such as (?(2)...) where
1605     subpattern 2 is defined later cause pcre_compile() to search forwards in
1606     the pattern for the relevant set of parentheses. This search went wrong
1607     when there were unescaped parentheses in a character class, parentheses
1608     escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1610     33. "Subroutine" calls and backreferences were previously restricted to
1611     referencing subpatterns earlier in the regex. This restriction has now
1612     been removed.
1614     34. Added a number of extra features that are going to be in Perl 5.10. On the
1615     whole, these are just syntactic alternatives for features that PCRE had
1616     previously implemented using the Python syntax or my own invention. The
1617     other formats are all retained for compatibility.
1619     (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1620     as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1621     also .NET compatible.
1623     (b) A recursion or subroutine call to a named group can now be defined as
1624     (?&name) as well as (?P>name).
1626     (c) A backreference to a named group can now be defined as \k<name> or
1627     \k'name' as well as (?P=name). The new forms, as well as being in Perl
1628     5.10, are also .NET compatible.
1630     (d) A conditional reference to a named group can now use the syntax
1631     (?(<name>) or (?('name') as well as (?(name).
1633     (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1634     groups (named and numbered) that are never evaluated inline, but can be
1635     called as "subroutines" from elsewhere. In effect, the DEFINE condition
1636     is always false. There may be only one alternative in such a group.
1638     (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1639     as the simple (?(R). The condition is true only if the most recent
1640     recursion is that of the given number or name. It does not search out
1641     through the entire recursion stack.
1643     (g) The escape \gN or \g{N} has been added, where N is a positive or
1644     negative number, specifying an absolute or relative reference.
1646     35. Tidied to get rid of some further signed/unsigned compiler warnings and
1647     some "unreachable code" warnings.
1649     36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1650     things, this adds five new scripts.
1652     37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1653     There were also incompatibilities regarding the handling of \Q..\E inside
1654     character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1655     hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1657     38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1658     matches an empty string, and forcibly breaks the loop. There were bugs in
1659     this code in non-simple cases. For a pattern such as ^(a()*)* matched
1660     against aaaa the result was just "a" rather than "aaaa", for example. Two
1661     separate and independent bugs (that affected different cases) have been
1662     fixed.
1664     39. Refactored the code to abolish the use of different opcodes for small
1665     capturing bracket numbers. This is a tidy that I avoided doing when I
1666     removed the limit on the number of capturing brackets for 3.5 back in 2001.
1667     The new approach is not only tidier, it makes it possible to reduce the
1668     memory needed to fix the previous bug (38).
1670     40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1671     sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1672     processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1673     mode.
1675     41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1676     report.
1678     42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1679     copy construction and assignment in the C++ wrapper.
1681     43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1682     couple of bugs that could have given wrong results in the "--newline=crlf"
1683     case.
1685     44. Added a number of casts and did some reorganization of signed/unsigned int
1686     variables following suggestions from Dair Grant. Also renamed the variable
1687     "this" as "item" because it is a C++ keyword.
1689     45. Arranged for dftables to add
1691     #include "pcre_internal.h"
1693     to pcre_chartables.c because without it, gcc 4.x may remove the array
1694     definition from the final binary if PCRE is built into a static library and
1695     dead code stripping is activated.
1697     46. For an unanchored pattern, if a match attempt fails at the start of a
1698     newline sequence, and the newline setting is CRLF or ANY, and the next two
1699     characters are CRLF, advance by two characters instead of one.
1702 nigel 91 Version 6.7 04-Jul-06
1703     ---------------------
1705     1. In order to handle tests when input lines are enormously long, pcretest has
1706     been re-factored so that it automatically extends its buffers when
1707     necessary. The code is crude, but this _is_ just a test program. The
1708     default size has been increased from 32K to 50K.
1710     2. The code in pcre_study() was using the value of the re argument before
1711     testing it for NULL. (Of course, in any sensible call of the function, it
1712     won't be NULL.)
1714     3. The memmove() emulation function in pcre_internal.h, which is used on
1715     systems that lack both memmove() and bcopy() - that is, hardly ever -
1716     was missing a "static" storage class specifier.
1718     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
1719     containing an extended class (one that cannot be represented by a bitmap
1720     because it contains high-valued characters or Unicode property items, e.g.
1721     [\pZ]). Almost always one would set UTF-8 mode when processing such a
1722     pattern, but PCRE should not loop if you do not (it no longer does).
1723     [Detail: two cases were found: (a) a repeated subpattern containing an
1724     extended class; (b) a recursive reference to a subpattern that followed a
1725     previous extended class. It wasn't skipping over the extended class
1726     correctly when UTF-8 mode was not set.]
1728     5. A negated single-character class was not being recognized as fixed-length
1729     in lookbehind assertions such as (?<=[^f]), leading to an incorrect
1730     compile error "lookbehind assertion is not fixed length".
1732     6. The RunPerlTest auxiliary script was showing an unexpected difference
1733     between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
1734     write a Perl script that can interpret lines of an input file either as
1735     byte characters or as UTF-8, which is what "perltest" was being required to
1736     do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
1737     can't do is switch easily at run time between having the "use utf8;" pragma
1738     or not. In the end, I fudged it by using the RunPerlTest script to insert
1739     "use utf8;" explicitly for the UTF-8 tests.
1741     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
1742     the end of the subject string, contrary to the documentation and to what
1743     Perl does. This was true of both matching functions. Now it matches only at
1744     the start of the subject and immediately after *internal* newlines.
1746     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
1747     a pointer to an int instead of a pointer to an unsigned long int. This
1748     caused problems on 64-bit systems.
1750     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
1751     instance of the 'standard' template library not being so standard".
1753     10. There was no check on the number of named subpatterns nor the maximum
1754     length of a subpattern name. The product of these values is used to compute
1755     the size of the memory block for a compiled pattern. By supplying a very
1756     long subpattern name and a large number of named subpatterns, the size
1757     computation could be caused to overflow. This is now prevented by limiting
1758     the length of names to 32 characters, and the number of named subpatterns
1759     to 10,000.
1761     11. Subpatterns that are repeated with specific counts have to be replicated in
1762     the compiled pattern. The size of memory for this was computed from the
1763     length of the subpattern and the repeat count. The latter is limited to
1764     65535, but there was no limit on the former, meaning that integer overflow
1765     could in principle occur. The compiled length of a repeated subpattern is
1766     now limited to 30,000 bytes in order to prevent this.
1768     12. Added the optional facility to have named substrings with the same name.
1770     13. Added the ability to use a named substring as a condition, using the
1771     Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
1772     are numbers (not recommended). Forward references are permitted.
1774     14. Added forward references in named backreferences (if you see what I mean).
1776     15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
1777     pattern could run off the end of the subject. For example, the pattern
1778     "(?s)(.{1,5})"8 did this with the subject "ab".
1780     16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
1781     PCRE_CASELESS was set when matching characters that were quantified with ?
1782     or *.
1784     17. A character class other than a single negated character that had a minimum
1785     but no maximum quantifier - for example [ab]{6,} - was not handled
1786     correctly by pce_dfa_exec(). It would match only one character.
1788     18. A valid (though odd) pattern that looked like a POSIX character
1789     class but used an invalid character after [ (for example [[,abc,]]) caused
1790     pcre_compile() to give the error "Failed: internal error: code overflow" or
1791     in some cases to crash with a glibc free() error. This could even happen if
1792     the pattern terminated after [[ but there just happened to be a sequence of
1793     letters, a binary zero, and a closing ] in the memory that followed.
1795     19. Perl's treatment of octal escapes in the range \400 to \777 has changed
1796     over the years. Originally (before any Unicode support), just the bottom 8
1797     bits were taken. Thus, for example, \500 really meant \100. Nowadays the
1798     output from "man perlunicode" includes this:
1800     The regular expression compiler produces polymorphic opcodes. That
1801     is, the pattern adapts to the data and automatically switches to
1802     the Unicode character scheme when presented with Unicode data--or
1803     instead uses a traditional byte scheme when presented with byte
1804     data.
1806     Sadly, a wide octal escape does not cause a switch, and in a string with
1807     no other multibyte characters, these octal escapes are treated as before.
1808     Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
1809     /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
1810     Unicode string.
1812     I have not perpetrated such confusion in PCRE. Up till now, it took just
1813     the bottom 8 bits, as in old Perl. I have now made octal escapes with
1814     values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
1815     translate to the appropriate multibyte character.
1817     29. Applied some refactoring to reduce the number of warnings from Microsoft
1818     and Borland compilers. This has included removing the fudge introduced
1819     seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
1820     a warning about an unused variable.
1822     21. PCRE has not included VT (character 0x0b) in the set of whitespace
1823     characters since release 4.0, because Perl (from release 5.004) does not.
1824     [Or at least, is documented not to: some releases seem to be in conflict
1825     with the documentation.] However, when a pattern was studied with
1826     pcre_study() and all its branches started with \s, PCRE still included VT
1827     as a possible starting character. Of course, this did no harm; it just
1828     caused an unnecessary match attempt.
1830     22. Removed a now-redundant internal flag bit that recorded the fact that case
1831     dependency changed within the pattern. This was once needed for "required
1832     byte" processing, but is no longer used. This recovers a now-scarce options
1833     bit. Also moved the least significant internal flag bit to the most-
1834     significant bit of the word, which was not previously used (hangover from
1835     the days when it was an int rather than a uint) to free up another bit for
1836     the future.
1838     23. Added support for CRLF line endings as well as CR and LF. As well as the
1839     default being selectable at build time, it can now be changed at runtime
1840     via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
1841     specify that it is scanning data with non-default line endings.
1843     24. Changed the definition of CXXLINK to make it agree with the definition of
1844     LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
1846     25. Applied Ian Taylor's patches to avoid using another stack frame for tail
1847     recursions. This makes a big different to stack usage for some patterns.
1849     26. If a subpattern containing a named recursion or subroutine reference such
1850     as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
1851     the space required for the compiled pattern went wrong and gave too small a
1852     value. Depending on the environment, this could lead to "Failed: internal
1853     error: code overflow at offset 49" or "glibc detected double free or
1854     corruption" errors.
1856     27. Applied patches from Google (a) to support the new newline modes and (b) to
1857     advance over multibyte UTF-8 characters in GlobalReplace.
1859     28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
1860     difference for some implementation of PCRE in some Windows version.
1862     29. Added some extra testing facilities to pcretest:
1864     \q<number> in a data line sets the "match limit" value
1865     \Q<number> in a data line sets the "match recursion limt" value
1866     -S <number> sets the stack size, where <number> is in megabytes
1868     The -S option isn't available for Windows.
1871 nigel 89 Version 6.6 06-Feb-06
1872     ---------------------
1874     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
1875     in pcreposix.h. I have copied the definition from pcre.h.
1877     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
1878     because pcre.h is no longer a built file.
1880     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
1881     not normally included in the compiled code.
1884 nigel 87 Version 6.5 01-Feb-06
1885     ---------------------
1887     1. When using the partial match feature with pcre_dfa_exec(), it was not
1888     anchoring the second and subsequent partial matches at the new starting
1889     point. This could lead to incorrect results. For example, with the pattern
1890     /1234/, partially matching against "123" and then "a4" gave a match.
1892     2. Changes to pcregrep:
1894     (a) All non-match returns from pcre_exec() were being treated as failures
1895     to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
1896     error message is output. Some extra information is given for the
1898     probably the only errors that are likely to be caused by users (by
1899     specifying a regex that has nested indefinite repeats, for instance).
1900     If there are more than 20 of these errors, pcregrep is abandoned.
1902     (b) A binary zero was treated as data while matching, but terminated the
1903     output line if it was written out. This has been fixed: binary zeroes
1904     are now no different to any other data bytes.
1906     (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
1907     used to set a locale for matching. The --locale=xxxx long option has
1908     been added (no short equivalent) to specify a locale explicitly on the
1909     pcregrep command, overriding the environment variables.
1911     (d) When -B was used with -n, some line numbers in the output were one less
1912     than they should have been.
1914     (e) Added the -o (--only-matching) option.
1916     (f) If -A or -C was used with -c (count only), some lines of context were
1917     accidentally printed for the final match.
1919     (g) Added the -H (--with-filename) option.
1921     (h) The combination of options -rh failed to suppress file names for files
1922     that were found from directory arguments.
1924     (i) Added the -D (--devices) and -d (--directories) options.
1926     (j) Added the -F (--fixed-strings) option.
1928     (k) Allow "-" to be used as a file name for -f as well as for a data file.
1930     (l) Added the --colo(u)r option.
1932     (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
1933     is not present by default.
1935     3. A nasty bug was discovered in the handling of recursive patterns, that is,
1936     items such as (?R) or (?1), when the recursion could match a number of
1937     alternatives. If it matched one of the alternatives, but subsequently,
1938     outside the recursion, there was a failure, the code tried to back up into
1939     the recursion. However, because of the way PCRE is implemented, this is not
1940     possible, and the result was an incorrect result from the match.
1942     In order to prevent this happening, the specification of recursion has
1943     been changed so that all such subpatterns are automatically treated as
1944     atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
1946     4. I had overlooked the fact that, in some locales, there are characters for
1947     which isalpha() is true but neither isupper() nor islower() are true. In
1948     the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
1949     and ordfeminine) are like this. This affected the treatment of \w and \W
1950     when they appeared in character classes, but not when they appeared outside
1951     a character class. The bit map for "word" characters is now created
1952     separately from the results of isalnum() instead of just taking it from the
1953     upper, lower, and digit maps. (Plus the underscore character, of course.)
1955     5. The above bug also affected the handling of POSIX character classes such as
1956     [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
1957     permanent tables. Instead, the bit maps for such a class were previously
1958     created as the appropriate unions of the upper, lower, and digit bitmaps.
1959     Now they are created by subtraction from the [[:word:]] class, which has
1960     its own bitmap.
1962     6. The [[:blank:]] character class matches horizontal, but not vertical space.
1963     It is created by subtracting the vertical space characters (\x09, \x0a,
1964     \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
1965     subtraction was done in the overall bitmap for a character class, meaning
1966     that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
1967     be recognized. This bug has been fixed.
1969     7. Patches from the folks at Google:
1971     (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
1972     real life, but is still worth protecting against".
1974     (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
1975     regular expressions".
1977     (c) pcre_scanner.cc: avoid use of std::count() because not all systems
1978     have it.
1980     (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
1981     "configure" and the latter not, in order to fix a problem somebody had
1982     with compiling the Arg class on HP-UX.
1984     (e) Improve the error-handling of the C++ wrapper a little bit.
1986     (f) New tests for checking recursion limiting.
1988     8. The pcre_memmove() function, which is used only if the environment does not
1989     have a standard memmove() function (and is therefore rarely compiled),
1990     contained two bugs: (a) use of int instead of size_t, and (b) it was not
1991     returning a result (though PCRE never actually uses the result).
1993     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
1994     large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
1995     returned instead of calling malloc() with an overflowing number that would
1996     most likely cause subsequent chaos.
1998     10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2000     11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2001     with this option is matched, the nmatch and pmatch options of regexec() are
2002     ignored.
2004     12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2005     provided in case anyone wants to the the POSIX interface with UTF-8
2006     strings.
2008     13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2009     C++ linking (needed for some HP-UX environments).
2011     14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2012     (unused parameter) and in the pcre_printint() function (omitted "default"
2013     switch label when the default is to do nothing).
2015     15. Added some code to make it possible, when PCRE is compiled as a C++
2016     library, to replace subject pointers for pcre_exec() with a smart pointer
2017     class, thus making it possible to process discontinuous strings.
2019     16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2020     much the same function. They were added by different people who were trying
2021     to make PCRE easy to compile on non-Unix systems. It has been suggested
2022     that PCRE_EXPORT be abolished now that there is more automatic apparatus
2023     for compiling on Windows systems. I have therefore replaced it with
2024     PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2025     defaults to "extern" for C or "extern C" for C++, which works fine on
2026     Unix-like systems. It is now possible to override the value of PCRE_DATA_
2027     SCOPE with something explicit in config.h. In addition:
2029     (a) pcreposix.h still had just "extern" instead of either of these macros;
2030     I have replaced it with PCRE_DATA_SCOPE.
2032     (b) Functions such as _pcre_xclass(), which are internal to the library,
2033     but external in the C sense, all had PCRE_EXPORT in their definitions.
2034     This is apparently wrong for the Windows case, so I have removed it.
2035     (It makes no difference on Unix-like systems.)
2037     17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2038     of recursive calls to match(). This is different to MATCH_LIMIT because
2039     that limits the total number of calls to match(), not all of which increase
2040     the depth of recursion. Limiting the recursion depth limits the amount of
2041     stack (or heap if NO_RECURSE is set) that is used. The default can be set
2042     when PCRE is compiled, and changed at run time. A patch from Google adds
2043     this functionality to the C++ interface.
2045     18. Changes to the handling of Unicode character properties:
2047     (a) Updated the table to Unicode 4.1.0.
2049     (b) Recognize characters that are not in the table as "Cn" (undefined).
2051     (c) I revised the way the table is implemented to a much improved format
2052     which includes recognition of ranges. It now supports the ranges that
2053     are defined in UnicodeData.txt, and it also amalgamates other
2054     characters into ranges. This has reduced the number of entries in the
2055     table from around 16,000 to around 3,000, thus reducing its size
2056     considerably. I realized I did not need to use a tree structure after
2057     all - a binary chop search is just as efficient. Having reduced the
2058     number of entries, I extended their size from 6 bytes to 8 bytes to
2059     allow for more data.
2061     (d) Added support for Unicode script names via properties such as \p{Han}.
2063     19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2064     matching that character.
2066     20. When matching a repeated Unicode property with a minimum greater than zero,
2067     (for example \pL{2,}), PCRE could look past the end of the subject if it
2068     reached it while seeking the minimum number of characters. This could
2069     happen only if some of the characters were more than one byte long, because
2070     there is a check for at least the minimum number of bytes.
2072     21. Refactored the implementation of \p and \P so as to be more general, to
2073     allow for more different types of property in future. This has changed the
2074     compiled form incompatibly. Anybody with saved compiled patterns that use
2075     \p or \P will have to recompile them.
2077     22. Added "Any" and "L&" to the supported property types.
2079     23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2080     but give a compile time error if the value is greater than 0xff.
2082     24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2083     accidentally not being installed or uninstalled.
2085     25. The pcre.h file was built from pcre.h.in, but the only changes that were
2086     made were to insert the current release number. This seemed silly, because
2087     it made things harder for people building PCRE on systems that don't run
2088     "configure". I have turned pcre.h into a distributed file, no longer built
2089     by "configure", with the version identification directly included. There is
2090     no longer a pcre.h.in file.
2092     However, this change necessitated a change to the pcre-config script as
2093     well. It is built from pcre-config.in, and one of the substitutions was the
2094     release number. I have updated configure.ac so that ./configure now finds
2095     the release number by grepping pcre.h.
2097     26. Added the ability to run the tests under valgrind.
2100 nigel 85 Version 6.4 05-Sep-05
2101     ---------------------
2103     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2104     "--" to be printed when multiple files were scanned, even when none of the
2105     -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2106     consider it to be a bug, and have restored the previous behaviour.
2108     2. A couple of code tidies to get rid of compiler warnings.
2110     3. The pcretest program used to cheat by referring to symbols in the library
2111     whose names begin with _pcre_. These are internal symbols that are not
2112     really supposed to be visible externally, and in some environments it is
2113     possible to suppress them. The cheating is now confined to including
2114     certain files from the library's source, which is a bit cleaner.
2116     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2117     file's purpose clearer.
2119     5. Reorganized pcre_ucp_findchar().
2122 nigel 83 Version 6.3 15-Aug-05
2123     ---------------------
2125     1. The file libpcre.pc.in did not have general read permission in the tarball.
2127     2. There were some problems when building without C++ support:
2129     (a) If C++ support was not built, "make install" and "make test" still
2130     tried to test it.
2132     (b) There were problems when the value of CXX was explicitly set. Some
2133     changes have been made to try to fix these, and ...
2135     (c) --disable-cpp can now be used to explicitly disable C++ support.
2137     (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2138     backslash in a target when C++ was disabled. This confuses some
2139     versions of "make", apparently. Using an intermediate variable solves
2140     this. (Same for CPP_LOBJ.)
2142     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2143     (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2144     necessary on certain architectures.
2146     4. Added a setting of -export-symbols-regex to the link command to remove
2147     those symbols that are exported in the C sense, but actually are local
2148     within the library, and not documented. Their names all begin with
2149     "_pcre_". This is not a perfect job, because (a) we have to except some
2150     symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2151     available (and never for static libraries). I have made a note to try to
2152     find a way round (a) in the future.
2155 nigel 81 Version 6.2 01-Aug-05
2156     ---------------------
2158     1. There was no test for integer overflow of quantifier values. A construction
2159     such as {1111111111111111} would give undefined results. What is worse, if
2160     a minimum quantifier for a parenthesized subpattern overflowed and became
2161     negative, the calculation of the memory size went wrong. This could have
2162     led to memory overwriting.
2164     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2166     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2167     operating environments where this matters.
2169     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2170     PCRE options from within the C++ wrapper.
2172     5. Named capturing subpatterns were not being correctly counted when a pattern
2173     was compiled. This caused two problems: (a) If there were more than 100
2174     such subpatterns, the calculation of the memory needed for the whole
2175     compiled pattern went wrong, leading to an overflow error. (b) Numerical
2176     back references of the form \12, where the number was greater than 9, were
2177     not recognized as back references, even though there were sufficient
2178     previous subpatterns.
2180     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2181     versions of gcc, e.g. 2.95.4.
2184 nigel 79 Version 6.1 21-Jun-05
2185     ---------------------
2187     1. There was one reference to the variable "posix" in pcretest.c that was not
2188     surrounded by "#if !defined NOPOSIX".
2190     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2191     the cross-check on the old pcre_info() function, for the benefit of the
2192     cut-down version of PCRE that is currently imported into Exim.
2194     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2195     allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2196     patterns that start (?i) but I don't think that matters. The use of (?i) is
2197     just an example; this all applies to the other options as well.
2199     4. Since libtool seems to echo the compile commands it is issuing, the output
2200     from "make" can be reduced a bit by putting "@" in front of each libtool
2201     compile command.
2203     5. Patch from the folks at Google for configure.in to be a bit more thorough
2204     in checking for a suitable C++ installation before trying to compile the
2205     C++ stuff. This should fix a reported problem when a compiler was present,
2206     but no suitable headers.
2208     6. The man pages all had just "PCRE" as their title. I have changed them to
2209     be the relevant file name. I have also arranged that these names are
2210     retained in the file doc/pcre.txt, which is a concatenation in text format
2211     of all the man pages except the little individual ones for each function.
2213     7. The NON-UNIX-USE file had not been updated for the different set of source
2214     files that come with release 6. I also added a few comments about the C++
2215     wrapper.
2218 nigel 77 Version 6.0 07-Jun-05
2219     ---------------------
2221     1. Some minor internal re-organization to help with my DFA experiments.
2223     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2224     didn't matter for the library itself when fully configured, but did matter
2225     when compiling without UCP support, or within Exim, where the ucp files are
2226     not imported.
2228     3. Refactoring of the library code to split up the various functions into
2229     different source modules. The addition of the new DFA matching code (see
2230     below) to a single monolithic source would have made it really too
2231     unwieldy, quite apart from causing all the code to be include in a
2232     statically linked application, when only some functions are used. This is
2233     relevant even without the DFA addition now that patterns can be compiled in
2234     one application and matched in another.
2236     The downside of splitting up is that there have to be some external
2237     functions and data tables that are used internally in different modules of
2238     the library but which are not part of the API. These have all had their
2239     names changed to start with "_pcre_" so that they are unlikely to clash
2240     with other external names.
2242     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2243     a different (DFA) algorithm. Although it is slower than the original
2244     function, it does have some advantages for certain types of matching
2245     problem.
2247     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2248     including restarting after a partial match.
2250     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2251     defined when compiling for Windows was sent to me. I have put it into the
2252     code, though I have no means of testing or verifying it.
2254     7. Added the pcre_refcount() auxiliary function.
2256     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2257     match before or at the first newline in the subject string. In pcretest,
2258     the /f option on a pattern can be used to set this.
2260     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2261     would behave wrongly. This has been present in PCRE since release 4.0.
2263     10. A number of changes to the pcregrep command:
2265     (a) Refactored how -x works; insert ^(...)$ instead of setting
2266     PCRE_ANCHORED and checking the length, in preparation for adding
2267     something similar for -w.
2269     (b) Added the -w (match as a word) option.
2271     (c) Refactored the way lines are read and buffered so as to have more
2272     than one at a time available.
2274     (d) Implemented a pcregrep test script.
2276     (e) Added the -M (multiline match) option. This allows patterns to match
2277     over several lines of the subject. The buffering ensures that at least
2278     8K, or the rest of the document (whichever is the shorter) is available
2279     for matching (and similarly the previous 8K for lookbehind assertions).
2281     (f) Changed the --help output so that it now says
2283     -w, --word-regex(p)
2285     instead of two lines, one with "regex" and the other with "regexp"
2286     because that confused at least one person since the short forms are the
2287     same. (This required a bit of code, as the output is generated
2288     automatically from a table. It wasn't just a text change.)
2290     (g) -- can be used to terminate pcregrep options if the next thing isn't an
2291     option but starts with a hyphen. Could be a pattern or a path name
2292     starting with a hyphen, for instance.
2294     (h) "-" can be given as a file name to represent stdin.
2296     (i) When file names are being printed, "(standard input)" is used for
2297     the standard input, for compatibility with GNU grep. Previously
2298     "<stdin>" was used.
2300     (j) The option --label=xxx can be used to supply a name to be used for
2301     stdin when file names are being printed. There is no short form.
2303     (k) Re-factored the options decoding logic because we are going to add
2304     two more options that take data. Such options can now be given in four
2305     different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2307     (l) Added the -A, -B, and -C options for requesting that lines of context
2308     around matches be printed.
2310     (m) Added the -L option to print the names of files that do not contain
2311     any matching lines, that is, the complement of -l.
2313     (n) The return code is 2 if any file cannot be opened, but pcregrep does
2314     continue to scan other files.
2316     (o) The -s option was incorrectly implemented. For compatibility with other
2317     greps, it now suppresses the error message for a non-existent or non-
2318     accessible file (but not the return code). There is a new option called
2319     -q that suppresses the output of matching lines, which was what -s was
2320     previously doing.
2322     (p) Added --include and --exclude options to specify files for inclusion
2323     and exclusion when recursing.
2325     11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2326     Hopefully, it now does.
2328     12. Missing cast in pcre_study().
2330     13. Added an "uninstall" target to the makefile.
2332     14. Replaced "extern" in the function prototypes in Makefile.in with
2333     "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2334     world, but is set differently for Windows.
2336     15. Added a second compiling function called pcre_compile2(). The only
2337     difference is that it has an extra argument, which is a pointer to an
2338     integer error code. When there is a compile-time failure, this is set
2339     non-zero, in addition to the error test pointer being set to point to an
2340     error message. The new argument may be NULL if no error number is required
2341     (but then you may as well call pcre_compile(), which is now just a
2342     wrapper). This facility is provided because some applications need a
2343     numeric error indication, but it has also enabled me to tidy up the way
2344     compile-time errors are handled in the POSIX wrapper.
2346     16. Added VPATH=.libs to the makefile; this should help when building with one
2347     prefix path and installing with another. (Or so I'm told by someone who
2348     knows more about this stuff than I do.)
2350     17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2351     passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2352     match everything, including newlines. This is not POSIX-compatible, but
2353     somebody wanted the feature. From pcretest it can be activated by using
2354     both the P and the s flags.
2356     18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2358     19. libpcre.pc was being incorrectly installed as executable.
2360     20. A couple of places in pcretest check for end-of-line by looking for '\n';
2361     it now also looks for '\r' so that it will work unmodified on Windows.
2363     21. Added Google's contributed C++ wrapper to the distribution.
2365     22. Added some untidy missing memory free() calls in pcretest, to keep
2366     Electric Fence happy when testing.
2370 nigel 75 Version 5.0 13-Sep-04
2371     ---------------------
2373     1. Internal change: literal characters are no longer packed up into items
2374     containing multiple characters in a single byte-string. Each character
2375     is now matched using a separate opcode. However, there may be more than one
2376     byte in the character in UTF-8 mode.
2378     2. The pcre_callout_block structure has two new fields: pattern_position and
2379     next_item_length. These contain the offset in the pattern to the next match
2380     item, and its length, respectively.
2382     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2383     insertion of callouts before each pattern item. Added the /C option to
2384     pcretest to make use of this.
2386     4. On the advice of a Windows user, the lines
2388     #if defined(_WIN32) || defined(WIN32)
2389     _setmode( _fileno( stdout ), 0x8000 );
2390     #endif /* defined(_WIN32) || defined(WIN32) */
2392     have been added to the source of pcretest. This apparently does useful
2393     magic in relation to line terminators.
2395     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2396     for the benefit of those environments where the "b" makes a difference.
2398     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2399     to know about it. I have put a hack into configure.in that adds in code
2400     to set GCC=yes if CC=icc. This seems to end up at a point in the
2401     generated configure script that is early enough to affect the setting of
2402     compiler options, which is what is needed, but I have no means of testing
2403     whether it really works. (The user who reported this had patched the
2404     generated configure script, which of course I cannot do.)
2406     LATER: After change 22 below (new libtool files), the configure script
2407     seems to know about icc (and also ecc). Therefore, I have commented out
2408     this hack in configure.in.
2410     7. Added support for pkg-config (2 patches were sent in).
2412     8. Negated POSIX character classes that used a combination of internal tables
2413     were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2414     [[:^ascii]]. Typically, they would match almost any characters. The other
2415     POSIX classes were not broken in this way.
2417     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2418     to find the match, as PCRE was deluded into thinking that the match had to
2419     start at the start point or following a newline. The same bug applied to
2420     patterns with negative forward assertions or any backward assertions
2421     preceding ".*" at the start, unless the pattern required a fixed first
2422     character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2424     10. In UTF-8 mode, when moving forwards in the subject after a failed match
2425     starting at the last subject character, bytes beyond the end of the subject
2426     string were read.
2428     11. Renamed the variable "class" as "classbits" to make life easier for C++
2429     users. (Previously there was a macro definition, but it apparently wasn't
2430     enough.)
2432     12. Added the new field "tables" to the extra data so that tables can be passed
2433     in at exec time, or the internal tables can be re-selected. This allows
2434     a compiled regex to be saved and re-used at a later time by a different
2435     program that might have everything at different addresses.
2437     13. Modified the pcre-config script so that, when run on Solaris, it shows a
2438     -R library as well as a -L library.
2440     14. The debugging options of pcretest (-d on the command line or D on a
2441     pattern) showed incorrect output for anything following an extended class
2442     that contained multibyte characters and which was followed by a quantifier.
2444     15. Added optional support for general category Unicode character properties
2445     via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2446     support. It adds about 90K to the size of the library. The meanings of the
2447     inbuilt class escapes such as \d and \s have NOT been changed.
2449     16. Updated pcredemo.c to include calls to free() to release the memory for the
2450     compiled pattern.
2452     17. The generated file chartables.c was being created in the source directory
2453     instead of in the building directory. This caused the build to fail if the
2454     source directory was different from the building directory, and was
2455     read-only.
2457     18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2458     file. No doubt somebody will tell me if they don't make sense... Also added
2459     Dan Mooney's comments about building on OpenVMS.
2461     19. Added support for partial matching via the PCRE_PARTIAL option for
2462     pcre_exec() and the \P data escape in pcretest.
2464     20. Extended pcretest with 3 new pattern features:
2466     (i) A pattern option of the form ">rest-of-line" causes pcretest to
2467     write the compiled pattern to the file whose name is "rest-of-line".
2468     This is a straight binary dump of the data, with the saved pointer to
2469     the character tables forced to be NULL. The study data, if any, is
2470     written too. After writing, pcretest reads a new pattern.
2472     (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2473     compiled pattern from the given file. There must not be any
2474     occurrences of "<" in the file name (pretty unlikely); if there are,
2475     pcretest will instead treat the initial "<" as a pattern delimiter.
2476     After reading in the pattern, pcretest goes on to read data lines as
2477     usual.
2479     (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2480     and 16-bit fields in a compiled pattern, to simulate a pattern that
2481     was compiled on a host of opposite endianness.
2483     21. The pcre-exec() function can now cope with patterns that were compiled on
2484     hosts of opposite endianness, with this restriction:
2486     As for any compiled expression that is saved and used later, the tables
2487     pointer field cannot be preserved; the extra_data field in the arguments
2488     to pcre_exec() should be used to pass in a tables address if a value
2489     other than the default internal tables were used at compile time.
2491     22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2492     now diagnosed as an error. Previously, most of the time, a negative number
2493     would have been treated as zero, but if in addition "ovector" was passed as
2494     NULL, a crash could occur.
2496     23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2497     new versions from the libtool 1.5 distribution (the last one is a copy of
2498     a file called libtool.m4). This seems to have fixed the need to patch
2499     "configure" to support Darwin 1.3 (which I used to do). However, I still
2500     had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2501     workstation).
2503     24. Changed the PCRE licence to be the more standard "BSD" licence.
2506 nigel 73 Version 4.5 01-Dec-03
2507     ---------------------
2509     1. There has been some re-arrangement of the code for the match() function so
2510     that it can be compiled in a version that does not call itself recursively.
2511     Instead, it keeps those local variables that need separate instances for
2512     each "recursion" in a frame on the heap, and gets/frees frames whenever it
2513     needs to "recurse". Keeping track of where control must go is done by means
2514     of setjmp/longjmp. The whole thing is implemented by a set of macros that
2515     hide most of the details from the main code, and operates only if
2516     NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2517     "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2518     operating.
2520     To make it easier for callers to provide specially tailored get/free
2521     functions for this usage, two new functions, pcre_stack_malloc, and
2522     pcre_stack_free, are used. They are always called in strict stacking order,
2523     and the size of block requested is always the same.
2525     The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2526     PCRE has been compiled to use the stack or the heap for recursion. The
2527     -C option of pcretest uses this to show which version is compiled.
2529     A new data escape \S, is added to pcretest; it causes the amounts of store
2530     obtained and freed by both kinds of malloc/free at match time to be added
2531     to the output.
2533     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2534     what's available on my current Linux desktop machine.
2536     3. When matching a UTF-8 string, the test for a valid string at the start has
2537     been extended. If start_offset is not zero, PCRE now checks that it points
2538     to a byte that is the start of a UTF-8 character. If not, it returns
2539     PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2540     this is necessary because there may be backward assertions in the pattern.
2541     When matching the same subject several times, it may save resources to use
2542     PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2544     4. The code for checking the validity of UTF-8 strings has been tightened so
2545     that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2546     containing "overlong sequences".
2548     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2549     I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2550     should have been "&", but it just so happened that all the cases this let
2551     through by mistake were picked up later in the function.
2553     6. I had used a variable called "isblank" - this is a C99 function, causing
2554     some compilers to warn. To avoid this, I renamed it (as "blankclass").
2556     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2557     prompting; (b) run "./pcretest /dev/null" at the start of the test script
2558     so the version is shown; (c) stop "make test" echoing "./RunTest".
2560     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2562     9. The prototype for memmove() for systems that don't have it was using
2563     size_t, but the inclusion of the header that defines size_t was later. I've
2564     moved the #includes for the C headers earlier to avoid this.
2566     10. Added some adjustments to the code to make it easier to compiler on certain
2567     special systems:
2569     (a) Some "const" qualifiers were missing.
2570     (b) Added the macro EXPORT before all exported functions; by default this
2571     is defined to be empty.
2572     (c) Changed the dftables auxiliary program (that builds chartables.c) so
2573     that it reads its output file name as an argument instead of writing
2574     to the standard output and assuming this can be redirected.
2576     11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2577     class containing characters with values greater than 255, PCRE compilation
2578     went into a loop.
2580     12. A recursive reference to a subpattern that was within another subpattern
2581     that had a minimum quantifier of zero caused PCRE to crash. For example,
2582     (x(y(?2))z)? provoked this bug with a subject that got as far as the
2583     recursion. If the recursively-called subpattern itself had a zero repeat,
2584     that was OK.
2586     13. In pcretest, the buffer for reading a data line was set at 30K, but the
2587     buffer into which it was copied (for escape processing) was still set at
2588     1024, so long lines caused crashes.
2590     14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2591     "internal error: code overflow...". This applied to any character class
2592     that was followed by a possessive quantifier.
2594     15. Modified the Makefile to add libpcre.la as a prerequisite for
2595     libpcreposix.la because I was told this is needed for a parallel build to
2596     work.
2598     16. If a pattern that contained .* following optional items at the start was
2599     studied, the wrong optimizing data was generated, leading to matching
2600     errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2601     matching string must start with a or b or c. The correct conclusion for
2602     this pattern is that a match can start with any character.
2605 nigel 71 Version 4.4 13-Aug-03
2606     ---------------------
2608     1. In UTF-8 mode, a character class containing characters with values between
2609     127 and 255 was not handled correctly if the compiled pattern was studied.
2610     In fixing this, I have also improved the studying algorithm for such
2611     classes (slightly).
2613     2. Three internal functions had redundant arguments passed to them. Removal
2614     might give a very teeny performance improvement.
2616     3. Documentation bug: the value of the capture_top field in a callout is *one
2617     more than* the number of the hightest numbered captured substring.
2619     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2620     in incorrectly linking with a previously installed version. They now link
2621     explicitly with libpcre.la.
2623     5. configure.in no longer needs to recognize Cygwin specially.
2625     6. A problem in pcre.in for Windows platforms is fixed.
2627     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2628     pcretest, it used to include the size of the study block as part of its
2629     output. Unfortunately, the structure contains a field that has a different
2630     size on different hardware architectures. This meant that the tests that
2631     showed this size failed. As the block is currently always of a fixed size,
2632     this information isn't actually particularly useful in pcretest output, so
2633     I have just removed it.
2635     8. Three pre-processor statements accidentally did not start in column 1.
2636     Sadly, there are *still* compilers around that complain, even though
2637     standard C has not required this for well over a decade. Sigh.
2639     9. In pcretest, the code for checking callouts passed small integers in the
2640     callout_data field, which is a void * field. However, some picky compilers
2641     complained about the casts involved for this on 64-bit systems. Now
2642     pcretest passes the address of the small integer instead, which should get
2643     rid of the warnings.
2645     10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2646     both compile and run time, and gives an error if an invalid UTF-8 sequence
2647     is found. There is a option for disabling this check in cases where the
2648     string is known to be correct and/or the maximum performance is wanted.
2650     11. In response to a bug report, I changed one line in Makefile.in from
2652     -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2653     to
2654     -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2656     to look similar to other lines, but I have no way of telling whether this
2657     is the right thing to do, as I do not use Windows. No doubt I'll get told
2658     if it's wrong...
2661 nigel 69 Version 4.3 21-May-03
2662     ---------------------
2664     1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2665     Makefile.
2667     2. Some refactoring to improve the quality of the code:
2669     (i) The utf8_table... variables are now declared "const".
2671     (ii) The code for \cx, which used the "case flipping" table to upper case
2672     lower case letters, now just substracts 32. This is ASCII-specific,
2673     but the whole concept of \cx is ASCII-specific, so it seems
2674     reasonable.
2676     (iii) PCRE was using its character types table to recognize decimal and
2677     hexadecimal digits in the pattern. This is silly, because it handles
2678     only 0-9, a-f, and A-F, but the character types table is locale-
2679     specific, which means strange things might happen. A private
2680     table is now used for this - though it costs 256 bytes, a table is
2681     much faster than multiple explicit tests. Of course, the standard
2682     character types table is still used for matching digits in subject
2683     strings against \d.
2685     (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2686     ending in _t are). So I've renamed it as ESC_tee.
2688     3. The first argument for regexec() in the POSIX wrapper should have been
2689     defined as "const".
2691     4. Changed pcretest to use malloc() for its buffers so that they can be
2692     Electric Fenced for debugging.
2694     5. There were several places in the code where, in UTF-8 mode, PCRE would try
2695     to read one or more bytes before the start of the subject string. Often this
2696     had no effect on PCRE's behaviour, but in some circumstances it could
2697     provoke a segmentation fault.
2699     6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
2700     to try to read one or more bytes before the start of the subject string.
2702     7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
2703     UTF-8 support could misbehave in various ways if the subject string
2704     contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
2705     area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
2706     back over UTF-8 characters.)
2709 nigel 67 Version 4.2 14-Apr-03
2710     ---------------------
2712     1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
2714     2. Changes to the building process, supplied by Ronald Landheer-Cieslak
2715     [ON_WINDOWS]: new variable, "#" on non-Windows platforms
2716     [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
2717     [WIN_PREFIX]: new variable, "cyg" for Cygwin
2718     * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
2719     and BUILD_EXEEXT
2720     Note: automatic setting of the BUILD variables is not yet working
2721     set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
2722     compile-time but not at link-time
2723     [LINK]: use for linking executables only
2724     make different versions for Windows and non-Windows
2725     [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
2726     libraries
2727     [LINK_FOR_BUILD]: new variable
2728     [OBJEXT]: use throughout
2729     [EXEEXT]: use throughout
2730     <winshared>: new target
2731     <wininstall>: new target
2732     <dftables.o>: use native compiler
2733     <dftables>: use native linker
2734     <install>: handle Windows platform correctly
2735     <clean>: ditto
2736     <check>: ditto
2737     copy DLL to top builddir before testing
2739     As part of these changes, -no-undefined was removed again. This was reported
2740     to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
2741     in any case.
2743     3. Some tidies to get rid of compiler warnings:
2745     . In the match_data structure, match_limit was an unsigned long int, whereas
2746     match_call_count was an int. I've made them both unsigned long ints.
2748     . In pcretest the fact that a const uschar * doesn't automatically cast to
2749     a void * provoked a warning.
2751     . Turning on some more compiler warnings threw up some "shadow" variables
2752     and a few more missing casts.
2754     4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
2755     option, a class that contained a single character with a value between 128
2756     and 255 (e.g. /[\xFF]/) caused PCRE to crash.
2758     5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
2759     option, a class that contained several characters, but with at least one
2760     whose value was between 128 and 255 caused PCRE to crash.
2763 nigel 65 Version 4.1 12-Mar-03
2764     ---------------------
2765 nigel 63
2766 nigel 65 1. Compiling with gcc -pedantic found a couple of places where casts were
2767     needed, and a string in dftables.c that was longer than standard compilers are
2768     required to support.
2770     2. Compiling with Sun's compiler found a few more places where the code could
2771     be tidied up in order to avoid warnings.
2773     3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
2774     first of these names is deprecated in the latest Autoconf in favour of the name
2775     CC_FOR_BUILD, because "host" is typically used to mean the system on which the
2776     compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
2777     analogy I have changed it to CFLAGS_FOR_BUILD.
2779     4. Added -no-undefined to the linking command in the Makefile, because this is
2780     apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
2781     linking step for the pcreposix library.
2783     5. PCRE was failing to diagnose the case of two named groups with the same
2784     name.
2786     6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
2787     literal character that is needed in the subject for a match, and scans along to
2788     ensure that it is present before embarking on the full matching process. This
2789     saves time in cases of nested unlimited repeats that are never going to match.
2790     Problem: the scan can take a lot of time if the subject is very long (e.g.
2791     megabytes), thus penalizing straightforward matches. It is now done only if the
2792     amount of subject to be scanned is less than 1000 bytes.
2794     7. A lesser problem with the same optimization is that it was recording the
2795     first character of an anchored pattern as "needed", thus provoking a search
2796     right along the subject, even when the first match of the pattern was going to
2797     fail. The "needed" character is now not set for anchored patterns, unless it
2798     follows something in the pattern that is of non-fixed length. Thus, it still
2799     fulfils its original purpose of finding quick non-matches in cases of nested
2800     unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
2803     Version 4.0 17-Feb-03
2804     ---------------------
2806 nigel 63 1. If a comment in an extended regex that started immediately after a meta-item
2807     extended to the end of string, PCRE compiled incorrect data. This could lead to
2808     all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
2810     2. Moved to autoconf 2.53 and libtool 1.4.2.
2812     3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
2813     the special perltest8 script is no longer needed - all the tests can be run
2814     from a single perltest script.
2816     4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
2817     by \s. It has now been removed in PCRE. This means it isn't recognized as
2818     whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
2819     class [:space:] *does* include VT, thereby creating a mess.
2821     5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
2822     space and tab.
2824     6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
2825     its new features into the main test script, reducing the number of scripts.
2827     7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
2828     were backward compatible, and made the (?i) apply to the whole pattern, as if
2829     /i were given. Now it behaves more logically, and applies the option setting
2830     only to what follows. PCRE has been changed to follow suit. However, if it
2831     finds options settings right at the start of the pattern, it extracts them into
2832     the global options, as before. Thus, they show up in the info data.
2834     8. Added support for the \Q...\E escape sequence. Characters in between are
2835     treated as literals. This is slightly different from Perl in that $ and @ are
2836     also handled as literals inside the quotes. In Perl, they will cause variable
2837     interpolation. Note the following examples:
2839     Pattern PCRE matches Perl matches
2841     \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
2842     \Qabc\$xyz\E abc\$xyz abc\$xyz
2843     \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
2845     For compatibility with Perl, \Q...\E sequences are recognized inside character
2846     classes as well as outside them.
2848     9. Re-organized 3 code statements in pcretest to avoid "overflow in
2849     floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
2850     (size_t) cast to one statement in pcretest and one in pcreposix to avoid
2851     signed/unsigned warnings.
2853     10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
2854     option for pcretest, so I've replaced it by a simple function that does just
2855     that job.
2857     11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
2858     "pcregrep -".
2860     12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
2861     Java package. This provides some syntactic sugar for simple cases of what my
2862     documentation calls "once-only subpatterns". A pattern such as x*+ is the same
2863     as (?>x*). In other words, if what is inside (?>...) is just a single repeated
2864     item, you can use this simplified notation. Note that only makes sense with
2865     greedy quantifiers. Consequently, the use of the possessive quantifier forces
2866     greediness, whatever the setting of the PCRE_UNGREEDY option.
2868     13. A change of greediness default within a pattern was not taking effect at
2869     the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
2870     subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
2871     was abstracted outside.
2873     14. PCRE now supports the \G assertion. It is true when the current matching
2874     position is at the start point of the match. This differs from \A when the
2875     starting offset is non-zero. Used with the /g option of pcretest (or similar
2876     code), it works in the same way as it does for Perl's /g option. If all
2877     alternatives of a regex begin with \G, the expression is anchored to the start
2878     match position, and the "anchored" flag is set in the compiled expression.
2880     15. Some bugs concerning the handling of certain option changes within patterns
2881     have been fixed. These applied to options other than (?ims). For example,
2882     "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
2883     been the other way round. Some of this was related to change 7 above.
2885     16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
2886     features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
2887     and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
2888     POSIX classes only within a class (e.g. /[[:alpha:]]/).
2890     17. Added support for Perl's \C escape. This matches one byte, even in UTF8
2891     mode. Unlike ".", it always matches newline, whatever the setting of
2892     PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
2893     assertions. Perl allows it, but it doesn't (in general) work because it can't
2894     calculate the length of the lookbehind. At least, that's the case for Perl
2895     5.8.0 - I've been told they are going to document that it doesn't work in
2896     future.
2898     18. Added an error diagnosis for escapes that PCRE does not support: these are
2899     \L, \l, \N, \P, \p, \U, \u, and \X.
2901     19. Although correctly diagnosing a missing ']' in a character class, PCRE was
2902     reading past the end of the pattern in cases such as /[abcd/.
2904     20. PCRE was getting more memory than necessary for patterns with classes that
2905     contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
2907     21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
2908     compiling PCRE for use with Virtual Pascal.
2910     22. Small fix to the Makefile to make it work properly if the build is done
2911     outside the source tree.
2913     23. Added a new extension: a condition to go with recursion. If a conditional
2914     subpattern starts with (?(R) the "true" branch is used if recursion has
2915     happened, whereas the "false" branch is used only at the top level.
2917     24. When there was a very long string of literal characters (over 255 bytes
2918     without UTF support, over 250 bytes with UTF support), the computation of how
2919     much memory was required could be incorrect, leading to segfaults or other
2920     strange effects.
2922     25. PCRE was incorrectly assuming anchoring (either to start of subject or to
2923     start of line for a non-DOTALL pattern) when a pattern started with (.*) and
2924     there was a subsequent back reference to those brackets. This meant that, for
2925     example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
2926     possible to check for precisely this case. All we can do is abandon the
2927     optimization if .* occurs inside capturing brackets when there are any back
2928     references whatsoever. (See below for a better fix that came later.)
2930     26. The handling of the optimization for finding the first character of a
2931     non-anchored pattern, and for finding a character that is required later in the
2932     match were failing in some cases. This didn't break the matching; it just
2933     failed to optimize when it could. The way this is done has been re-implemented.
2935     27. Fixed typo in error message for invalid (?R item (it said "(?p").
2937     28. Added a new feature that provides some of the functionality that Perl
2938     provides with (?{...}). The facility is termed a "callout". The way it is done
2939     in PCRE is for the caller to provide an optional function, by setting
2940     pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
2941     global variable. By default it is unset, which disables all calling out. To get
2942     the function called, the regex must include (?C) at appropriate points. This
2943     is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
2944     This provides a means of identifying different callout points. When PCRE
2945     reaches such a point in the regex, if pcre_callout has been set, the external
2946     function is called. It is provided with data in a structure called
2947     pcre_callout_block, which is defined in pcre.h. If the function returns 0,
2948     matching continues; if it returns a non-zero value, the match at the current
2949     point fails. However, backtracking will occur if possible. [This was changed
2950     later and other features added - see item 49 below.]
2952     29. pcretest is upgraded to test the callout functionality. It provides a
2953     callout function that displays information. By default, it shows the start of
2954     the match and the current position in the text. There are some new data escapes
2955     to vary what happens:
2957     \C+ in addition, show current contents of captured substrings
2958     \C- do not supply a callout function
2959     \C!n return 1 when callout number n is reached
2960     \C!n!m return 1 when callout number n is reached for the mth time
2962     30. If pcregrep was called with the -l option and just a single file name, it
2963     output "<stdin>" if a match was found, instead of the file name.
2965     31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
2966     slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
2967     pcre_exec(). This saves a malloc/free per call. The default value of
2968     POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
2969     when configuring.
2971     32. The default maximum size of a compiled pattern is 64K. There have been a
2972     few cases of people hitting this limit. The code now uses macros to handle the
2973     storing of links as offsets within the compiled pattern. It defaults to 2-byte
2974     links, but this can be changed to 3 or 4 bytes by --with-link-size when
2975     configuring. Tests 2 and 5 work only with 2-byte links because they output
2976     debugging information about compiled patterns.
2978     33. Internal code re-arrangements:
2980     (a) Moved the debugging function for printing out a compiled regex into
2981     its own source file (printint.c) and used #include to pull it into
2982     pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
2983     separate copies.
2985     (b) Defined the list of op-code names for debugging as a macro in
2986     internal.h so that it is next to the definition of the opcodes.
2988     (c) Defined a table of op-code lengths for simpler skipping along compiled
2989     code. This is again a macro in internal.h so that it is next to the
2990     definition of the opcodes.
2992     34. Added support for recursive calls to individual subpatterns, along the
2993     lines of Robin Houston's patch (but implemented somewhat differently).
2995     35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
2996     allow it to read and process whole directories in Win32. This code was
2997     contributed by Lionel Fourquaux; it has not been tested by me.
2999     36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3000     used to name a group. Names consist of alphanumerics and underscores, and must
3001     be unique. Back references use the syntax (?P=name) and recursive calls use
3002     (?P>name) which is a PCRE extension to the Python extension. Groups still have
3003     numbers. The function pcre_fullinfo() can be used after compilation to extract
3004     a name/number map. There are three relevant calls:
3006     PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3007     PCRE_INFO_NAMECOUNT yields the number of entries
3008     PCRE_INFO_NAMETABLE yields a pointer to the map.
3010     The map is a vector of fixed-size entries. The size of each entry depends on
3011     the length of the longest name used. The first two bytes of each entry are the
3012     group number, most significant byte first. There follows the corresponding
3013     name, zero terminated. The names are in alphabetical order.
3015     37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3016     case instead of 255. Making it the same both with and without UTF-8 support
3017     means that the same test output works with both.
3019     38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3020     calling malloc() with a zero argument.
3022     39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3023     optimization. I've improved things by keeping a bitmap of backreferences with
3024     numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3025     fact referenced, the optimization can be applied. It is unlikely that a
3026     relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3027     the match to follow \n) will appear inside brackets with a number greater than
3028     31, but if it does, any back reference > 31 suppresses the optimization.
3030     40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3031     of disabling numbered capturing parentheses. Any opening parenthesis that is
3032     not followed by ? behaves as if it were followed by ?: but named parentheses
3033     can still be used for capturing (and they will acquire numbers in the usual
3034     way).
3036     41. Redesigned the return codes from the match() function into yes/no/error so
3037     that errors can be passed back from deep inside the nested calls. A malloc
3038     failure while inside a recursive subpattern call now causes the
3039     PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3041     42. It is now possible to set a limit on the number of times the match()
3042     function is called in a call to pcre_exec(). This facility makes it possible to
3043     limit the amount of recursion and backtracking, though not in a directly
3044     obvious way, because the match() function is used in a number of different
3045     circumstances. The count starts from zero for each position in the subject
3046     string (for non-anchored patterns). The default limit is, for compatibility, a
3047     large number, namely 10 000 000. You can change this in two ways:
3049     (a) When configuring PCRE before making, you can use --with-match-limit=n
3050     to set a default value for the compiled library.
3052     (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3053     a different value is set. See 45 below.
3055     If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3057     43. Added a new function pcre_config(int, void *) to enable run-time extraction
3058     of things that can be changed at compile time. The first argument specifies
3059     what is wanted and the second points to where the information is to be placed.
3060     The current list of available information is:
3064     The output is an integer that is set to one if UTF-8 support is available;
3065     otherwise it is set to zero.
3069     The output is an integer that it set to the value of the code that is used for
3070     newline. It is either LF (10) or CR (13).
3074     The output is an integer that contains the number of bytes used for internal
3075     linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3079     The output is an integer that contains the threshold above which the POSIX
3080     interface uses malloc() for output vectors. See item 31 above.
3084     The output is an unsigned integer that contains the default limit of the number
3085     of match() calls in a pcre_exec() execution. See 42 above.
3087     44. pcretest has been upgraded by the addition of the -C option. This causes it
3088     to extract all the available output from the new pcre_config() function, and to
3089     output it. The program then exits immediately.
3091     45. A need has arisen to pass over additional data with calls to pcre_exec() in
3092     order to support additional features. One way would have been to define
3093     pcre_exec2() (for example) with extra arguments, but this would not have been
3094     extensible, and would also have required all calls to the original function to
3095     be mapped to the new one. Instead, I have chosen to extend the mechanism that
3096     is used for passing in "extra" data from pcre_study().
3098     The pcre_extra structure is now exposed and defined in pcre.h. It currently
3099     contains the following fields:
3101     flags a bitmap indicating which of the following fields are set
3102     study_data opaque data from pcre_study()
3103     match_limit a way of specifying a limit on match() calls for a specific
3104     call to pcre_exec()
3105     callout_data data for callouts (see 49 below)
3107     The flag bits are also defined in pcre.h, and are
3113     The pcre_study() function now returns one of these new pcre_extra blocks, with
3114     the actual study data pointed to by the study_data field, and the
3115     PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3116     before. That is, this change is entirely upwards-compatible and requires no
3117     change to existing code.
3119     If you want to pass in additional data to pcre_exec(), you can either place it
3120     in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3121     block.
3123     46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3124     data string contains the escape sequence \M, pcretest calls pcre_exec() several
3125     times with different match limits, until it finds the minimum value needed for
3126     pcre_exec() to complete. The value is then output. This can be instructive; for
3127     most simple matches the number is quite small, but for pathological cases it
3128     gets very large very quickly.
3130     47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3131     returns the size of the data block pointed to by the study_data field in a
3132     pcre_extra block, that is, the value that was passed as the argument to
3133     pcre_malloc() when PCRE was getting memory in which to place the information
3134     created by pcre_study(). The fourth argument should point to a size_t variable.
3135     pcretest has been extended so that this information is shown after a successful
3136     pcre_study() call when information about the compiled regex is being displayed.
3138     48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3139     because what follows is always an absolute path. (Later: it turns out that this
3140     is more than cosmetic for MinGW, because it doesn't like empty path
3141     components.)
3143     49. Some changes have been made to the callout feature (see 28 above):
3145     (i) A callout function now has three choices for what it returns:
3147     0 => success, carry on matching
3148     > 0 => failure at this point, but backtrack if possible
3149     < 0 => serious error, return this value from pcre_exec()
3151     Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3152     values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3153     "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3154     use by callout functions. It will never be used by PCRE itself.
3156     (ii) The pcre_extra structure (see 45 above) has a void * field called
3157     callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3158     pcre_callout_block structure has a field of the same name. The contents of
3159     the field passed in the pcre_extra structure are passed to the callout
3160     function in the corresponding field in the callout block. This makes it
3161     easier to use the same callout-containing regex from multiple threads. For
3162     testing, the pcretest program has a new data escape
3164     \C*n pass the number n (may be negative) as callout_data
3166     If the callout function in pcretest receives a non-zero value as
3167     callout_data, it returns that value.
3169     50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3170     there were some redundant $(CFLAGS) in commands that are now specified as
3171     $(LINK), which already includes $(CFLAGS).
3173     51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3174     has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3175     with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3176     one-byte characters throughout. Note that case-insensitive matching applies
3177     only to characters whose values are less than 256. PCRE doesn't support the
3178     notion of cases for higher-valued characters.
3180     (i) A character class whose characters are all within 0-255 is handled as
3181     a bit map, and the map is inverted for negative classes. Previously, a
3182     character > 255 always failed to match such a class; however it should
3183     match if the class was a negative one (e.g. [^ab]). This has been fixed.
3185     (ii) A negated character class with a single character < 255 is coded as
3186     "not this character" (OP_NOT). This wasn't working properly when the test
3187     character was multibyte, either singly or repeated.
3189     (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3190     mode, for example: \x{100}{2,3}.
3192     (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3193     singly or repeated) now correctly test multibyte characters. However,
3194     PCRE doesn't recognize any characters with values greater than 255 as
3195     digits, spaces, or word characters. Such characters always match \D, \S,
3196     and \W, and never match \d, \s, or \w.
3198     (v) Classes may now contain characters and character ranges with values
3199     greater than 255. For example: [ab\x{100}-\x{400}].
3201     (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3202     PCRE in UTF-8 mode.
3204     52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3205     PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3206     retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3207     value.)
3209     53. The single man page has become too large. I have therefore split it up into
3210     a number of separate man pages. These also give rise to individual HTML pages;
3211     these are now put in a separate directory, and there is an index.html page that
3212     lists them all. Some hyperlinking between the pages has been installed.
3214     54. Added convenience functions for handling named capturing parentheses.
3216     55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3217     aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3218     true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3219     are faulted.
3221     56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3222     calling configure. These values are used when compiling the dftables.c program
3223     which is run to generate the source of the default character tables. They
3224     default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3225     you will need to set these values.
3227     57. Updated the building process for Windows DLL, as provided by Fred Cox.
3230     Version 3.9 02-Jan-02
3231 nigel 61 ---------------------
3233     1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3235     2. If --disable-static was given, the building process failed when trying to
3236     build pcretest and pcregrep. (For some reason it was using libtool to compile
3237     them, which is not right, as they aren't part of the library.)
3240 nigel 59 Version 3.8 18-Dec-01
3241     ---------------------
3243     1. The experimental UTF-8 code was completely screwed up. It was packing the
3244     bytes in the wrong order. How dumb can you get?
3247 nigel 57 Version 3.7 29-Oct-01
3248     ---------------------
3250     1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3251     This caused pcretest, when used on the test data, to segfault. Unfortunately,
3252     this didn't happen under Solaris 8, where I normally test things.
3254 nigel 59 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3255     doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3256     isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3257     this fix an hour or so after the initial 3.7 release.)
3258 nigel 57
3259 nigel 59
3260 nigel 55 Version 3.6 23-Oct-01
3261     ---------------------
3263     1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3264     offsets passed as NULL with zero offset count.
3266     2. The config.guess and config.sub files had not been updated when I moved to
3267     the latest autoconf.
3270 nigel 53 Version 3.5 15-Aug-01
3271     ---------------------
3272 nigel 5
3273 nigel 53 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3274     had been forgotten.
3276     2. By using declared but undefined structures, we can avoid using "void"
3277     definitions in pcre.h while keeping the internal definitions of the structures
3278     private.
3280     3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3281     user point of view, this means that both static and shared libraries are built
3282     by default, but this can be individually controlled. More of the work of
3283     handling this static/shared cases is now inside libtool instead of PCRE's make
3284     file.
3286     4. The pcretest utility is now installed along with pcregrep because it is
3287     useful for users (to test regexs) and by doing this, it automatically gets
3288     relinked by libtool. The documentation has been turned into a man page, so
3289     there are now .1, .txt, and .html versions in /doc.
3291     5. Upgrades to pcregrep:
3292     (i) Added long-form option names like gnu grep.
3293     (ii) Added --help to list all options with an explanatory phrase.
3294     (iii) Added -r, --recursive to recurse into sub-directories.
3295     (iv) Added -f, --file to read patterns from a file.
3297     6. pcre_exec() was referring to its "code" argument before testing that
3298     argument for NULL (and giving an error if it was NULL).
3300     7. Upgraded Makefile.in to allow for compiling in a different directory from
3301     the source directory.
3303     8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3304     options bits, the pointer it was passed was to an int instead of to an unsigned
3305     long int. This mattered only on 64-bit systems.
3307     9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3308     generated) instead of pcre.in, which it its source. Also made the same change
3309     in several of the .c files.
3311     10. A new release of gcc defines printf() as a macro, which broke pcretest
3312     because it had an ifdef in the middle of a string argument for printf(). Fixed
3313     by using separate calls to printf().
3315     11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3316     script, to force use of CR or LF instead of \n in the source. On non-Unix
3317     systems, the value can be set in config.h.
3319     12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3320     absolute limit. Changed the text of the error message to make this clear, and
3321     likewise updated the man page.
3323     13. The limit of 99 on the number of capturing subpatterns has been removed.
3324     The new limit is 65535, which I hope will not be a "real" limit.
3327 nigel 51 Version 3.4 22-Aug-00
3328     ---------------------
3330     1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3332     2. Diagnose condition (?(0) as an error instead of crashing on matching.
3335 nigel 49 Version 3.3 01-Aug-00
3336     ---------------------
3338     1. If an octal character was given, but the value was greater than \377, it
3339     was not getting masked to the least significant bits, as documented. This could
3340     lead to crashes in some systems.
3342     2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3343     the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3345     3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3346     These just pass their arguments on to (pcre_free)(), but they are provided
3347     because some uses of PCRE bind it to non-C systems that can call its functions,
3348     but cannot call free() or pcre_free() directly.
3350     4. Add "make test" as a synonym for "make check". Corrected some comments in
3351     the Makefile.
3353     5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3354     Makefile.
3356     6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3357     command called pgrep for grepping around the active processes.
3359     7. Added the beginnings of support for UTF-8 character strings.
3361     8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3362     RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3363     the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3364     out for the ar command.)
3367 nigel 47 Version 3.2 12-May-00
3368     ---------------------
3370     This is purely a bug fixing release.
3372     1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3373     of ZA. This was just one example of several cases that could provoke this bug,
3374     which was introduced by change 9 of version 2.00. The code for breaking
3375     infinite loops after an iteration that matches an empty string was't working
3376     correctly.
3378     2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3379     when matched against abbab (for example). After matching an empty string, it
3380     wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3381     caused it to match further down the string than it should.
3383     3. The code contained an inclusion of sys/types.h. It isn't clear why this
3384     was there because it doesn't seem to be needed, and it causes trouble on some
3385     systems, as it is not a Standard C header. It has been removed.
3387     4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3388     were reported on the Macintosh. The changes were from
3390     while ((c = *(++ptr)) != 0 && c != '\n');
3391     to
3392     while ((c = *(++ptr)) != 0 && c != '\n') ;
3394     Totally extraordinary, but if that's what it takes...
3396     5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3397     available. Added HAVE_BCOPY and an autoconf test for it; if neither
3398     HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3399     assumes the way PCRE uses memmove() (always moving upwards).
3401     6. PCRE is being used in one environment where strchr() is not available. There
3402     was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3403     faster code anyway.
3406 nigel 45 Version 3.1 09-Feb-00
3407     ---------------------
3409     The only change in this release is the fixing of some bugs in Makefile.in for
3410     the "install" target:
3412     (1) It was failing to install pcreposix.h.
3414     (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3417 nigel 43 Version 3.0 01-Feb-00
3418     ---------------------
3419 nigel 41
3420     1. Add support for the /+ modifier to perltest (to output $` like it does in
3421     pcretest).
3423     2. Add support for the /g modifier to perltest.
3425     3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3426     matches null strings.
3428     4. Fix perltest so that it doesn't do unwanted things when fed an empty
3429     pattern. Perl treats empty patterns specially - it reuses the most recent
3430     pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3431     effect.
3433     5. The POSIX interface was broken in that it was just handing over the POSIX
3434     captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3435     required a bigger vector, with some working space on the end. This means that
3436     the POSIX wrapper now has to get and free some memory, and copy the results.
3438 nigel 43 6. Added some simple autoconf support, placing the test data and the
3439     documentation in separate directories, re-organizing some of the
3440     information files, and making it build pcre-config (a GNU standard). Also added
3441     libtool support for building PCRE as a shared library, which is now the
3442     default.
3443 nigel 41
3444 nigel 43 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
3445     09 are not valid octal constants. Single digits will be used for minor values
3446     less than 10.
3448     8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
3449     existing programs that set these in the POSIX interface can use PCRE without
3450     modification.
3452     9. Added a new function, pcre_fullinfo() with an extensible interface. It can
3453     return all that pcre_info() returns, plus additional data. The pcre_info()
3454     function is retained for compatibility, but is considered to be obsolete.
3456     10. Added experimental recursion feature (?R) to handle one common case that
3457     Perl 5.6 will be able to do with (?p{...}).
3459     11. Added support for POSIX character classes like [:alpha:], which Perl is
3460     adopting.
3463 nigel 39 Version 2.08 31-Aug-99
3464     ----------------------
3466     1. When startoffset was not zero and the pattern began with ".*", PCRE was not
3467     trying to match at the startoffset position, but instead was moving forward to
3468     the next newline as if a previous match had failed.
3470     2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
3471     and could get into a loop if a null string was matched other than at the start
3472     of the subject.
3474     3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
3475     be distinguished at compile time, and for completeness also added PCRE_DATE.
3477     5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
3478     in GnuWin32 environments.
3481 nigel 37 Version 2.07 29-Jul-99
3482     ----------------------
3484     1. The documentation is now supplied in plain text form and HTML as well as in
3485     the form of man page sources.
3487     2. C++ compilers don't like assigning (void *) values to other pointer types.
3488     In particular this affects malloc(). Although there is no problem in Standard
3489     C, I've put in casts to keep C++ compilers happy.
3491     3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
3492     should be (const char *).
3494     4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
3495     be useful for non-Unix systems who don't want to bother with the POSIX stuff.
3496     However, I haven't made this a standard facility. The documentation doesn't
3497     mention it, and the Makefile doesn't support it.
3499     5. The Makefile now contains an "install" target, with editable destinations at
3500     the top of the file. The pcretest program is not installed.
3502     6. pgrep -V now gives the PCRE version number and date.
3504     7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
3505     causing the entire string to be ignored, instead of just the last character.
3507     8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
3508     non-matching string, it can take a very, very long time, even for strings of
3509     quite modest length, because of the nested recursion. PCRE now does better in
3510     some of these cases. It does this by remembering the last required literal
3511     character in the pattern, and pre-searching the subject to ensure it is present
3512     before running the real match. In other words, it applies a heuristic to detect
3513     some types of certain failure quickly, and in the above example, if presented
3514     with a string that has no trailing " it gives "no match" very quickly.
3516     9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
3517     other alternatives are tried instead.
3520 nigel 35 Version 2.06 09-Jun-99
3521     ----------------------
3523     1. Change pcretest's output for amount of store used to show just the code
3524     space, because the remainder (the data block) varies in size between 32-bit and
3525     64-bit systems.
3527     2. Added an extra argument to pcre_exec() to supply an offset in the subject to
3528     start matching at. This allows lookbehinds to work when searching for multiple
3529     occurrences in a string.
3531     3. Added additional options to pcretest for testing multiple occurrences:
3533     /+ outputs the rest of the string that follows a match
3534     /g loops for multiple occurrences, using the new startoffset argument
3535     /G loops for multiple occurrences by passing an incremented pointer
3537     4. PCRE wasn't doing the "first character" optimization for patterns starting
3538     with \b or \B, though it was doing it for other lookbehind assertions. That is,
3539     it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
3540     the letter 'x'. On long subject strings, this gives a significant speed-up.
3543 nigel 33 Version 2.05 21-Apr-99
3544     ----------------------
3546     1. Changed the type of magic_number from int to long int so that it works
3547     properly on 16-bit systems.
3549     2. Fixed a bug which caused patterns starting with .* not to work correctly
3550     when the subject string contained newline characters. PCRE was assuming
3551     anchoring for such patterns in all cases, which is not correct because .* will
3552     not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
3553     DOTALL is set at top level; otherwise it knows that patterns starting with .*
3554     must be retried after every newline in the subject.
3557 nigel 31 Version 2.04 18-Feb-99
3558     ----------------------
3560     1. For parenthesized subpatterns with repeats whose minimum was zero, the
3561     computation of the store needed to hold the pattern was incorrect (too large).
3562     If such patterns were nested a few deep, this could multiply and become a real
3563     problem.
3565     2. Added /M option to pcretest to show the memory requirement of a specific
3566     pattern. Made -m a synonym of -s (which does this globally) for compatibility.
3568     3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
3569     compiled in such a way that the backtracking after subsequent failure was
3570     pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
3571     ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
3574 nigel 29 Version 2.03 02-Feb-99
3575     ----------------------
3577     1. Fixed typo and small mistake in man page.
3579 nigel 31 2. Added 4th condition (GPL supersedes if conflict) and created separate
3580     LICENCE file containing the conditions.
3581 nigel 29
3582     3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
3583     Perl, that is the internal \ allows the delimiter to be included in the
3584     pattern. Locked out the use of \ as a delimiter. If \ immediately follows
3585     the final delimiter, add \ to the end of the pattern (to test the error).
3587     4. Added the convenience functions for extracting substrings after a successful
3588     match. Updated pcretest to make it able to test these functions.
3591 nigel 27 Version 2.02 14-Jan-99
3592     ----------------------
3594     1. Initialized the working variables associated with each extraction so that
3595     their saving and restoring doesn't refer to uninitialized store.
3597     2. Put dummy code into study.c in order to trick the optimizer of the IBM C
3598     compiler for OS/2 into generating correct code. Apparently IBM isn't going to
3599     fix the problem.
3601     3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
3602     calls, and wasn't printing the correct value for compiling calls. Increased the
3603     default value of LOOPREPEAT, and the number of significant figures in the
3604     times.
3606     4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
3608     5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
3609     a building problem on Windows NT with a FAT file system.
3612 nigel 25 Version 2.01 21-Oct-98
3613     ----------------------
3615     1. Changed the API for pcre_compile() to allow for the provision of a pointer
3616     to character tables built by pcre_maketables() in the current locale. If NULL
3617     is passed, the default tables are used.
3620 nigel 23 Version 2.00 24-Sep-98
3621 nigel 21 ----------------------
3623 nigel 23 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
3624     it any more.
3626     2. Allow quantification of (?>) groups, and make it work correctly.
3628     3. The first character computation wasn't working for (?>) groups.
3630     4. Correct the implementation of \Z (it is permitted to match on the \n at the
3631     end of the subject) and add 5.005's \z, which really does match only at the
3632     very end of the subject.
3634     5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
3636     6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
3637     DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
3638     localized options. All options to pcre_study() were also removed.
3640     7. Add other new features from 5.005:
3642     $(?<= positive lookbehind
3643     $(?<! negative lookbehind
3644     (?imsx-imsx) added the unsetting capability
3645     such a setting is global if at outer level; local otherwise
3646     (?imsx-imsx:) non-capturing groups with option setting
3647     (?(cond)re|re) conditional pattern matching
3649     A backreference to itself in a repeated group matches the previous
3650     captured string.
3652     8. General tidying up of studying (both automatic and via "study")
3653     consequential on the addition of new assertions.
3655     9. As in 5.005, unlimited repeated groups that could match an empty substring
3656     are no longer faulted at compile time. Instead, the loop is forcibly broken at
3657     runtime if any iteration does actually match an empty substring.
3659     10. Include the RunTest script in the distribution.
3661     11. Added tests from the Perl 5.005_02 distribution. This showed up a few
3662     discrepancies, some of which were old and were also with respect to 5.004. They
3663     have now been fixed.
3666     Version 1.09 28-Apr-98
3667     ----------------------
3669 nigel 21 1. A negated single character class followed by a quantifier with a minimum
3670     value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
3671     program crashes, or just wrong answers. This did not apply to negated classes
3672     containing more than one character, or to minima other than one.
3675 nigel 19 Version 1.08 27-Mar-98
3676     ----------------------
3678     1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
3680     2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
3681     latter must appear before anything that relies on it in the pattern.
3684 nigel 17 Version 1.07 16-Feb-98
3685     ----------------------
3687     1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
3688     repeat of a potentially empty string).
3691 nigel 15 Version 1.06 23-Jan-98
3692     ----------------------
3694     1. Added Markus Oberhumer's little patches for C++.
3696     2. Literal strings longer than 255 characters were broken.
3699 nigel 13 Version 1.05 23-Dec-97
3700     ----------------------
3702     1. Negated character classes containing more than one character were failing if
3703     PCRE_CASELESS was set at run time.
3706 nigel 11 Version 1.04 19-Dec-97
3707     ----------------------