ViewVC logotype

Diff of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 57 by nigel, Sat Feb 24 21:39:50 2007 UTC revision 629 by ph10, Fri Jul 22 09:18:11 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
4    Version 8.13 30-Apr-2011
5    ------------------------
7    1.  The Unicode data tables have been updated to Unicode 6.0.0.
9    2.  Two minor typos in pcre_internal.h have been fixed.
11    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
12        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
13        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
15    4.  There were a number of related bugs in the code for matching backrefences
16        caselessly in UTF-8 mode when codes for the characters concerned were
17        different numbers of bytes. For example, U+023A and U+2C65 are an upper
18        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
19        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
20        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
21        2-byte code at the end of the subject (it thought there wasn't enough data
22        left).
24    5.  Comprehensive information about what went wrong is now returned by
25        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
26        as the output vector has at least 2 elements. The offset of the start of
27        the failing character and a reason code are placed in the vector.
29    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
30        now returned is for the first byte of the failing character, instead of the
31        last byte inspected. This is an incompatible change, but I hope it is small
32        enough not to be a problem. It makes the returned offset consistent with
33        pcre_exec() and pcre_dfa_exec().
35    7.  pcretest now gives a text phrase as well as the error number when
36        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
37        failure, the offset and reason code are output.
39    8.  When \R was used with a maximizing quantifier it failed to skip backwards
40        over a \r\n pair if the subsequent match failed. Instead, it just skipped
41        back over a single character (\n). This seems wrong (because it treated the
42        two characters as a single entity when going forwards), conflicts with the
43        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
44        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
45        has been changed.
47    9.  Some internal refactoring has changed the processing so that the handling
48        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
49        time (the PCRE_DOTALL option was changed this way some time ago: version
50        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
51        which was always a bit of a fudge. It also means that there is one less
52        argument for the match() function, which reduces its stack requirements
53        slightly. This change also fixes an incompatibility with Perl: the pattern
54        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
56    10. More internal refactoring has drastically reduced the number of recursive
57        calls to match() for possessively repeated groups such as (abc)++ when
58        using pcre_exec().
60    11. While implementing 10, a number of bugs in the handling of groups were
61        discovered and fixed:
63        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
64        (a|)*(?1) gave a compile-time internal error.
65        ((a|)+)+  did not notice that the outer group could match an empty string.
66        (^a|^)+   was not marked as anchored.
67        (.*a|.*)+ was not marked as matching at start or after a newline.
69    12. Yet more internal refactoring has removed another argument from the match()
70        function. Special calls to this function are now indicated by setting a
71        value in a variable in the "match data" data block.
73    13. Be more explicit in pcre_study() instead of relying on "default" for
74        opcodes that mean there is no starting character; this means that when new
75        ones are added and accidentally left out of pcre_study(), testing should
76        pick them up.
78    14. The -s option of pcretest has been documented for ages as being an old
79        synonym of -m (show memory usage). I have changed it to mean "force study
80        for every regex", that is, assume /S for every regex. This is similar to -i
81        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
82        using it. It makes it easier to run collections of tests with and without
83        study enabled, and thereby test pcre_study() more easily. All the standard
84        tests are now run with and without -s (but some patterns can be marked as
85        "never study" - see 20 below).
87    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
88        restoration of the capturing data to the outer values was not happening
89        correctly.
91    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
92        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
93        pattern had matched an empty string, and so incorrectly returned a no
94        match.
96    17. There was optimizing code for the last branch of non-capturing parentheses,
97        and also for the obeyed branch of a conditional subexpression, which used
98        tail recursion to cut down on stack usage. Unfortunately, not that there is
99        the possibility of (*THEN) occurring in these branches, tail recursion is
100        no longer possible because the return has to be checked for (*THEN). These
101        two optimizations have therefore been removed.
103    18. If a pattern containing \R was studied, it was assumed that \R always
104        matched two bytes, thus causing the minimum subject length to be
105        incorrectly computed because \R can also match just one byte.
107    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
108        was incorrectly computed.
110    20. If /S is present twice on a test pattern in pcretest input, it *disables*
111        studying, thereby overriding the use of -s on the command line. This is
112        necessary for one or two tests to keep the output identical in both cases.
114    21. When (*ACCEPT) was used in an assertion that matched an empty string and
115        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
117    22. When an atomic group that contained a capturing parenthesis was
118        successfully matched, but the branch in which it appeared failed, the
119        capturing was not being forgotten if a higher numbered group was later
120        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
121        group 1 to "a", when in fact it should be unset. This applied to multi-
122        branched capturing and non-capturing groups, repeated or not, and also to
123        positive assertions (capturing in negative assertions is not well defined
124        in PCRE) and also to nested atomic groups.
126    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
127        subject after a captured substring (to make it easier to tell which of a
128        number of identical substrings has been captured).
130    24. The way atomic groups are processed by pcre_exec() has been changed so that
131        if they are repeated, backtracking one repetition now resets captured
132        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
133        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
134        "aaa". Previously, it would have been "a". As part of this code
135        refactoring, the way recursive calls are handled has also been changed.
137    24. If an assertion condition captured any substrings, they were not passed
138        back unless some other capturing happened later. For example, if
139        (?(?=(a))a) was matched against "a", no capturing was returned.
141    25. When studying a pattern that contained subroutine calls or assertions,
142        the code for finding the minimum length of a possible match was handling
143        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
144        group 1 called group 2 while simultaneously a separate group 2 called group
145        1). A stack overflow occurred in this case. I have fixed this by limiting
146        the recursion depth to 10.
148    26. Updated RunTest.bat in the distribution to the version supplied by Tom
149        Fortmann. This supports explicit test numbers on the command line, and has
150        argument validation and error reporting.
152    27. An instance of \X with an unlimited repeat could fail if at any point the
153        first character it looked at was a mark character.
155    28. Some minor code refactoring concerning Unicode properties and scripts
156        should reduce the stack requirement of match() slightly.
158    29. Added the '=' option to pcretest to check the setting of unused capturing
159        slots at the end of the pattern, which are documented as being -1, but are
160        not included in the return count.
162    30. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
163        compiled something random. Now it gives a compile-time error (as does
164        Perl).
167    Version 8.12 15-Jan-2011
168    ------------------------
170    1.  Fixed some typos in the markup of the man pages, and wrote a script that
171        checks for such things as part of the documentation building process.
173    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
174        --match-limit and --recursion-limit options (added for 8.11). In
175        particular, this made one of the standard tests fail. (The integer value
176        went into the wrong half of a long int.)
178    3.  If the --colour option was given to pcregrep with -v (invert match), it
179        did strange things, either producing crazy output, or crashing. It should,
180        of course, ignore a request for colour when reporting lines that do not
181        match.
183    4.  Another pcregrep bug caused similar problems if --colour was specified with
184        -M (multiline) and the pattern match finished with a line ending.
186    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
187        matched in multiline mode, the following line was shown as part of the
188        match. This seems wrong, so I have changed it.
190    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
191        the check for further matches in the same line (so they could be coloured)
192        to overrun the end of the current line. If another match was found, it was
193        incorrectly shown (and then shown again when found in the next line).
195    7.  If pcregrep was compiled under Windows, there was a reference to the
196        function pcregrep_exit() before it was defined. I am assuming this was
197        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
198        reported by a user. I've moved the definition above the reference.
201    Version 8.11 10-Dec-2010
202    ------------------------
204    1.  (*THEN) was not working properly if there were untried alternatives prior
205        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
206        backtracked to try for "b" instead of moving to the next alternative branch
207        at the same level (in this case, to look for "c"). The Perl documentation
208        is clear that when (*THEN) is backtracked onto, it goes to the "next
209        alternative in the innermost enclosing group".
211    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
212        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
213        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
214        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
215        (*THEN).
217    3.  If \s appeared in a character class, it removed the VT character from
218        the class, even if it had been included by some previous item, for example
219        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
220        of \s, but is part of the POSIX "space" class.)
222    4.  A partial match never returns an empty string (because you can always
223        match an empty string at the end of the subject); however the checking for
224        an empty string was starting at the "start of match" point. This has been
225        changed to the "earliest inspected character" point, because the returned
226        data for a partial match starts at this character. This means that, for
227        example, /(?<=abc)def/ gives a partial match for the subject "abc"
228        (previously it gave "no match").
230    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
231        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
232        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
233        has an implication that the given string is incomplete (because a partial
234        match is preferred over a full match). For this reason, these items now
235        give a partial match in this situation. [Aside: previously, the one case
236        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
237        match rather than a full match, which was wrong by the old rules, but is
238        now correct.]
240    6.  There was a bug in the handling of #-introduced comments, recognized when
241        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
242        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
243        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
244        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
245        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
246        places in pcre_compile().
248    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
249        comments when looking ahead for named forward references to subpatterns,
250        the only newline sequence it recognized was NL. It now handles newlines
251        according to the set newline convention.
253    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
254        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
255        cater for a lack of strerror(). These oversights have been fixed.
257    9.  Added --match-limit and --recursion-limit to pcregrep.
259    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
261    11. When the -o option was used, pcregrep was setting a return code of 1, even
262        when matches were found, and --line-buffered was not being honoured.
264    12. Added an optional parentheses number to the -o and --only-matching options
265        of pcregrep.
267    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
268        can match an empty string. The code to do it in pcretest and pcredemo
269        needed fixing:
271        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
272            only one byte after an empty string match just before CRLF (this case
273            just got forgotten; "any" and "anycrlf" were OK).
275        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
276            mode when an empty string match preceded an ASCII character followed by
277            a non-ASCII character. (The code for advancing by one character rather
278            than one byte was nonsense.)
280        (c) The pcredemo.c sample program did not have any code at all to handle
281            the cases when CRLF is a valid newline sequence.
283    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
284        as a starting offset was within the subject string. There is now a new
285        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
286        negative or greater than the length of the string. In order to test this,
287        pcretest is extended to allow the setting of negative starting offsets.
289    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
290        starting offset points to the beginning of a UTF-8 character was
291        unnecessarily clumsy. I tidied it up.
293    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
294        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
296    17. Nobody had reported that the --include_dir option, which was added in
297        release 7.7 should have been called --include-dir (hyphen, not underscore)
298        for compatibility with GNU grep. I have changed it to --include-dir, but
299        left --include_dir as an undocumented synonym, and the same for
300        --exclude-dir, though that is not available in GNU grep, at least as of
301        release 2.5.4.
303    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
304        characters from a string of bytes) have been redefined so as not to use
305        loops, in order to improve performance in some environments. At the same
306        time, I abstracted some of the common code into auxiliary macros to save
307        repetition (this should not affect the compiled code).
309    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
310        compile-time error is now given if \c is not followed by an ASCII
311        character, that is, a byte less than 128. (In EBCDIC mode, the code is
312        different, and any byte value is allowed.)
314    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
315        START_OPTIMIZE option, which is now allowed at compile time - but just
316        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
317        to pcregrep and other applications that have no direct access to PCRE
318        options. The new /Y option in pcretest sets this option when calling
319        pcre_compile().
321    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
322        back references. Groups containing recursive back references were forced to
323        be atomic by that change, but in the case of named groups, the amount of
324        memory required was incorrectly computed, leading to "Failed: internal
325        error: code overflow". This has been fixed.
327    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
328        pcretest.c, to avoid build problems in some Borland environments.
331    Version 8.10 25-Jun-2010
332    ------------------------
334    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
335        THEN.
337    2.  (*ACCEPT) was not working when inside an atomic group.
339    3.  Inside a character class, \B is treated as a literal by default, but
340        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
341        causes the error). The code is unchanged, but I tidied the documentation.
343    4.  Inside a character class, PCRE always treated \R and \X as literals,
344        whereas Perl faults them if its -w option is set. I have changed PCRE so
345        that it faults them when PCRE_EXTRA is set.
347    5.  Added support for \N, which always matches any character other than
348        newline. (It is the same as "." when PCRE_DOTALL is not set.)
350    6.  When compiling pcregrep with newer versions of gcc which may have
351        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
352        declared with attribute warn_unused_result" were given. Just casting the
353        result to (void) does not stop the warnings; a more elaborate fudge is
354        needed. I've used a macro to implement this.
356    7.  Minor change to pcretest.c to avoid a compiler warning.
358    8.  Added four artifical Unicode properties to help with an option to make
359        \s etc use properties (see next item). The new properties are: Xan
360        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
362    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
363        use Unicode properties. (*UCP) at the start of a pattern can be used to set
364        this option. Modified pcretest to add /W to test this facility. Added
365        REG_UCP to make it available via the POSIX interface.
367    10. Added --line-buffered to pcregrep.
369    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
370        studied, and the match started with a letter with a code point greater than
371        127 whose first byte was different to the first byte of the other case of
372        the letter, the other case of this starting letter was not recognized
373        (#976).
375    12. If a pattern that was studied started with a repeated Unicode property
376        test, for example, \p{Nd}+, there was the theoretical possibility of
377        setting up an incorrect bitmap of starting bytes, but fortunately it could
378        not have actually happened in practice until change 8 above was made (it
379        added property types that matched character-matching opcodes).
381    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
382        possible starting bytes for non-anchored patterns.
384    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
385        \R, and also a number of cases that involve Unicode properties, both
386        explicit and implicit when PCRE_UCP is set.
388    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
389        input, it could crash or give wrong results if characters with values
390        greater than 0xc0 were present in the subject string. (Detail: it assumed
391        UTF-8 input when processing these items.)
393    16. Added a lot of (int) casts to avoid compiler warnings in systems where
394        size_t is 64-bit (#991).
396    17. Added a check for running out of memory when PCRE is compiled with
397        --disable-stack-for-recursion (#990).
399    18. If the last data line in a file for pcretest does not have a newline on
400        the end, a newline was missing in the output.
402    19. The default pcre_chartables.c file recognizes only ASCII characters (values
403        less than 128) in its various bitmaps. However, there is a facility for
404        generating tables according to the current locale when PCRE is compiled. It
405        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
406        characters, are recognized by isspace() and therefore were getting set in
407        these tables, and indeed these tables seem to approximate to ISO 8859. This
408        caused a problem in UTF-8 mode when pcre_study() was used to create a list
409        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
410        which of course cannot start UTF-8 characters. I have changed the code so
411        that only real ASCII characters (less than 128) and the correct starting
412        bytes for UTF-8 encodings are set for characters greater than 127 when in
413        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
414        altogether.)
416    20. Added the /T option to pcretest so as to be able to run tests with non-
417        standard character tables, thus making it possible to include the tests
418        used for 19 above in the standard set of tests.
420    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
421        reference to a subpattern the other side of a comment that contains an
422        opening parenthesis caused either an internal compiling error, or a
423        reference to the wrong subpattern.
426    Version 8.02 19-Mar-2010
427    ------------------------
429    1.  The Unicode data tables have been updated to Unicode 5.2.0.
431    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
432        configured.
434    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
435        original author of that file, following a query about its status.
437    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
438        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
440    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
441        quantifier applied to a forward-referencing subroutine call, could compile
442        incorrect code or give the error "internal error: previously-checked
443        referenced subpattern not found".
445    6.  Both MS Visual Studio and Symbian OS have problems with initializing
446        variables to point to external functions. For these systems, therefore,
447        pcre_malloc etc. are now initialized to local functions that call the
448        relevant global functions.
450    7.  There were two entries missing in the vectors called coptable and poptable
451        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
452        I've fixed the data, and added a kludgy way of testing at compile time that
453        the lengths are correct (equal to the number of opcodes).
455    8.  Following on from 7, I added a similar kludge to check the length of the
456        eint vector in pcreposix.c.
458    9.  Error texts for pcre_compile() are held as one long string to avoid too
459        much relocation at load time. To find a text, the string is searched,
460        counting zeros. There was no check for running off the end of the string,
461        which could happen if a new error number was added without updating the
462        string.
464    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
466    11. \K was not working if it appeared in an atomic group or in a group that
467        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
468        \K is "not well defined" if used in an assertion. PCRE now accepts it if
469        the assertion is positive, but not if it is negative.
471    12. Change 11 fortuitously reduced the size of the stack frame used in the
472        "match()" function of pcre_exec.c by one pointer. Forthcoming
473        implementation of support for (*MARK) will need an extra pointer on the
474        stack; I have reserved it now, so that the stack frame size does not
475        decrease.
477    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
478        item in branch that calls a recursion is a subroutine call - as in the
479        second branch in the above example - was incorrectly given the compile-
480        time error "recursive call could loop indefinitely" because pcre_compile()
481        was not correctly checking the subroutine for matching a non-empty string.
483    14. The checks for overrunning compiling workspace could trigger after an
484        overrun had occurred. This is a "should never occur" error, but it can be
485        triggered by pathological patterns such as hundreds of nested parentheses.
486        The checks now trigger 100 bytes before the end of the workspace.
488    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
491    Version 8.01 19-Jan-2010
492    ------------------------
494    1.  If a pattern contained a conditional subpattern with only one branch (in
495        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
496        computed the wrong minimum data length (which is of course zero for such
497        subpatterns). This could cause incorrect "no match" results.
499    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
500        the pattern is reset in the first branch, pcre_compile() failed with
501        "internal error: code overflow at offset...". This happened only when
502        the reset was to the original external option setting. (An optimization
503        abstracts leading options settings into an external setting, which was the
504        cause of this.)
506    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
507        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
508        assertion pattern did not match (meaning that the assertion was true), it
509        was incorrectly treated as false if the SKIP had been reached during the
510        matching. This also applied to assertions used as conditions.
512    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
513        assertion subpattern, including such a pattern used as a condition,
514        unpredictable results occurred, instead of the error return
517    5.  The C++ GlobalReplace function was not working like Perl for the special
518        situation when an empty string is matched. It now does the fancy magic
519        stuff that is necessary.
521    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
522        removed. (These were left over from very, very early versions of PCRE.)
524    7.  Some cosmetic changes to the code to make life easier when compiling it
525        as part of something else:
527        (a) Change DEBUG to PCRE_DEBUG.
529        (b) In pcre_compile(), rename the member of the "branch_chain" structure
530            called "current" as "current_branch", to prevent a collision with the
531            Linux macro when compiled as a kernel module.
533        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
534            prevent a collision with the Linux macro when compiled as a kernel
535            module.
537    8.  In pcre_compile() there are some checks for integer overflows that used to
538        cast potentially large values to (double). This has been changed to that
539        when building, a check for int64_t is made, and if it is found, it is used
540        instead, thus avoiding the use of floating point arithmetic. (There is no
541        other use of FP in PCRE.) If int64_t is not found, the fallback is to
542        double.
544    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
545        2005 (difference between two addresses compared to an unsigned value).
547    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
548        custom one, because of the following reported problem in Windows:
550          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
551              under Win32.
552          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
553              therefore missing the function definition.
554          - The compiler thus generates a "C" signature for the test function.
555          - The linker fails to find the "C" function.
556          - PCRE fails to configure if asked to do so against libbz2.
558    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
559        messages were output:
561          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
562          rerunning libtoolize, to keep the correct libtool macros in-tree.
563          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
565        I have done both of these things.
567    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
568        most of the time, it *can* run out if it is given a pattern that contains a
569        runaway infinite recursion. I updated the discussion in the pcrestack man
570        page.
572    13. Now that we have gone to the x.xx style of version numbers, the minor
573        version may start with zero. Using 08 or 09 is a bad idea because users
574        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
575        interpreted as invalid octal numbers. I've updated the previous comment in
576        configure.ac, and also added a check that gives an error if 08 or 09 are
577        used.
579    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
580        causing partial matching to fail when the end of the subject matched \W
581        in a UTF-8 pattern where \W was quantified with a minimum of 3.
583    15. There were some discrepancies between the declarations in pcre_internal.h
584        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
585        their definitions. The declarations used "const uschar *" and the
586        definitions used USPTR. Even though USPTR is normally defined as "const
587        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
588        reported that: "This difference in casting confuses some C++ compilers, for
589        example, SunCC recognizes above declarations as different functions and
590        generates broken code for hbpcre." I have changed the declarations to use
591        USPTR.
593    16. GNU libtool is named differently on some systems. The autogen.sh script now
594        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
595        (FreeBSD).
597    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
598        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
599        comment: "Figure out how to create a longlong from a string: strtoll and
600        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
601        instance, but it only takes 2 args instead of 3!"
603    18. A subtle bug concerned with back references has been fixed by a change of
604        specification, with a corresponding code fix. A pattern such as
605        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
606        refers, was giving matches when it shouldn't. For example, xa=xaaa would
607        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
608        same bug. Such groups have to be quantified to be useful, or contained
609        inside another quantified group. (If there's no repetition, the reference
610        can never match.) The problem arises because, having left the group and
611        moved on to the rest of the pattern, a later failure that backtracks into
612        the group uses the captured value from the final iteration of the group
613        rather than the correct earlier one. I have fixed this in PCRE by forcing
614        any group that contains a reference to itself to be an atomic group; that
615        is, there cannot be any backtracking into it once it has completed. This is
616        similar to recursive and subroutine calls.
619    Version 8.00 19-Oct-09
620    ----------------------
622    1.  The table for translating pcre_compile() error codes into POSIX error codes
623        was out-of-date, and there was no check on the pcre_compile() error code
624        being within the table. This could lead to an OK return being given in
625        error.
627    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
628        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
629        in a Windows environment.
631    3.  The pcregrep --count option prints the count for each file even when it is
632        zero, as does GNU grep. However, pcregrep was also printing all files when
633        --files-with-matches was added. Now, when both options are given, it prints
634        counts only for those files that have at least one match. (GNU grep just
635        prints the file name in this circumstance, but including the count seems
636        more useful - otherwise, why use --count?) Also ensured that the
637        combination -clh just lists non-zero counts, with no names.
639    4.  The long form of the pcregrep -F option was incorrectly implemented as
640        --fixed_strings instead of --fixed-strings. This is an incompatible change,
641        but it seems right to fix it, and I didn't think it was worth preserving
642        the old behaviour.
644    5.  The command line items --regex=pattern and --regexp=pattern were not
645        recognized by pcregrep, which required --regex pattern or --regexp pattern
646        (with a space rather than an '='). The man page documented the '=' forms,
647        which are compatible with GNU grep; these now work.
649    6.  No libpcreposix.pc file was created for pkg-config; there was just
650        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
652    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
653        when UCP support is not needed, by modifying the Python script that
654        generates it from Unicode data files. This should not matter if the module
655        is correctly used as a library, but I received one complaint about 50K of
656        unwanted data. My guess is that the person linked everything into his
657        program rather than using a library. Anyway, it does no harm.
659    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
660        was a minimum greater than 1 for a wide character in a possessive
661        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
662        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
663        character. Chaos in the form of incorrect output or a compiling loop could
664        result.
666    9.  The restrictions on what a pattern can contain when partial matching is
667        requested for pcre_exec() have been removed. All patterns can now be
668        partially matched by this function. In addition, if there are at least two
669        slots in the offset vector, the offset of the earliest inspected character
670        for the match and the offset of the end of the subject are set in them when
671        PCRE_ERROR_PARTIAL is returned.
673    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
674        synonymous with PCRE_PARTIAL, for backwards compatibility, and
675        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
676        and may be more useful for multi-segment matching.
678    11. Partial matching with pcre_exec() is now more intuitive. A partial match
679        used to be given if ever the end of the subject was reached; now it is
680        given only if matching could not proceed because another character was
681        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
682        string "Z", which now yields "no match" instead of "partial match". In the
683        case of pcre_dfa_exec(), "no match" is given if every matching path for the
684        final character ended with (*FAIL).
686    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
687        if the pattern had a "must contain" character that was already found in the
688        earlier partial match, unless partial matching was again requested. For
689        example, with the pattern /dog.(body)?/, the "must contain" character is
690        "g". If the first part-match was for the string "dog", restarting with
691        "sbody" failed. This bug has been fixed.
693    13. The string returned by pcre_dfa_exec() after a partial match has been
694        changed so that it starts at the first inspected character rather than the
695        first character of the match. This makes a difference only if the pattern
696        starts with a lookbehind assertion or \b or \B (\K is not supported by
697        pcre_dfa_exec()). It's an incompatible change, but it makes the two
698        matching functions compatible, and I think it's the right thing to do.
700    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
701        so that the demonstration program is easily available in environments where
702        PCRE has not been installed from source.
704    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
705        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
706        library.
708    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
709        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
710        is not the first non-POSIX option to be added. Clearly some people find
711        these options useful.
713    17. If a caller to the POSIX matching function regexec() passes a non-zero
714        value for nmatch with a NULL value for pmatch, the value of
715        nmatch is forced to zero.
717    18. RunGrepTest did not have a test for the availability of the -u option of
718        the diff command, as RunTest does. It now checks in the same way as
719        RunTest, and also checks for the -b option.
721    19. If an odd number of negated classes containing just a single character
722        interposed, within parentheses, between a forward reference to a named
723        subpattern and the definition of the subpattern, compilation crashed with
724        an internal error, complaining that it could not find the referenced
725        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
726        [The bug was that it was starting one character too far in when skipping
727        over the character class, thus treating the ] as data rather than
728        terminating the class. This meant it could skip too much.]
730    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
731        /g option in pcretest when the pattern contains \K, which makes it possible
732        to have an empty string match not at the start, even when the pattern is
733        anchored. Updated pcretest and pcredemo to use this option.
735    21. If the maximum number of capturing subpatterns in a recursion was greater
736        than the maximum at the outer level, the higher number was returned, but
737        with unset values at the outer level. The correct (outer level) value is
738        now given.
740    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
741        PCRE did not set those parentheses (unlike Perl). I have now found a way to
742        make it do so. The string so far is captured, making this feature
743        compatible with Perl.
745    23. The tests have been re-organized, adding tests 11 and 12, to make it
746        possible to check the Perl 5.10 features against Perl 5.10.
748    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
749        pattern matches a fixed length string. PCRE did not allow this; now it
750        does. Neither allows recursion.
752    25. I finally figured out how to implement a request to provide the minimum
753        length of subject string that was needed in order to match a given pattern.
754        (It was back references and recursion that I had previously got hung up
755        on.) This code has now been added to pcre_study(); it finds a lower bound
756        to the length of subject needed. It is not necessarily the greatest lower
757        bound, but using it to avoid searching strings that are too short does give
758        some useful speed-ups. The value is available to calling programs via
759        pcre_fullinfo().
761    26. While implementing 25, I discovered to my embarrassment that pcretest had
762        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
763        study optimizations had never been tested with that matching function.
764        Oops. What is worse, even when it was passed study data, there was a bug in
765        pcre_dfa_exec() that meant it never actually used it. Double oops. There
766        were also very few tests of studied patterns with pcre_dfa_exec().
768    27. If (?| is used to create subpatterns with duplicate numbers, they are now
769        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
770        on the other side of the coin, they are no longer allowed to have different
771        names, because these cannot be distinguished in PCRE, and this has caused
772        confusion. (This is a difference from Perl.)
774    28. When duplicate subpattern names are present (necessarily with different
775        numbers, as required by 27 above), and a test is made by name in a
776        conditional pattern, either for a subpattern having been matched, or for
777        recursion in such a pattern, all the associated numbered subpatterns are
778        tested, and the overall condition is true if the condition is true for any
779        one of them. This is the way Perl works, and is also more like the way
780        testing by number works.
783    Version 7.9 11-Apr-09
784    ---------------------
786    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
787        (pcretest), all targets were linked against these libraries. This included
788        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
789        libraries. This caused unwanted dependencies to be created. This problem
790        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
791        pcretest is linked with readline.
793    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
794        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
795        moved outside it again, because FALSE and TRUE are already defined in AIX,
796        but BOOL is not.
798    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
799        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
801    4.  The pcregrep documentation said spaces were inserted as well as colons (or
802        hyphens) following file names and line numbers when outputting matching
803        lines. This is not true; no spaces are inserted. I have also clarified the
804        wording for the --colour (or --color) option.
806    5.  In pcregrep, when --colour was used with -o, the list of matching strings
807        was not coloured; this is different to GNU grep, so I have changed it to be
808        the same.
810    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
811        each matching line was coloured. Now it goes on to look for further matches
812        of any of the test patterns, which is the same behaviour as GNU grep.
814    7.  A pattern that could match an empty string could cause pcregrep to loop; it
815        doesn't make sense to accept an empty string match in pcregrep, so I have
816        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
817        seems to be how GNU grep behaves.
819    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
820        start or after a newline", because the conditional assertion was not being
821        correctly handled. The rule now is that both the assertion and what follows
822        in the first alternative must satisfy the test.
824    9.  If auto-callout was enabled in a pattern with a conditional group whose
825        condition was an assertion, PCRE could crash during matching, both with
826        pcre_exec() and pcre_dfa_exec().
828    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
829        used for matching.
831    11. Unicode property support in character classes was not working for
832        characters (bytes) greater than 127 when not in UTF-8 mode.
834    12. Added the -M command line option to pcretest.
836    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
838    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
840    16. Added comments and documentation about mis-use of no_arg in the C++
841        wrapper.
843    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
844        from Martin Jerabek that uses macro names for all relevant character and
845        string constants.
847    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
848        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
849        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
850        these, but not everybody uses configure.
852    19. A conditional group that had only one branch was not being correctly
853        recognized as an item that could match an empty string. This meant that an
854        enclosing group might also not be so recognized, causing infinite looping
855        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
856        with the subject "ab", where knowledge that the repeated group can match
857        nothing is needed in order to break the loop.
859    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
860        exec(), but without supplying a callout function, matching went wrong.
862    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
863        leak if the size of the offset vector was greater than 30. When the vector
864        is smaller, the saved offsets during recursion go onto a local stack
865        vector, but for larger vectors malloc() is used. It was failing to free
866        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
867        error, in fact).
869    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
870        heapframe that is used only when UTF-8 support is enabled. This caused no
871        problem, but was untidy.
873    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
874        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
875        included within another project.
877    24. Steven Van Ingelgem's patches to add more options to the CMake support,
878        slightly modified by me:
880          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
881              not building pcregrep.
883          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
884              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
886    25. Forward references, both numeric and by name, in patterns that made use of
887        duplicate group numbers, could behave incorrectly or give incorrect errors,
888        because when scanning forward to find the reference group, PCRE was not
889        taking into account the duplicate group numbers. A pattern such as
890        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
892    26. Changed a few more instances of "const unsigned char *" to USPTR, making
893        the feature of a custom pointer more persuasive (as requested by a user).
895    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
896        pcretest.c, inside #ifndefs, because it seems they are sometimes already
897        pre-defined.
899    28. Added support for (*UTF8) at the start of a pattern.
901    29. Arrange for flags added by the "release type" setting in CMake to be shown
902        in the configuration summary.
905    Version 7.8 05-Sep-08
906    ---------------------
908    1.  Replaced UCP searching code with optimized version as implemented for Ad
909        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
910        stage table and inline lookup instead of a function, giving speed ups of 2
911        to 5 times on some simple patterns that I tested. Permission was given to
912        distribute the MultiStage2.py script that generates the tables (it's not in
913        the tarball, but is in the Subversion repository).
915    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
916        scripts.
918    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
919        a group with a zero qualifier. The result of the study could be incorrect,
920        or the function might crash, depending on the pattern.
922    4.  Caseless matching was not working for non-ASCII characters in back
923        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
924        It now works when Unicode Property Support is available.
926    5.  In pcretest, an escape such as \x{de} in the data was always generating
927        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
928        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
929        truncation.
931    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
933    7.  Added two (int) casts to pcregrep when printing the difference of two
934        pointers, in case they are 64-bit values.
936    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
937        test 2 if it fails.
939    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
940        and a #define of that name to empty if it is not externally set. This is to
941        allow users of MSVC to set it if necessary.
943    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
944        the convenience functions in the pcre_get.c source file.
946    11. An option change at the start of a pattern that had top-level alternatives
947        could cause overwriting and/or a crash. This command provoked a crash in
948        some environments:
950          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
952        This potential security problem was recorded as CVE-2008-2371.
954    12. For a pattern where the match had to start at the beginning or immediately
955        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
956        pcre_dfa_exec() could read past the end of the passed subject if there was
957        no match. To help with detecting such bugs (e.g. with valgrind), I modified
958        pcretest so that it places the subject at the end of its malloc-ed buffer.
960    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
961        exec() might read past the end of the data buffer in UTF-8 mode.
963    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
964        the data contained the byte 0x85 as part of a UTF-8 character within its
965        first line. This applied both to normal and DFA matching.
967    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
968        /^[^d]*?$/8 failed to match "abc".
970    16. Added a missing copyright notice to pcrecpp_internal.h.
972    17. Make it more clear in the documentation that values returned from
973        pcre_exec() in ovector are byte offsets, not character counts.
975    18. Tidied a few places to stop certain compilers from issuing warnings.
977    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
978        supplied by Stefan Weber. I made a further small update for 7.8 because
979        there is a change of source arrangements: the pcre_searchfuncs.c module is
980        replaced by pcre_ucd.c.
983    Version 7.7 07-May-08
984    ---------------------
986    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
987        a string to a long long, pretend we don't even have a long long." This is
988        done by checking for the strtoq, strtoll, and _strtoi64 functions.
990    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
991        pre-7.6 versions, which defined a global no_arg variable instead of putting
992        it in the RE class. (See also #8 below.)
994    3.  Remove a line of dead code, identified by coverity and reported by Nuno
995        Lopes.
997    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
999        (1) The include/exclude patterns were being applied to the whole pathnames
1000            of files, instead of just to the final components.
1002        (2) If there was more than one level of directory, the subdirectories were
1003            skipped unless they satisfied the include/exclude conditions. This is
1004            inconsistent with GNU grep (and could even be seen as contrary to the
1005            pcregrep specification - which I improved to make it absolutely clear).
1006            The action now is always to scan all levels of directory, and just
1007            apply the include/exclude patterns to regular files.
1009    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1010        --exclude_dir in the tests to avoid scanning .svn directories.
1012    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1013        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1014        doesn't support NULs in patterns.
1016    7.  Added some missing "const"s to declarations of static tables in
1017        pcre_compile.c and pcre_dfa_exec.c.
1019    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1020        caused by fix #2  above. (Subsequently also a second patch to fix the
1021        first patch. And a third patch - this was a messy problem.)
1023    9.  Applied Craig's patch to remove the use of push_back().
1025    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1026        matching function regexec().
1028    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1029        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1030        references. PCRE supports relative numbers with this syntax (I don't think
1031        Oniguruma does).
1033    12. Previously, a group with a zero repeat such as (...){0} was completely
1034        omitted from the compiled regex. However, this means that if the group
1035        was called as a subroutine from elsewhere in the pattern, things went wrong
1036        (an internal error was given). Such groups are now left in the compiled
1037        pattern, with a new opcode that causes them to be skipped at execution
1038        time.
1040    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1041        to the way PCRE behaves:
1043        (a) A lone ] character is dis-allowed (Perl treats it as data).
1045        (b) A back reference to an unmatched subpattern matches an empty string
1046            (Perl fails the current match path).
1048        (c) A data ] in a character class must be notated as \] because if the
1049            first data character in a class is ], it defines an empty class. (In
1050            Perl it is not possible to have an empty class.) The empty class []
1051            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1052            The negative empty class [^] matches any one character, independently
1053            of the DOTALL setting.
1055    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1056        non-existent subpattern following a character class starting with ']' and
1057        containing () gave an internal compiling error instead of "reference to
1058        non-existent subpattern". Fortunately, when the pattern did exist, the
1059        compiled code was correct. (When scanning forwards to check for the
1060        existencd of the subpattern, it was treating the data ']' as terminating
1061        the class, so got the count wrong. When actually compiling, the reference
1062        was subsequently set up correctly.)
1064    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1065        it was being rejected as not supported by pcre_dfa_exec(), even though
1066        other assertions are supported. I have made pcre_dfa_exec() support
1067        (*FAIL).
1069    16. The implementation of 13c above involved the invention of a new opcode,
1070        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1071        cannot be changed at match time, I realized I could make a small
1072        improvement to matching performance by compiling OP_ALLANY instead of
1073        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1074        on the OP_ANY path.
1076    17. Compiling pcretest on Windows with readline support failed without the
1077        following two fixes: (1) Make the unistd.h include conditional on
1078        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1080    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1081        ncurses library to be included for pcretest when ReadLine support is
1082        requested, but also to allow for it to be overridden. This patch came from
1083        Daniel Bergström.
1085    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1086        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1087        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1088        spotting this.
1091    Version 7.6 28-Jan-08
1092    ---------------------
1094    1.  A character class containing a very large number of characters with
1095        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1096        overflow.
1098    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1099        HAVE_LONG_LONG is not defined.
1101    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1102        bring them up to date and include new features. This patch includes:
1104        - Fixed PH's badly added libz and libbz2 support.
1105        - Fixed a problem with static linking.
1106        - Added pcredemo. [But later removed - see 7 below.]
1107        - Fixed dftables problem and added an option.
1108        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1109            HAVE_LONG_LONG.
1110        - Added readline support for pcretest.
1111        - Added an listing of the option settings after cmake has run.
1113    4.  A user submitted a patch to Makefile that makes it easy to create
1114        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1115        Makefile.am that cause it to include this special target, without
1116        affecting anything else. Note that the same mingw target plus all
1117        the other distribution libraries and programs are now supported
1118        when configuring with CMake (see 6 below) instead of with
1119        Configure/Make.
1121    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1122        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1123        exported in the Windows port". It has not yet been confirmed that the patch
1124        solves the problem, but it does no harm.
1126    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1127        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1128        with CMake, and also correct the comment about stack recursion.
1130    7.  Remove the automatic building of pcredemo from the ./configure system and
1131        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1132        of a program that users should build themselves after PCRE is installed, so
1133        building it automatically is not really right. What is more, it gave
1134        trouble in some build environments.
1136    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1139    Version 7.5 10-Jan-08
1140    ---------------------
1142    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1143        values in parens when parsing an RE using the C++ wrapper."
1145    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1146        Characters greater than 255 were excluded from the class instead of being
1147        included.
1149    3.  The same bug as (2) above applied to negated POSIX classes such as
1150        [:^space:].
1152    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1153        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1154        I have changed it.
1156    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1157        first named subpattern) and a construct such as (?&a) would reference the
1158        first named subpattern whose name started with "a" (in other words, the
1159        length check was missing). Both these problems are fixed. "Subpattern name
1160        expected" is now given for (?&) (a zero-length name), and this patch also
1161        makes it give the same error for \k'' (previously it complained that that
1162        was a reference to a non-existent subpattern).
1164    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1165        this is right because (?- can be followed by option settings as well as by
1166        digits. I have, however, made the messages clearer.
1168    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1169        than the number used in the conditional) now cause a compile-time error.
1170        This is actually not compatible with Perl, which accepts such patterns, but
1171        treats the conditional as always being FALSE (as PCRE used to), but it
1172        seems to me that giving a diagnostic is better.
1174    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1175        and messages.
1177    9.  Fix two occurrences of "backslash" in comments that should have been
1178        "backspace".
1180    10. Remove two redundant lines of code that can never be obeyed (their function
1181        was moved elsewhere).
1183    11. The program that makes PCRE's Unicode character property table had a bug
1184        which caused it to generate incorrect table entries for sequences of
1185        characters that have the same character type, but are in different scripts.
1186        It amalgamated them into a single range, with the script of the first of
1187        them. In other words, some characters were in the wrong script. There were
1188        thirteen such cases, affecting characters in the following ranges:
1190          U+002b0 - U+002c1
1191          U+0060c - U+0060d
1192          U+0061e - U+00612
1193          U+0064b - U+0065e
1194          U+0074d - U+0076d
1195          U+01800 - U+01805
1196          U+01d00 - U+01d77
1197          U+01d9b - U+01dbf
1198          U+0200b - U+0200f
1199          U+030fc - U+030fe
1200          U+03260 - U+0327f
1201          U+0fb46 - U+0fbb1
1202          U+10450 - U+1049d
1204    12. The -o option (show only the matching part of a line) for pcregrep was not
1205        compatible with GNU grep in that, if there was more than one match in a
1206        line, it showed only the first of them. It now behaves in the same way as
1207        GNU grep.
1209    13. If the -o and -v options were combined for pcregrep, it printed a blank
1210        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1211        does the same. The return code can be used to tell if there were any
1212        non-matching lines.
1214    14. Added --file-offsets and --line-offsets to pcregrep.
1216    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1217        infinitely looping recursion. The bug was that positive lookaheads were not
1218        being skipped when checking for a possible empty match (negative lookaheads
1219        and both kinds of lookbehind were skipped).
1221    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1222        inclusion of <windows.h> to before rather than after the definition of
1223        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1225    17. Specifying a possessive quantifier with a specific limit for a Unicode
1226        character property caused pcre_compile() to compile bad code, which led at
1227        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1228        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1229        caused the error; without that there was no problem.
1231    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1233    19. Added --enable-pcretest-libreadline.
1235    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1236        RE::GlobalReplace(). As a result, the number of replacements returned was
1237        double what it should be. I removed one of the increments, but Craig sent a
1238        later patch that removed the other one (the right fix) and added unit tests
1239        that check the return values (which was not done before).
1241    21. Several CMake things:
1243        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1244            the names libpcre and libpcreposix, not just pcre and pcreposix.
1246        (2) The above change means that pcretest and pcregrep are now correctly
1247            linked with the newly-built libraries, not previously installed ones.
1251    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1252        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1253        UTF-8 newline character). The key issue is that the pattern starts .*;
1254        this means that the match must be either at the beginning, or after a
1255        newline. The bug was in the code for advancing after a failed match and
1256        checking that the new position followed a newline. It was not taking
1257        account of UTF-8 characters correctly.
1259    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1260        character classes. PCRE was not treating the sequence [:...:] as a
1261        character class unless the ... were all letters. Perl, however, seems to
1262        allow any characters between [: and :], though of course it rejects as
1263        unknown any "names" that contain non-letters, because all the known class
1264        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1265        for example, whereas PCRE did not - it did not recognize a POSIX character
1266        class. This seemed a bit dangerous, so the code has been changed to be
1267        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1268        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1269        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1270        Perl does, and where it didn't before.
1272    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1273        Windows environments %n is disabled by default.
1276    Version 7.4 21-Sep-07
1277    ---------------------
1279    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1280        means that a class such as [\s] counted as "explicit reference to CR or
1281        LF". That isn't really right - the whole point of the change was to try to
1282        help when there was an actual mention of one of the two characters. So now
1283        the change happens only if \r or \n (or a literal CR or LF) character is
1284        encountered.
1286    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1287        of both had grown to the point where there were only 3 bits left.
1288        Fortunately, there was spare space in the data structure, and so I have
1289        moved the internal flags into a new 16-bit field to free up more option
1290        bits.
1292    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1293        but did not set the internal JCHANGED flag - either of these is enough to
1294        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1295        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1296        start sets both bits.
1298    4.  Added options (at build time, compile time, exec time) to change \R from
1299        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1301    5.  doc/pcresyntax.html was missing from the distribution.
1303    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1304        compatibility, even though it is no longer used.
1306    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1307        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1308        windows.h file is present (where different names are used). [This was
1309        reversed later after testing - see 16 below.]
1311    8.  Changed all #include <config.h> to #include "config.h". There were also
1312        some further <pcre.h> cases that I changed to "pcre.h".
1314    9.  When pcregrep was used with the --colour option, it missed the line ending
1315        sequence off the lines that it output.
1317    10. It was pointed out to me that arrays of string pointers cause lots of
1318        relocations when a shared library is dynamically loaded. A technique of
1319        using a single long string with a table of offsets can drastically reduce
1320        these. I have refactored PCRE in four places to do this. The result is
1321        dramatic:
1323          Originally:                          290
1324          After changing UCP table:            187
1325          After changing error message table:   43
1326          After changing table of "verbs"       36
1327          After changing table of Posix names   22
1329        Thanks to the folks working on Gregex for glib for this insight.
1331    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1332        unicode-properties was also set.
1334    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1336    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1337        checked only for CRLF.
1339    14. Added casts to pcretest.c to avoid compiler warnings.
1341    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1343    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1344        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1345        entirely. This removes changes made in 7 above.
1347    17. The CMake files have been updated, and there is now more information about
1348        building with CMake in the NON-UNIX-USE document.
1351    Version 7.3 28-Aug-07
1352    ---------------------
1354     1. In the rejigging of the build system that eventually resulted in 7.1, the
1355        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1356        brackets there is not right, since it causes compilers to look for an
1357        installed pcre.h, not the version that is in the source that is being
1358        compiled (which of course may be different). I have changed it back to:
1360          #include "pcre.h"
1362        I have a vague recollection that the change was concerned with compiling in
1363        different directories, but in the new build system, that is taken care of
1364        by the VPATH setting the Makefile.
1366     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1367        when the subject happened to end in the byte 0x85 (e.g. if the last
1368        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1369        characters but of course it shouldn't be taken as a newline when it is part
1370        of another character. The bug was that, for an unlimited repeat of . in
1371        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1372        characters when looking for a newline.
1374     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1376     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1377        in debug output.
1379     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1380        long printing in the pcrecpp unittest when running under MinGW.
1382     6. ESC_K was left out of the EBCDIC table.
1384     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1385        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1386        limit also applies to "virtual nesting" when a pattern is recursive, and in
1387        this case 1000 isn't so big. I have been able to remove this limit at the
1388        expense of backing off one optimization in certain circumstances. Normally,
1389        when pcre_exec() would call its internal match() function recursively and
1390        immediately return the result unconditionally, it uses a "tail recursion"
1391        feature to save stack. However, when a subpattern that can match an empty
1392        string has an unlimited repetition quantifier, it no longer makes this
1393        optimization. That gives it a stack frame in which to save the data for
1394        checking that an empty string has been matched. Previously this was taken
1395        from the 1000-entry workspace that had been reserved. So now there is no
1396        explicit limit, but more stack is used.
1398     8. Applied Daniel's patches to solve problems with the import/export magic
1399        syntax that is required for Windows, and which was going wrong for the
1400        pcreposix and pcrecpp parts of the library. These were overlooked when this
1401        problem was solved for the main library.
1403     9. There were some crude static tests to avoid integer overflow when computing
1404        the size of patterns that contain repeated groups with explicit upper
1405        limits. As the maximum quantifier is 65535, the maximum group length was
1406        set at 30,000 so that the product of these two numbers did not overflow a
1407        32-bit integer. However, it turns out that people want to use groups that
1408        are longer than 30,000 bytes (though not repeat them that many times).
1409        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1410        made it possible to implement the integer overflow checks in a much more
1411        dynamic way, which I have now done. The artificial limitation on group
1412        length has been removed - we now have only the limit on the total length of
1413        the compiled pattern, which depends on the LINK_SIZE setting.
1415    10. Fixed a bug in the documentation for get/copy named substring when
1416        duplicate names are permitted. If none of the named substrings are set, the
1417        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1418        empty string.
1420    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1421        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1422        because the ] is interpreted as the first data character and the
1423        terminating ] is not found. PCRE has been made compatible with Perl in this
1424        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1425        cause memory overwriting.
1427    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1428        string has been matched (to stop an infinite loop). It was not recognizing
1429        a conditional subpattern that could match an empty string if that
1430        subpattern was within another subpattern. For example, it looped when
1431        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1432        condition was not nested. This bug has been fixed.
1434    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1435        past the start of the subject in the presence of bytes with the top bit
1436        set, for example "\x8aBCD".
1438    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1439        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1441    14. Optimized (?!) to (*FAIL).
1443    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1444        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1445        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1446        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1447        does: it's just the validity check that is more restrictive.
1449    16. Inserted checks for integer overflows during escape sequence (backslash)
1450        processing, and also fixed erroneous offset values for syntax errors during
1451        backslash processing.
1453    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1454        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1456    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1457        caused an overrun.
1459    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1460        something other than just ASCII characters) inside a group that had an
1461        unlimited repeat caused a loop at compile time (while checking to see
1462        whether the group could match an empty string).
1464    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1465        [\P{Any}] did so. (Error in the code for printing property names.)
1467    21. An orphan \E inside a character class could cause a crash.
1469    22. A repeated capturing bracket such as (A)? could cause a wild memory
1470        reference during compilation.
1472    23. There are several functions in pcre_compile() that scan along a compiled
1473        expression for various reasons (e.g. to see if it's fixed length for look
1474        behind). There were bugs in these functions when a repeated \p or \P was
1475        present in the pattern. These operators have additional parameters compared
1476        with \d, etc, and these were not being taken into account when moving along
1477        the compiled data. Specifically:
1479        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1480            length.
1482        (b) An item such as \pL+ within a repeated group could cause crashes or
1483            loops.
1485        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1486            "reference to non-existent subpattern" error.
1488        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1490    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1491        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1493    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1495    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1496        character were causing crashes (broken optimization).
1498    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1499        \p or \P) caused a compile-time loop.
1501    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1502        break. For example, the unstudied pattern [\r\n]A does not match the string
1503        "\r\nA" because change 7.0/46 below moves the current point on by two
1504        characters after failing to match at the start. However, the pattern \nA
1505        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1506        the same is true. There doesn't seem any very clean way out of this, but
1507        what I have chosen to do makes the common cases work: PCRE now takes note
1508        of whether there can be an explicit match for \r or \n anywhere in the
1509        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1510        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1511        pattern has explicit CR or LF references.
1513    29. Added (*CR) etc for changing newline setting at start of pattern.
1516    Version 7.2 19-Jun-07
1517    ---------------------
1519     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1520        which is apparently normally available under Windows.
1522     2. Re-jig the pcregrep tests with different newline settings in an attempt
1523        to make them independent of the local environment's newline setting.
1525     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1527     4. Some of the "internals" tests were previously cut out when the link size
1528        was not 2, because the output contained actual offsets. The recent new
1529        "Z" feature of pcretest means that these can be cut out, making the tests
1530        usable with all link sizes.
1532     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1533        stack recursion. This gives a massive performance boost under BSD, but just
1534        a small improvement under Linux. However, it saves one field in the frame
1535        in all cases.
1537     6. Added more features from the forthcoming Perl 5.10:
1539        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1540            recursion call. It refers to the nth most recently opened parentheses.
1542        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1543            to be opened parentheses.
1545        (c) Conditions that refer to capturing parentheses can be specified
1546            relatively, for example, (?(-2)... or (?(+3)...
1548        (d) \K resets the start of the current match so that everything before
1549            is not part of it.
1551        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1553        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1554            reference syntax.
1556        (g) (?| introduces a group in which the numbering of parentheses in each
1557            alternative starts with the same number.
1559        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1561     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1564     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1565        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1566        for detecting groups that can match an empty string.
1568     9. A pattern with a very large number of alternatives (more than several
1569        hundred) was running out of internal workspace during the pre-compile
1570        phase, where pcre_compile() figures out how much memory will be needed. A
1571        bit of new cunning has reduced the workspace needed for groups with
1572        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1573        workspace instead of running out of the 4096 that are available.
1575    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1577    11. Applied patch from Google to remove an optimization that didn't quite work.
1578        The report of the bug said:
1580          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1581          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1582          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1584    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1585        it matched the wrong number of bytes.
1588    Version 7.1 24-Apr-07
1589    ---------------------
1591     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1592        that is more "standard", making use of automake and other Autotools. There
1593        is some re-arrangement of the files and adjustment of comments consequent
1594        on this.
1596     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1597        for recursive directory scanning broke on some systems because the files
1598        are not scanned in any specific order and on different systems the order
1599        was different. A call to "sort" has been inserted into RunGrepTest for the
1600        approprate test as a short-term fix. In the longer term there may be an
1601        alternative.
1603     3. I had an email from Eric Raymond about problems translating some of PCRE's
1604        man pages to HTML (despite the fact that I distribute HTML pages, some
1605        people do their own conversions for various reasons). The problems
1606        concerned the use of low-level troff macros .br and .in. I have therefore
1607        removed all such uses from the man pages (some were redundant, some could
1608        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1609        HTML has been updated to handle .nf/.fi and to complain if it encounters
1610        .br or .in.
1612     4. Updated comments in configure.ac that get placed in config.h.in and also
1613        arranged for config.h to be included in the distribution, with the name
1614        config.h.generic, for the benefit of those who have to compile without
1615        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1617     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1618        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1619        makevp.bat for the current PCRE, using the additional files
1620        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1622     6. A Windows user reported a minor discrepancy with test 2, which turned out
1623        to be caused by a trailing space on an input line that had got lost in his
1624        copy. The trailing space was an accident, so I've just removed it.
1626     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1627        that is needed.
1629     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1630        as "const" (a) because they are and (b) because it helps the PHP
1631        maintainers who have recently made a script to detect big data structures
1632        in the php code that should be moved to the .rodata section. I remembered
1633        to update Builducptable as well, so it won't revert if ucptable.h is ever
1634        re-created.
1636     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1637        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1638        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1639        support is not required. This saves 1.5-2K of code, which is important in
1640        some applications.
1642        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1643        so as not to refer to the tables, even though these functions will never be
1644        called when UTF-8 support is disabled. Otherwise there are problems with a
1645        shared library.
1647    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1649        (a) It was defining its arguments as char * instead of void *.
1651        (b) It was assuming that all moves were upwards in memory; this was true
1652            a long time ago when I wrote it, but is no longer the case.
1654        The emulated memove() is provided for those environments that have neither
1655        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1656        is clearly not the case, as these two bugs were recently reported.
1658    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1659        and Detrail to create the HTML documentation, the .txt form of the man
1660        pages, and it removes trailing spaces from listed files. It also creates
1661        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1662        case, it wraps all the #defines with #ifndefs. This script should be run
1663        before "make dist".
1665    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1666        with Unicode property support.
1668        (a) For a maximizing quantifier, if the two different cases of the
1669            character were of different lengths in their UTF-8 codings (there are
1670            some cases like this - I found 11), and the matching function had to
1671            back up over a mixture of the two cases, it incorrectly assumed they
1672            were both the same length.
1674        (b) When PCRE was configured to use the heap rather than the stack for
1675            recursion during matching, it was not correctly preserving the data for
1676            the other case of a UTF-8 character when checking ahead for a match
1677            while processing a minimizing repeat. If the check also involved
1678            matching a wide character, but failed, corruption could cause an
1679            erroneous result when trying to check for a repeat of the original
1680            character.
1682    13. Some tidying changes to the testing mechanism:
1684        (a) The RunTest script now detects the internal link size and whether there
1685            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1686            values substituted by "configure". (The RunGrepTest script already did
1687            this for UTF-8.) The configure.ac script no longer substitutes the
1688            relevant variables.
1690        (b) The debugging options /B and /D in pcretest show the compiled bytecode
1691            with length and offset values. This means that the output is different
1692            for different internal link sizes. Test 2 is skipped for link sizes
1693            other than 2 because of this, bypassing the problem. Unfortunately,
1694            there was also a test in test 3 (the locale tests) that used /B and
1695            failed for link sizes other than 2. Rather than cut the whole test out,
1696            I have added a new /Z option to pcretest that replaces the length and
1697            offset values with spaces. This is now used to make test 3 independent
1698            of link size. (Test 2 will be tidied up later.)
1700    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1701        segmentation fault instead of returning the appropriate error message.
1703    15. In multiline mode when the newline sequence was set to "any", the pattern
1704        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1705        This doesn't seem right; it now treats the CRLF combination as the line
1706        ending, and so does not match in that case. It's only a pattern such as ^$
1707        that would hit this one: something like ^ABC$ would have failed after \r
1708        and then tried again after \r\n.
1710    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1711        in an attempt to make files that differ only in their line terminators
1712        compare equal. This works on Linux.
1714    17. Under certain error circumstances pcregrep might try to free random memory
1715        as it exited. This is now fixed, thanks to valgrind.
1717    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1718        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1719        was because its rules for how to advance for /g after matching an empty
1720        string at the end of a line did not allow for this case. They now check for
1721        it specially.
1723    20. pcretest is supposed to handle patterns and data of any length, by
1724        extending its buffers when necessary. It was getting this wrong when the
1725        buffer for a data line had to be extended.
1727    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1728        CRLF as a newline sequence.
1730    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1731        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1732        I have nevertheless tidied it up.
1734    23. Added some casts to kill warnings from HP-UX ia64 compiler.
1736    24. Added a man page for pcre-config.
1739    Version 7.0 19-Dec-06
1740    ---------------------
1742     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1743        moving to gcc 4.1.1.
1745     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1746        sys/time.h, which is documented as needed for this function. It doesn't
1747        seem to matter on Linux, but it showed up on some releases of OS X.
1749     3. It seems that there are systems where bytes whose values are greater than
1750        127 match isprint() in the "C" locale. The "C" locale should be the
1751        default when a C program starts up. In most systems, only ASCII printing
1752        characters match isprint(). This difference caused the output from pcretest
1753        to vary, making some of the tests fail. I have changed pcretest so that:
1755        (a) When it is outputting text in the compiled version of a pattern, bytes
1756            other than 32-126 are always shown as hex escapes.
1758        (b) When it is outputting text that is a matched part of a subject string,
1759            it does the same, unless a different locale has been set for the match
1760            (using the /L modifier). In this case, it uses isprint() to decide.
1762     4. Fixed a major bug that caused incorrect computation of the amount of memory
1763        required for a compiled pattern when options that changed within the
1764        pattern affected the logic of the preliminary scan that determines the
1765        length. The relevant options are -x, and -i in UTF-8 mode. The result was
1766        that the computed length was too small. The symptoms of this bug were
1767        either the PCRE error "internal error: code overflow" from pcre_compile(),
1768        or a glibc crash with a message such as "pcretest: free(): invalid next
1769        size (fast)". Examples of patterns that provoked this bug (shown in
1770        pcretest format) are:
1772          /(?-x: )/x
1773          /(?x)(?-x: \s*#\s*)/
1774          /((?i)[\x{c0}])/8
1775          /(?i:[\x{c0}])/8
1777        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
1778        is now done differently.
1780     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
1781        wrapper classes; (b) implement a new function in the C++ scanner that is
1782        more efficient than the old way of doing things because it avoids levels of
1783        recursion in the regex matching; (c) add a paragraph to the documentation
1784        for the FullMatch() function.
1786     6. The escape sequence \n was being treated as whatever was defined as
1787        "newline". Not only was this contrary to the documentation, which states
1788        that \n is character 10 (hex 0A), but it also went horribly wrong when
1789        "newline" was defined as CRLF. This has been fixed.
1791     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
1792        was being set to -1 for the "end of line" case (supposedly a value that no
1793        character can have). Though this value is never used (the check for end of
1794        line is "zero bytes in current character"), it caused compiler complaints.
1795        I've changed it to 0xffffffff.
1797     8. In pcre_version.c, the version string was being built by a sequence of
1798        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
1799        string (as it is for production releases) called a macro with an empty
1800        argument. The C standard says the result of this is undefined. The gcc
1801        compiler treats it as an empty string (which was what was wanted) but it is
1802        reported that Visual C gives an error. The source has been hacked around to
1803        avoid this problem.
1805     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
1806        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
1807        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
1808        of them did).
1810    10. Originally, pcretest opened its input and output without "b"; then I was
1811        told that "b" was needed in some environments, so it was added for release
1812        5.0 to both the input and output. (It makes no difference on Unix-like
1813        systems.) Later I was told that it is wrong for the input on Windows. I've
1814        now abstracted the modes into two macros, to make it easier to fiddle with
1815        them, and removed "b" from the input mode under Windows.
1817    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
1819    12. Added -help and --help to pcretest as an official way of being reminded
1820        of the options.
1822    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
1823        and pcrecpp.cc because they annoy compilers at high warning levels.
1825    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
1827    15. Fixed an occurrence of == in configure.ac that should have been = (shell
1828        scripts are not C programs :-) and which was not noticed because it works
1829        on Linux.
1831    16. pcretest is supposed to handle any length of pattern and data line (as one
1832        line or as a continued sequence of lines) by extending its input buffer if
1833        necessary. This feature was broken for very long pattern lines, leading to
1834        a string of junk being passed to pcre_compile() if the pattern was longer
1835        than about 50K.
1837    17. I have done a major re-factoring of the way pcre_compile() computes the
1838        amount of memory needed for a compiled pattern. Previously, there was code
1839        that made a preliminary scan of the pattern in order to do this. That was
1840        OK when PCRE was new, but as the facilities have expanded, it has become
1841        harder and harder to keep it in step with the real compile phase, and there
1842        have been a number of bugs (see for example, 4 above). I have now found a
1843        cunning way of running the real compile function in a "fake" mode that
1844        enables it to compute how much memory it would need, while actually only
1845        ever using a few hundred bytes of working memory and without too many
1846        tests of the mode. This should make future maintenance and development
1847        easier. A side effect of this work is that the limit of 200 on the nesting
1848        depth of parentheses has been removed (though this was never a serious
1849        limitation, I suspect). However, there is a downside: pcre_compile() now
1850        runs more slowly than before (30% or more, depending on the pattern). I
1851        hope this isn't a big issue. There is no effect on runtime performance.
1853    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
1854        newline (only possible for the last line of a file) and it was a
1855        pattern that set a locale (followed by /Lsomething), pcretest crashed.
1857    19. Added additional timing features to pcretest. (1) The -tm option now times
1858        matching only, not compiling. (2) Both -t and -tm can be followed, as a
1859        separate command line item, by a number that specifies the number of
1860        repeats to use when timing. The default is 50000; this gives better
1861        precision, but takes uncomfortably long for very large patterns.
1863    20. Extended pcre_study() to be more clever in cases where a branch of a
1864        subpattern has no definite first character. For example, (a*|b*)[cd] would
1865        previously give no result from pcre_study(). Now it recognizes that the
1866        first character must be a, b, c, or d.
1868    21. There was an incorrect error "recursive call could loop indefinitely" if
1869        a subpattern (or the entire pattern) that was being tested for matching an
1870        empty string contained only one non-empty item after a nested subpattern.
1871        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
1872        incorrectly, because the \d was being skipped in the check.
1874    22. The pcretest program now has a new pattern option /B and a command line
1875        option -b, which is equivalent to adding /B to every pattern. This causes
1876        it to show the compiled bytecode, without the additional information that
1877        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
1878        is the same as /B/I).
1880    23. A new optimization is now able automatically to treat some sequences such
1881        as a*b as a*+b. More specifically, if something simple (such as a character
1882        or a simple class like \d) has an unlimited quantifier, and is followed by
1883        something that cannot possibly match the quantified thing, the quantifier
1884        is automatically "possessified".
1886    24. A recursive reference to a subpattern whose number was greater than 39
1887        went wrong under certain circumstances in UTF-8 mode. This bug could also
1888        have affected the operation of pcre_study().
1890    25. Realized that a little bit of performance could be had by replacing
1891        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
1893    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
1895    27. Possessive quantifiers such as a++ were previously implemented by turning
1896        them into atomic groups such as ($>a+). Now they have their own opcodes,
1897        which improves performance. This includes the automatically created ones
1898        from 23 above.
1900    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
1901        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
1902        the first matched character to be a colon. This applied both to named and
1903        numbered groups.
1905    29. The ucpinternal.h header file was missing its idempotency #ifdef.
1907    30. I was sent a "project" file called libpcre.a.dev which I understand makes
1908        building PCRE on Windows easier, so I have included it in the distribution.
1910    31. There is now a check in pcretest against a ridiculously large number being
1911        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
1912        loop, the loop is abandoned.
1914    32. Forward references to subpatterns in conditions such as (?(2)...) where
1915        subpattern 2 is defined later cause pcre_compile() to search forwards in
1916        the pattern for the relevant set of parentheses. This search went wrong
1917        when there were unescaped parentheses in a character class, parentheses
1918        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
1920    33. "Subroutine" calls and backreferences were previously restricted to
1921        referencing subpatterns earlier in the regex. This restriction has now
1922        been removed.
1924    34. Added a number of extra features that are going to be in Perl 5.10. On the
1925        whole, these are just syntactic alternatives for features that PCRE had
1926        previously implemented using the Python syntax or my own invention. The
1927        other formats are all retained for compatibility.
1929        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
1930            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
1931            also .NET compatible.
1933        (b) A recursion or subroutine call to a named group can now be defined as
1934            (?&name) as well as (?P>name).
1936        (c) A backreference to a named group can now be defined as \k<name> or
1937            \k'name' as well as (?P=name). The new forms, as well as being in Perl
1938            5.10, are also .NET compatible.
1940        (d) A conditional reference to a named group can now use the syntax
1941            (?(<name>) or (?('name') as well as (?(name).
1943        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
1944            groups (named and numbered) that are never evaluated inline, but can be
1945            called as "subroutines" from elsewhere. In effect, the DEFINE condition
1946            is always false. There may be only one alternative in such a group.
1948        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
1949            as the simple (?(R). The condition is true only if the most recent
1950            recursion is that of the given number or name. It does not search out
1951            through the entire recursion stack.
1953        (g) The escape \gN or \g{N} has been added, where N is a positive or
1954            negative number, specifying an absolute or relative reference.
1956    35. Tidied to get rid of some further signed/unsigned compiler warnings and
1957        some "unreachable code" warnings.
1959    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
1960        things, this adds five new scripts.
1962    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
1963        There were also incompatibilities regarding the handling of \Q..\E inside
1964        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
1965        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
1967    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
1968        matches an empty string, and forcibly breaks the loop. There were bugs in
1969        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
1970        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
1971        separate and independent bugs (that affected different cases) have been
1972        fixed.
1974    39. Refactored the code to abolish the use of different opcodes for small
1975        capturing bracket numbers. This is a tidy that I avoided doing when I
1976        removed the limit on the number of capturing brackets for 3.5 back in 2001.
1977        The new approach is not only tidier, it makes it possible to reduce the
1978        memory needed to fix the previous bug (38).
1980    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
1981        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
1982        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
1983        mode.
1985    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
1986        report.
1988    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
1989        copy construction and assignment in the C++ wrapper.
1991    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
1992        couple of bugs that could have given wrong results in the "--newline=crlf"
1993        case.
1995    44. Added a number of casts and did some reorganization of signed/unsigned int
1996        variables following suggestions from Dair Grant. Also renamed the variable
1997        "this" as "item" because it is a C++ keyword.
1999    45. Arranged for dftables to add
2001          #include "pcre_internal.h"
2003        to pcre_chartables.c because without it, gcc 4.x may remove the array
2004        definition from the final binary if PCRE is built into a static library and
2005        dead code stripping is activated.
2007    46. For an unanchored pattern, if a match attempt fails at the start of a
2008        newline sequence, and the newline setting is CRLF or ANY, and the next two
2009        characters are CRLF, advance by two characters instead of one.
2012    Version 6.7 04-Jul-06
2013    ---------------------
2015     1. In order to handle tests when input lines are enormously long, pcretest has
2016        been re-factored so that it automatically extends its buffers when
2017        necessary. The code is crude, but this _is_ just a test program. The
2018        default size has been increased from 32K to 50K.
2020     2. The code in pcre_study() was using the value of the re argument before
2021        testing it for NULL. (Of course, in any sensible call of the function, it
2022        won't be NULL.)
2024     3. The memmove() emulation function in pcre_internal.h, which is used on
2025        systems that lack both memmove() and bcopy() - that is, hardly ever -
2026        was missing a "static" storage class specifier.
2028     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2029        containing an extended class (one that cannot be represented by a bitmap
2030        because it contains high-valued characters or Unicode property items, e.g.
2031        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2032        pattern, but PCRE should not loop if you do not (it no longer does).
2033        [Detail: two cases were found: (a) a repeated subpattern containing an
2034        extended class; (b) a recursive reference to a subpattern that followed a
2035        previous extended class. It wasn't skipping over the extended class
2036        correctly when UTF-8 mode was not set.]
2038     5. A negated single-character class was not being recognized as fixed-length
2039        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2040        compile error "lookbehind assertion is not fixed length".
2042     6. The RunPerlTest auxiliary script was showing an unexpected difference
2043        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2044        write a Perl script that can interpret lines of an input file either as
2045        byte characters or as UTF-8, which is what "perltest" was being required to
2046        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2047        can't do is switch easily at run time between having the "use utf8;" pragma
2048        or not. In the end, I fudged it by using the RunPerlTest script to insert
2049        "use utf8;" explicitly for the UTF-8 tests.
2051     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2052        the end of the subject string, contrary to the documentation and to what
2053        Perl does. This was true of both matching functions. Now it matches only at
2054        the start of the subject and immediately after *internal* newlines.
2056     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2057        a pointer to an int instead of a pointer to an unsigned long int. This
2058        caused problems on 64-bit systems.
2060     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2061        instance of the 'standard' template library not being so standard".
2063    10. There was no check on the number of named subpatterns nor the maximum
2064        length of a subpattern name. The product of these values is used to compute
2065        the size of the memory block for a compiled pattern. By supplying a very
2066        long subpattern name and a large number of named subpatterns, the size
2067        computation could be caused to overflow. This is now prevented by limiting
2068        the length of names to 32 characters, and the number of named subpatterns
2069        to 10,000.
2071    11. Subpatterns that are repeated with specific counts have to be replicated in
2072        the compiled pattern. The size of memory for this was computed from the
2073        length of the subpattern and the repeat count. The latter is limited to
2074        65535, but there was no limit on the former, meaning that integer overflow
2075        could in principle occur. The compiled length of a repeated subpattern is
2076        now limited to 30,000 bytes in order to prevent this.
2078    12. Added the optional facility to have named substrings with the same name.
2080    13. Added the ability to use a named substring as a condition, using the
2081        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2082        are numbers (not recommended). Forward references are permitted.
2084    14. Added forward references in named backreferences (if you see what I mean).
2086    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2087        pattern could run off the end of the subject. For example, the pattern
2088        "(?s)(.{1,5})"8 did this with the subject "ab".
2090    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2091        PCRE_CASELESS was set when matching characters that were quantified with ?
2092        or *.
2094    17. A character class other than a single negated character that had a minimum
2095        but no maximum quantifier - for example [ab]{6,} - was not handled
2096        correctly by pce_dfa_exec(). It would match only one character.
2098    18. A valid (though odd) pattern that looked like a POSIX character
2099        class but used an invalid character after [ (for example [[,abc,]]) caused
2100        pcre_compile() to give the error "Failed: internal error: code overflow" or
2101        in some cases to crash with a glibc free() error. This could even happen if
2102        the pattern terminated after [[ but there just happened to be a sequence of
2103        letters, a binary zero, and a closing ] in the memory that followed.
2105    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2106        over the years. Originally (before any Unicode support), just the bottom 8
2107        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2108        output from "man perlunicode" includes this:
2110          The regular expression compiler produces polymorphic opcodes.  That
2111          is, the pattern adapts to the data and automatically switches to
2112          the Unicode character scheme when presented with Unicode data--or
2113          instead uses a traditional byte scheme when presented with byte
2114          data.
2116        Sadly, a wide octal escape does not cause a switch, and in a string with
2117        no other multibyte characters, these octal escapes are treated as before.
2118        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2119        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2120        Unicode string.
2122        I have not perpetrated such confusion in PCRE. Up till now, it took just
2123        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2124        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2125        translate to the appropriate multibyte character.
2127    29. Applied some refactoring to reduce the number of warnings from Microsoft
2128        and Borland compilers. This has included removing the fudge introduced
2129        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2130        a warning about an unused variable.
2132    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2133        characters since release 4.0, because Perl (from release 5.004) does not.
2134        [Or at least, is documented not to: some releases seem to be in conflict
2135        with the documentation.] However, when a pattern was studied with
2136        pcre_study() and all its branches started with \s, PCRE still included VT
2137        as a possible starting character. Of course, this did no harm; it just
2138        caused an unnecessary match attempt.
2140    22. Removed a now-redundant internal flag bit that recorded the fact that case
2141        dependency changed within the pattern. This was once needed for "required
2142        byte" processing, but is no longer used. This recovers a now-scarce options
2143        bit. Also moved the least significant internal flag bit to the most-
2144        significant bit of the word, which was not previously used (hangover from
2145        the days when it was an int rather than a uint) to free up another bit for
2146        the future.
2148    23. Added support for CRLF line endings as well as CR and LF. As well as the
2149        default being selectable at build time, it can now be changed at runtime
2150        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2151        specify that it is scanning data with non-default line endings.
2153    24. Changed the definition of CXXLINK to make it agree with the definition of
2154        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2156    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2157        recursions. This makes a big different to stack usage for some patterns.
2159    26. If a subpattern containing a named recursion or subroutine reference such
2160        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2161        the space required for the compiled pattern went wrong and gave too small a
2162        value. Depending on the environment, this could lead to "Failed: internal
2163        error: code overflow at offset 49" or "glibc detected double free or
2164        corruption" errors.
2166    27. Applied patches from Google (a) to support the new newline modes and (b) to
2167        advance over multibyte UTF-8 characters in GlobalReplace.
2169    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2170        difference for some implementation of PCRE in some Windows version.
2172    29. Added some extra testing facilities to pcretest:
2174        \q<number>   in a data line sets the "match limit" value
2175        \Q<number>   in a data line sets the "match recursion limt" value
2176        -S <number>  sets the stack size, where <number> is in megabytes
2178        The -S option isn't available for Windows.
2181    Version 6.6 06-Feb-06
2182    ---------------------
2184     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2185        in pcreposix.h. I have copied the definition from pcre.h.
2187     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2188        because pcre.h is no longer a built file.
2190     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2191        not normally included in the compiled code.
2194    Version 6.5 01-Feb-06
2195    ---------------------
2197     1. When using the partial match feature with pcre_dfa_exec(), it was not
2198        anchoring the second and subsequent partial matches at the new starting
2199        point. This could lead to incorrect results. For example, with the pattern
2200        /1234/, partially matching against "123" and then "a4" gave a match.
2202     2. Changes to pcregrep:
2204        (a) All non-match returns from pcre_exec() were being treated as failures
2205            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2206            error message is output. Some extra information is given for the
2207            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2208            probably the only errors that are likely to be caused by users (by
2209            specifying a regex that has nested indefinite repeats, for instance).
2210            If there are more than 20 of these errors, pcregrep is abandoned.
2212        (b) A binary zero was treated as data while matching, but terminated the
2213            output line if it was written out. This has been fixed: binary zeroes
2214            are now no different to any other data bytes.
2216        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2217            used to set a locale for matching. The --locale=xxxx long option has
2218            been added (no short equivalent) to specify a locale explicitly on the
2219            pcregrep command, overriding the environment variables.
2221        (d) When -B was used with -n, some line numbers in the output were one less
2222            than they should have been.
2224        (e) Added the -o (--only-matching) option.
2226        (f) If -A or -C was used with -c (count only), some lines of context were
2227            accidentally printed for the final match.
2229        (g) Added the -H (--with-filename) option.
2231        (h) The combination of options -rh failed to suppress file names for files
2232            that were found from directory arguments.
2234        (i) Added the -D (--devices) and -d (--directories) options.
2236        (j) Added the -F (--fixed-strings) option.
2238        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2240        (l) Added the --colo(u)r option.
2242        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2243            is not present by default.
2245     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2246        items such as (?R) or (?1), when the recursion could match a number of
2247        alternatives. If it matched one of the alternatives, but subsequently,
2248        outside the recursion, there was a failure, the code tried to back up into
2249        the recursion. However, because of the way PCRE is implemented, this is not
2250        possible, and the result was an incorrect result from the match.
2252        In order to prevent this happening, the specification of recursion has
2253        been changed so that all such subpatterns are automatically treated as
2254        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2256     4. I had overlooked the fact that, in some locales, there are characters for
2257        which isalpha() is true but neither isupper() nor islower() are true. In
2258        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2259        and ordfeminine) are like this. This affected the treatment of \w and \W
2260        when they appeared in character classes, but not when they appeared outside
2261        a character class. The bit map for "word" characters is now created
2262        separately from the results of isalnum() instead of just taking it from the
2263        upper, lower, and digit maps. (Plus the underscore character, of course.)
2265     5. The above bug also affected the handling of POSIX character classes such as
2266        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2267        permanent tables. Instead, the bit maps for such a class were previously
2268        created as the appropriate unions of the upper, lower, and digit bitmaps.
2269        Now they are created by subtraction from the [[:word:]] class, which has
2270        its own bitmap.
2272     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2273        It is created by subtracting the vertical space characters (\x09, \x0a,
2274        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2275        subtraction was done in the overall bitmap for a character class, meaning
2276        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2277        be recognized. This bug has been fixed.
2279     7. Patches from the folks at Google:
2281          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2282          real life, but is still worth protecting against".
2284          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2285          regular expressions".
2287          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2288          have it.
2290          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2291          "configure" and the latter not, in order to fix a problem somebody had
2292          with compiling the Arg class on HP-UX.
2294          (e) Improve the error-handling of the C++ wrapper a little bit.
2296          (f) New tests for checking recursion limiting.
2298     8. The pcre_memmove() function, which is used only if the environment does not
2299        have a standard memmove() function (and is therefore rarely compiled),
2300        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2301        returning a result (though PCRE never actually uses the result).
2303     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2304        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2305        returned instead of calling malloc() with an overflowing number that would
2306        most likely cause subsequent chaos.
2308    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2310    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2311        with this option is matched, the nmatch and pmatch options of regexec() are
2312        ignored.
2314    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2315        provided in case anyone wants to the the POSIX interface with UTF-8
2316        strings.
2318    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2319        C++ linking (needed for some HP-UX environments).
2321    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2322        (unused parameter) and in the pcre_printint() function (omitted "default"
2323        switch label when the default is to do nothing).
2325    15. Added some code to make it possible, when PCRE is compiled as a C++
2326        library, to replace subject pointers for pcre_exec() with a smart pointer
2327        class, thus making it possible to process discontinuous strings.
2329    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2330        much the same function. They were added by different people who were trying
2331        to make PCRE easy to compile on non-Unix systems. It has been suggested
2332        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2333        for compiling on Windows systems. I have therefore replaced it with
2334        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2335        defaults to "extern" for C or "extern C" for C++, which works fine on
2336        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2337        SCOPE with something explicit in config.h. In addition:
2339        (a) pcreposix.h still had just "extern" instead of either of these macros;
2340            I have replaced it with PCRE_DATA_SCOPE.
2342        (b) Functions such as _pcre_xclass(), which are internal to the library,
2343            but external in the C sense, all had PCRE_EXPORT in their definitions.
2344            This is apparently wrong for the Windows case, so I have removed it.
2345            (It makes no difference on Unix-like systems.)
2347    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2348        of recursive calls to match(). This is different to MATCH_LIMIT because
2349        that limits the total number of calls to match(), not all of which increase
2350        the depth of recursion. Limiting the recursion depth limits the amount of
2351        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2352        when PCRE is compiled, and changed at run time. A patch from Google adds
2353        this functionality to the C++ interface.
2355    18. Changes to the handling of Unicode character properties:
2357        (a) Updated the table to Unicode 4.1.0.
2359        (b) Recognize characters that are not in the table as "Cn" (undefined).
2361        (c) I revised the way the table is implemented to a much improved format
2362            which includes recognition of ranges. It now supports the ranges that
2363            are defined in UnicodeData.txt, and it also amalgamates other
2364            characters into ranges. This has reduced the number of entries in the
2365            table from around 16,000 to around 3,000, thus reducing its size
2366            considerably. I realized I did not need to use a tree structure after
2367            all - a binary chop search is just as efficient. Having reduced the
2368            number of entries, I extended their size from 6 bytes to 8 bytes to
2369            allow for more data.
2371        (d) Added support for Unicode script names via properties such as \p{Han}.
2373    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2374        matching that character.
2376    20. When matching a repeated Unicode property with a minimum greater than zero,
2377        (for example \pL{2,}), PCRE could look past the end of the subject if it
2378        reached it while seeking the minimum number of characters. This could
2379        happen only if some of the characters were more than one byte long, because
2380        there is a check for at least the minimum number of bytes.
2382    21. Refactored the implementation of \p and \P so as to be more general, to
2383        allow for more different types of property in future. This has changed the
2384        compiled form incompatibly. Anybody with saved compiled patterns that use
2385        \p or \P will have to recompile them.
2387    22. Added "Any" and "L&" to the supported property types.
2389    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2390        but give a compile time error if the value is greater than 0xff.
2392    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2393        accidentally not being installed or uninstalled.
2395    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2396        made were to insert the current release number. This seemed silly, because
2397        it made things harder for people building PCRE on systems that don't run
2398        "configure". I have turned pcre.h into a distributed file, no longer built
2399        by "configure", with the version identification directly included. There is
2400        no longer a pcre.h.in file.
2402        However, this change necessitated a change to the pcre-config script as
2403        well. It is built from pcre-config.in, and one of the substitutions was the
2404        release number. I have updated configure.ac so that ./configure now finds
2405        the release number by grepping pcre.h.
2407    26. Added the ability to run the tests under valgrind.
2410    Version 6.4 05-Sep-05
2411    ---------------------
2413     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2414        "--" to be printed when multiple files were scanned, even when none of the
2415        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2416        consider it to be a bug, and have restored the previous behaviour.
2418     2. A couple of code tidies to get rid of compiler warnings.
2420     3. The pcretest program used to cheat by referring to symbols in the library
2421        whose names begin with _pcre_. These are internal symbols that are not
2422        really supposed to be visible externally, and in some environments it is
2423        possible to suppress them. The cheating is now confined to including
2424        certain files from the library's source, which is a bit cleaner.
2426     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2427        file's purpose clearer.
2429     5. Reorganized pcre_ucp_findchar().
2432    Version 6.3 15-Aug-05
2433    ---------------------
2435     1. The file libpcre.pc.in did not have general read permission in the tarball.
2437     2. There were some problems when building without C++ support:
2439        (a) If C++ support was not built, "make install" and "make test" still
2440            tried to test it.
2442        (b) There were problems when the value of CXX was explicitly set. Some
2443            changes have been made to try to fix these, and ...
2445        (c) --disable-cpp can now be used to explicitly disable C++ support.
2447        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2448            backslash in a target when C++ was disabled. This confuses some
2449            versions of "make", apparently. Using an intermediate variable solves
2450            this. (Same for CPP_LOBJ.)
2452     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2453        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2454        necessary on certain architectures.
2456     4. Added a setting of -export-symbols-regex to the link command to remove
2457        those symbols that are exported in the C sense, but actually are local
2458        within the library, and not documented. Their names all begin with
2459        "_pcre_". This is not a perfect job, because (a) we have to except some
2460        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2461        available (and never for static libraries). I have made a note to try to
2462        find a way round (a) in the future.
2465    Version 6.2 01-Aug-05
2466    ---------------------
2468     1. There was no test for integer overflow of quantifier values. A construction
2469        such as {1111111111111111} would give undefined results. What is worse, if
2470        a minimum quantifier for a parenthesized subpattern overflowed and became
2471        negative, the calculation of the memory size went wrong. This could have
2472        led to memory overwriting.
2474     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2476     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2477        operating environments where this matters.
2479     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2480        PCRE options from within the C++ wrapper.
2482     5. Named capturing subpatterns were not being correctly counted when a pattern
2483        was compiled. This caused two problems: (a) If there were more than 100
2484        such subpatterns, the calculation of the memory needed for the whole
2485        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2486        back references of the form \12, where the number was greater than 9, were
2487        not recognized as back references, even though there were sufficient
2488        previous subpatterns.
2490     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2491        versions of gcc, e.g. 2.95.4.
2494    Version 6.1 21-Jun-05
2495    ---------------------
2497     1. There was one reference to the variable "posix" in pcretest.c that was not
2498        surrounded by "#if !defined NOPOSIX".
2500     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2501        the cross-check on the old pcre_info() function, for the benefit of the
2502        cut-down version of PCRE that is currently imported into Exim.
2504     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2505        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2506        patterns that start (?i) but I don't think that matters. The use of (?i) is
2507        just an example; this all applies to the other options as well.
2509     4. Since libtool seems to echo the compile commands it is issuing, the output
2510        from "make" can be reduced a bit by putting "@" in front of each libtool
2511        compile command.
2513     5. Patch from the folks at Google for configure.in to be a bit more thorough
2514        in checking for a suitable C++ installation before trying to compile the
2515        C++ stuff. This should fix a reported problem when a compiler was present,
2516        but no suitable headers.
2518     6. The man pages all had just "PCRE" as their title. I have changed them to
2519        be the relevant file name. I have also arranged that these names are
2520        retained in the file doc/pcre.txt, which is a concatenation in text format
2521        of all the man pages except the little individual ones for each function.
2523     7. The NON-UNIX-USE file had not been updated for the different set of source
2524        files that come with release 6. I also added a few comments about the C++
2525        wrapper.
2528    Version 6.0 07-Jun-05
2529    ---------------------
2531     1. Some minor internal re-organization to help with my DFA experiments.
2533     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2534        didn't matter for the library itself when fully configured, but did matter
2535        when compiling without UCP support, or within Exim, where the ucp files are
2536        not imported.
2538     3. Refactoring of the library code to split up the various functions into
2539        different source modules. The addition of the new DFA matching code (see
2540        below) to a single monolithic source would have made it really too
2541        unwieldy, quite apart from causing all the code to be include in a
2542        statically linked application, when only some functions are used. This is
2543        relevant even without the DFA addition now that patterns can be compiled in
2544        one application and matched in another.
2546        The downside of splitting up is that there have to be some external
2547        functions and data tables that are used internally in different modules of
2548        the library but which are not part of the API. These have all had their
2549        names changed to start with "_pcre_" so that they are unlikely to clash
2550        with other external names.
2552     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2553        a different (DFA) algorithm. Although it is slower than the original
2554        function, it does have some advantages for certain types of matching
2555        problem.
2557     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2558        including restarting after a partial match.
2560     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2561        defined when compiling for Windows was sent to me. I have put it into the
2562        code, though I have no means of testing or verifying it.
2564     7. Added the pcre_refcount() auxiliary function.
2566     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2567        match before or at the first newline in the subject string. In pcretest,
2568        the /f option on a pattern can be used to set this.
2570     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2571        would behave wrongly. This has been present in PCRE since release 4.0.
2573    10. A number of changes to the pcregrep command:
2575        (a) Refactored how -x works; insert ^(...)$ instead of setting
2576            PCRE_ANCHORED and checking the length, in preparation for adding
2577            something similar for -w.
2579        (b) Added the -w (match as a word) option.
2581        (c) Refactored the way lines are read and buffered so as to have more
2582            than one at a time available.
2584        (d) Implemented a pcregrep test script.
2586        (e) Added the -M (multiline match) option. This allows patterns to match
2587            over several lines of the subject. The buffering ensures that at least
2588            8K, or the rest of the document (whichever is the shorter) is available
2589            for matching (and similarly the previous 8K for lookbehind assertions).
2591        (f) Changed the --help output so that it now says
2593              -w, --word-regex(p)
2595            instead of two lines, one with "regex" and the other with "regexp"
2596            because that confused at least one person since the short forms are the
2597            same. (This required a bit of code, as the output is generated
2598            automatically from a table. It wasn't just a text change.)
2600        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2601            option but starts with a hyphen. Could be a pattern or a path name
2602            starting with a hyphen, for instance.
2604        (h) "-" can be given as a file name to represent stdin.
2606        (i) When file names are being printed, "(standard input)" is used for
2607            the standard input, for compatibility with GNU grep. Previously
2608            "<stdin>" was used.
2610        (j) The option --label=xxx can be used to supply a name to be used for
2611            stdin when file names are being printed. There is no short form.
2613        (k) Re-factored the options decoding logic because we are going to add
2614            two more options that take data. Such options can now be given in four
2615            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2617        (l) Added the -A, -B, and -C options for requesting that lines of context
2618            around matches be printed.
2620        (m) Added the -L option to print the names of files that do not contain
2621            any matching lines, that is, the complement of -l.
2623        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2624            continue to scan other files.
2626        (o) The -s option was incorrectly implemented. For compatibility with other
2627            greps, it now suppresses the error message for a non-existent or non-
2628            accessible file (but not the return code). There is a new option called
2629            -q that suppresses the output of matching lines, which was what -s was
2630            previously doing.
2632        (p) Added --include and --exclude options to specify files for inclusion
2633            and exclusion when recursing.
2635    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2636        Hopefully, it now does.
2638    12. Missing cast in pcre_study().
2640    13. Added an "uninstall" target to the makefile.
2642    14. Replaced "extern" in the function prototypes in Makefile.in with
2643        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2644        world, but is set differently for Windows.
2646    15. Added a second compiling function called pcre_compile2(). The only
2647        difference is that it has an extra argument, which is a pointer to an
2648        integer error code. When there is a compile-time failure, this is set
2649        non-zero, in addition to the error test pointer being set to point to an
2650        error message. The new argument may be NULL if no error number is required
2651        (but then you may as well call pcre_compile(), which is now just a
2652        wrapper). This facility is provided because some applications need a
2653        numeric error indication, but it has also enabled me to tidy up the way
2654        compile-time errors are handled in the POSIX wrapper.
2656    16. Added VPATH=.libs to the makefile; this should help when building with one
2657        prefix path and installing with another. (Or so I'm told by someone who
2658        knows more about this stuff than I do.)
2660    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2661        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2662        match everything, including newlines. This is not POSIX-compatible, but
2663        somebody wanted the feature. From pcretest it can be activated by using
2664        both the P and the s flags.
2666    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2668    19. libpcre.pc was being incorrectly installed as executable.
2670    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2671        it now also looks for '\r' so that it will work unmodified on Windows.
2673    21. Added Google's contributed C++ wrapper to the distribution.
2675    22. Added some untidy missing memory free() calls in pcretest, to keep
2676        Electric Fence happy when testing.
2680    Version 5.0 13-Sep-04
2681    ---------------------
2683     1. Internal change: literal characters are no longer packed up into items
2684        containing multiple characters in a single byte-string. Each character
2685        is now matched using a separate opcode. However, there may be more than one
2686        byte in the character in UTF-8 mode.
2688     2. The pcre_callout_block structure has two new fields: pattern_position and
2689        next_item_length. These contain the offset in the pattern to the next match
2690        item, and its length, respectively.
2692     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2693        insertion of callouts before each pattern item. Added the /C option to
2694        pcretest to make use of this.
2696     4. On the advice of a Windows user, the lines
2698          #if defined(_WIN32) || defined(WIN32)
2699          _setmode( _fileno( stdout ), 0x8000 );
2700          #endif  /* defined(_WIN32) || defined(WIN32) */
2702        have been added to the source of pcretest. This apparently does useful
2703        magic in relation to line terminators.
2705     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2706        for the benefit of those environments where the "b" makes a difference.
2708     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2709        to know about it. I have put a hack into configure.in that adds in code
2710        to set GCC=yes if CC=icc. This seems to end up at a point in the
2711        generated configure script that is early enough to affect the setting of
2712        compiler options, which is what is needed, but I have no means of testing
2713        whether it really works. (The user who reported this had patched the
2714        generated configure script, which of course I cannot do.)
2716        LATER: After change 22 below (new libtool files), the configure script
2717        seems to know about icc (and also ecc). Therefore, I have commented out
2718        this hack in configure.in.
2720     7. Added support for pkg-config (2 patches were sent in).
2722     8. Negated POSIX character classes that used a combination of internal tables
2723        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2724        [[:^ascii]]. Typically, they would match almost any characters. The other
2725        POSIX classes were not broken in this way.
2727     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2728        to find the match, as PCRE was deluded into thinking that the match had to
2729        start at the start point or following a newline. The same bug applied to
2730        patterns with negative forward assertions or any backward assertions
2731        preceding ".*" at the start, unless the pattern required a fixed first
2732        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2734    10. In UTF-8 mode, when moving forwards in the subject after a failed match
2735        starting at the last subject character, bytes beyond the end of the subject
2736        string were read.
2738    11. Renamed the variable "class" as "classbits" to make life easier for C++
2739        users. (Previously there was a macro definition, but it apparently wasn't
2740        enough.)
2742    12. Added the new field "tables" to the extra data so that tables can be passed
2743        in at exec time, or the internal tables can be re-selected. This allows
2744        a compiled regex to be saved and re-used at a later time by a different
2745        program that might have everything at different addresses.
2747    13. Modified the pcre-config script so that, when run on Solaris, it shows a
2748        -R library as well as a -L library.
2750    14. The debugging options of pcretest (-d on the command line or D on a
2751        pattern) showed incorrect output for anything following an extended class
2752        that contained multibyte characters and which was followed by a quantifier.
2754    15. Added optional support for general category Unicode character properties
2755        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
2756        support. It adds about 90K to the size of the library. The meanings of the
2757        inbuilt class escapes such as \d and \s have NOT been changed.
2759    16. Updated pcredemo.c to include calls to free() to release the memory for the
2760        compiled pattern.
2762    17. The generated file chartables.c was being created in the source directory
2763        instead of in the building directory. This caused the build to fail if the
2764        source directory was different from the building directory, and was
2765        read-only.
2767    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
2768        file. No doubt somebody will tell me if they don't make sense... Also added
2769        Dan Mooney's comments about building on OpenVMS.
2771    19. Added support for partial matching via the PCRE_PARTIAL option for
2772        pcre_exec() and the \P data escape in pcretest.
2774    20. Extended pcretest with 3 new pattern features:
2776        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
2777              write the compiled pattern to the file whose name is "rest-of-line".
2778              This is a straight binary dump of the data, with the saved pointer to
2779              the character tables forced to be NULL. The study data, if any, is
2780              written too. After writing, pcretest reads a new pattern.
2782        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
2783              compiled pattern from the given file. There must not be any
2784              occurrences of "<" in the file name (pretty unlikely); if there are,
2785              pcretest will instead treat the initial "<" as a pattern delimiter.
2786              After reading in the pattern, pcretest goes on to read data lines as
2787              usual.
2789        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
2790              and 16-bit fields in a compiled pattern, to simulate a pattern that
2791              was compiled on a host of opposite endianness.
2793    21. The pcre-exec() function can now cope with patterns that were compiled on
2794        hosts of opposite endianness, with this restriction:
2796          As for any compiled expression that is saved and used later, the tables
2797          pointer field cannot be preserved; the extra_data field in the arguments
2798          to pcre_exec() should be used to pass in a tables address if a value
2799          other than the default internal tables were used at compile time.
2801    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
2802        now diagnosed as an error. Previously, most of the time, a negative number
2803        would have been treated as zero, but if in addition "ovector" was passed as
2804        NULL, a crash could occur.
2806    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
2807        new versions from the libtool 1.5 distribution (the last one is a copy of
2808        a file called libtool.m4). This seems to have fixed the need to patch
2809        "configure" to support Darwin 1.3 (which I used to do). However, I still
2810        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
2811        workstation).
2813    24. Changed the PCRE licence to be the more standard "BSD" licence.
2816    Version 4.5 01-Dec-03
2817    ---------------------
2819     1. There has been some re-arrangement of the code for the match() function so
2820        that it can be compiled in a version that does not call itself recursively.
2821        Instead, it keeps those local variables that need separate instances for
2822        each "recursion" in a frame on the heap, and gets/frees frames whenever it
2823        needs to "recurse". Keeping track of where control must go is done by means
2824        of setjmp/longjmp. The whole thing is implemented by a set of macros that
2825        hide most of the details from the main code, and operates only if
2826        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
2827        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
2828        operating.
2830        To make it easier for callers to provide specially tailored get/free
2831        functions for this usage, two new functions, pcre_stack_malloc, and
2832        pcre_stack_free, are used. They are always called in strict stacking order,
2833        and the size of block requested is always the same.
2835        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
2836        PCRE has been compiled to use the stack or the heap for recursion. The
2837        -C option of pcretest uses this to show which version is compiled.
2839        A new data escape \S, is added to pcretest; it causes the amounts of store
2840        obtained and freed by both kinds of malloc/free at match time to be added
2841        to the output.
2843     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
2844        what's available on my current Linux desktop machine.
2846     3. When matching a UTF-8 string, the test for a valid string at the start has
2847        been extended. If start_offset is not zero, PCRE now checks that it points
2848        to a byte that is the start of a UTF-8 character. If not, it returns
2849        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
2850        this is necessary because there may be backward assertions in the pattern.
2851        When matching the same subject several times, it may save resources to use
2852        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
2854     4. The code for checking the validity of UTF-8 strings has been tightened so
2855        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
2856        containing "overlong sequences".
2858     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
2859        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
2860        should have been "&", but it just so happened that all the cases this let
2861        through by mistake were picked up later in the function.
2863     6. I had used a variable called "isblank" - this is a C99 function, causing
2864        some compilers to warn. To avoid this, I renamed it (as "blankclass").
2866     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
2867        prompting; (b) run "./pcretest /dev/null" at the start of the test script
2868        so the version is shown; (c) stop "make test" echoing "./RunTest".
2870     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
2872     9. The prototype for memmove() for systems that don't have it was using
2873        size_t, but the inclusion of the header that defines size_t was later. I've
2874        moved the #includes for the C headers earlier to avoid this.
2876    10. Added some adjustments to the code to make it easier to compiler on certain
2877        special systems:
2879          (a) Some "const" qualifiers were missing.
2880          (b) Added the macro EXPORT before all exported functions; by default this
2881              is defined to be empty.
2882          (c) Changed the dftables auxiliary program (that builds chartables.c) so
2883              that it reads its output file name as an argument instead of writing
2884              to the standard output and assuming this can be redirected.
2886    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
2887        class containing characters with values greater than 255, PCRE compilation
2888        went into a loop.
2890    12. A recursive reference to a subpattern that was within another subpattern
2891        that had a minimum quantifier of zero caused PCRE to crash. For example,
2892        (x(y(?2))z)? provoked this bug with a subject that got as far as the
2893        recursion. If the recursively-called subpattern itself had a zero repeat,
2894        that was OK.
2896    13. In pcretest, the buffer for reading a data line was set at 30K, but the
2897        buffer into which it was copied (for escape processing) was still set at
2898        1024, so long lines caused crashes.
2900    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
2901        "internal error: code overflow...". This applied to any character class
2902        that was followed by a possessive quantifier.
2904    15. Modified the Makefile to add libpcre.la as a prerequisite for
2905        libpcreposix.la because I was told this is needed for a parallel build to
2906        work.
2908    16. If a pattern that contained .* following optional items at the start was
2909        studied, the wrong optimizing data was generated, leading to matching
2910        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
2911        matching string must start with a or b or c. The correct conclusion for
2912        this pattern is that a match can start with any character.
2915    Version 4.4 13-Aug-03
2916    ---------------------
2918     1. In UTF-8 mode, a character class containing characters with values between
2919        127 and 255 was not handled correctly if the compiled pattern was studied.
2920        In fixing this, I have also improved the studying algorithm for such
2921        classes (slightly).
2923     2. Three internal functions had redundant arguments passed to them. Removal
2924        might give a very teeny performance improvement.
2926     3. Documentation bug: the value of the capture_top field in a callout is *one
2927        more than* the number of the hightest numbered captured substring.
2929     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
2930        in incorrectly linking with a previously installed version. They now link
2931        explicitly with libpcre.la.
2933     5. configure.in no longer needs to recognize Cygwin specially.
2935     6. A problem in pcre.in for Windows platforms is fixed.
2937     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
2938        pcretest, it used to include the size of the study block as part of its
2939        output. Unfortunately, the structure contains a field that has a different
2940        size on different hardware architectures. This meant that the tests that
2941        showed this size failed. As the block is currently always of a fixed size,
2942        this information isn't actually particularly useful in pcretest output, so
2943        I have just removed it.
2945     8. Three pre-processor statements accidentally did not start in column 1.
2946        Sadly, there are *still* compilers around that complain, even though
2947        standard C has not required this for well over a decade. Sigh.
2949     9. In pcretest, the code for checking callouts passed small integers in the
2950        callout_data field, which is a void * field. However, some picky compilers
2951        complained about the casts involved for this on 64-bit systems. Now
2952        pcretest passes the address of the small integer instead, which should get
2953        rid of the warnings.
2955    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
2956        both compile and run time, and gives an error if an invalid UTF-8 sequence
2957        is found. There is a option for disabling this check in cases where the
2958        string is known to be correct and/or the maximum performance is wanted.
2960    11. In response to a bug report, I changed one line in Makefile.in from
2962            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
2963        to
2964            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
2966        to look similar to other lines, but I have no way of telling whether this
2967        is the right thing to do, as I do not use Windows. No doubt I'll get told
2968        if it's wrong...
2971    Version 4.3 21-May-03
2972    ---------------------
2974    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
2975       Makefile.
2977    2. Some refactoring to improve the quality of the code:
2979       (i)   The utf8_table... variables are now declared "const".
2981       (ii)  The code for \cx, which used the "case flipping" table to upper case
2982             lower case letters, now just substracts 32. This is ASCII-specific,
2983             but the whole concept of \cx is ASCII-specific, so it seems
2984             reasonable.
2986       (iii) PCRE was using its character types table to recognize decimal and
2987             hexadecimal digits in the pattern. This is silly, because it handles
2988             only 0-9, a-f, and A-F, but the character types table is locale-
2989             specific, which means strange things might happen. A private
2990             table is now used for this - though it costs 256 bytes, a table is
2991             much faster than multiple explicit tests. Of course, the standard
2992             character types table is still used for matching digits in subject
2993             strings against \d.
2995       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
2996             ending in _t are). So I've renamed it as ESC_tee.
2998    3. The first argument for regexec() in the POSIX wrapper should have been
2999       defined as "const".
3001    4. Changed pcretest to use malloc() for its buffers so that they can be
3002       Electric Fenced for debugging.
3004    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3005       to read one or more bytes before the start of the subject string. Often this
3006       had no effect on PCRE's behaviour, but in some circumstances it could
3007       provoke a segmentation fault.
3009    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3010       to try to read one or more bytes before the start of the subject string.
3012    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3013       UTF-8 support could misbehave in various ways if the subject string
3014       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3015       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3016       back over UTF-8 characters.)
3019    Version 4.2 14-Apr-03
3020    ---------------------
3022    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3024    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3025         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3026         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3027         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3028         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3029           and BUILD_EXEEXT
3030         Note: automatic setting of the BUILD variables is not yet working
3031         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3032           compile-time but not at link-time
3033         [LINK]: use for linking executables only
3034         make different versions for Windows and non-Windows
3035         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3036           libraries
3037         [LINK_FOR_BUILD]: new variable
3038         [OBJEXT]: use throughout
3039         [EXEEXT]: use throughout
3040         <winshared>: new target
3041         <wininstall>: new target
3042         <dftables.o>: use native compiler
3043         <dftables>: use native linker
3044         <install>: handle Windows platform correctly
3045         <clean>: ditto
3046         <check>: ditto
3047         copy DLL to top builddir before testing
3049       As part of these changes, -no-undefined was removed again. This was reported
3050       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3051       in any case.
3053    3. Some tidies to get rid of compiler warnings:
3055       . In the match_data structure, match_limit was an unsigned long int, whereas
3056         match_call_count was an int. I've made them both unsigned long ints.
3058       . In pcretest the fact that a const uschar * doesn't automatically cast to
3059         a void * provoked a warning.
3061       . Turning on some more compiler warnings threw up some "shadow" variables
3062         and a few more missing casts.
3064    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3065       option, a class that contained a single character with a value between 128
3066       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3068    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3069       option, a class that contained several characters, but with at least one
3070       whose value was between 128 and 255 caused PCRE to crash.
3073    Version 4.1 12-Mar-03
3074    ---------------------
3076    1. Compiling with gcc -pedantic found a couple of places where casts were
3077    needed, and a string in dftables.c that was longer than standard compilers are
3078    required to support.
3080    2. Compiling with Sun's compiler found a few more places where the code could
3081    be tidied up in order to avoid warnings.
3083    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3084    first of these names is deprecated in the latest Autoconf in favour of the name
3085    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3086    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3087    analogy I have changed it to CFLAGS_FOR_BUILD.
3089    4. Added -no-undefined to the linking command in the Makefile, because this is
3090    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3091    linking step for the pcreposix library.
3093    5. PCRE was failing to diagnose the case of two named groups with the same
3094    name.
3096    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3097    literal character that is needed in the subject for a match, and scans along to
3098    ensure that it is present before embarking on the full matching process. This
3099    saves time in cases of nested unlimited repeats that are never going to match.
3100    Problem: the scan can take a lot of time if the subject is very long (e.g.
3101    megabytes), thus penalizing straightforward matches. It is now done only if the
3102    amount of subject to be scanned is less than 1000 bytes.
3104    7. A lesser problem with the same optimization is that it was recording the
3105    first character of an anchored pattern as "needed", thus provoking a search
3106    right along the subject, even when the first match of the pattern was going to
3107    fail. The "needed" character is now not set for anchored patterns, unless it
3108    follows something in the pattern that is of non-fixed length. Thus, it still
3109    fulfils its original purpose of finding quick non-matches in cases of nested
3110    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3113    Version 4.0 17-Feb-03
3114    ---------------------
3116    1. If a comment in an extended regex that started immediately after a meta-item
3117    extended to the end of string, PCRE compiled incorrect data. This could lead to
3118    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3120    2. Moved to autoconf 2.53 and libtool 1.4.2.
3122    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3123    the special perltest8 script is no longer needed - all the tests can be run
3124    from a single perltest script.
3126    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3127    by \s. It has now been removed in PCRE. This means it isn't recognized as
3128    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3129    class [:space:] *does* include VT, thereby creating a mess.
3131    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3132    space and tab.
3134    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3135    its new features into the main test script, reducing the number of scripts.
3137    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3138    were backward compatible, and made the (?i) apply to the whole pattern, as if
3139    /i were given. Now it behaves more logically, and applies the option setting
3140    only to what follows. PCRE has been changed to follow suit. However, if it
3141    finds options settings right at the start of the pattern, it extracts them into
3142    the global options, as before. Thus, they show up in the info data.
3144    8. Added support for the \Q...\E escape sequence. Characters in between are
3145    treated as literals. This is slightly different from Perl in that $ and @ are
3146    also handled as literals inside the quotes. In Perl, they will cause variable
3147    interpolation. Note the following examples:
3149        Pattern            PCRE matches      Perl matches
3151        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3152        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3153        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3155    For compatibility with Perl, \Q...\E sequences are recognized inside character
3156    classes as well as outside them.
3158    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3159    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3160    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3161    signed/unsigned warnings.
3163    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3164    option for pcretest, so I've replaced it by a simple function that does just
3165    that job.
3167    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3168    "pcregrep -".
3170    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3171    Java package. This provides some syntactic sugar for simple cases of what my
3172    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3173    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3174    item, you can use this simplified notation. Note that only makes sense with
3175    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3176    greediness, whatever the setting of the PCRE_UNGREEDY option.
3178    13. A change of greediness default within a pattern was not taking effect at
3179    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3180    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3181    was abstracted outside.
3183    14. PCRE now supports the \G assertion. It is true when the current matching
3184    position is at the start point of the match. This differs from \A when the
3185    starting offset is non-zero. Used with the /g option of pcretest (or similar
3186    code), it works in the same way as it does for Perl's /g option. If all
3187    alternatives of a regex begin with \G, the expression is anchored to the start
3188    match position, and the "anchored" flag is set in the compiled expression.
3190    15. Some bugs concerning the handling of certain option changes within patterns
3191    have been fixed. These applied to options other than (?ims). For example,
3192    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3193    been the other way round. Some of this was related to change 7 above.
3195    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3196    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3197    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3198    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3200    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3201    mode. Unlike ".", it always matches newline, whatever the setting of
3202    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3203    assertions. Perl allows it, but it doesn't (in general) work because it can't
3204    calculate the length of the lookbehind. At least, that's the case for Perl
3205    5.8.0 - I've been told they are going to document that it doesn't work in
3206    future.
3208    18. Added an error diagnosis for escapes that PCRE does not support: these are
3209    \L, \l, \N, \P, \p, \U, \u, and \X.
3211    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3212    reading past the end of the pattern in cases such as /[abcd/.
3214    20. PCRE was getting more memory than necessary for patterns with classes that
3215    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3217    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3218    compiling PCRE for use with Virtual Pascal.
3220    22. Small fix to the Makefile to make it work properly if the build is done
3221    outside the source tree.
3223    23. Added a new extension: a condition to go with recursion. If a conditional
3224    subpattern starts with (?(R) the "true" branch is used if recursion has
3225    happened, whereas the "false" branch is used only at the top level.
3227    24. When there was a very long string of literal characters (over 255 bytes
3228    without UTF support, over 250 bytes with UTF support), the computation of how
3229    much memory was required could be incorrect, leading to segfaults or other
3230    strange effects.
3232    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3233    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3234    there was a subsequent back reference to those brackets. This meant that, for
3235    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3236    possible to check for precisely this case. All we can do is abandon the
3237    optimization if .* occurs inside capturing brackets when there are any back
3238    references whatsoever. (See below for a better fix that came later.)
3240    26. The handling of the optimization for finding the first character of a
3241    non-anchored pattern, and for finding a character that is required later in the
3242    match were failing in some cases. This didn't break the matching; it just
3243    failed to optimize when it could. The way this is done has been re-implemented.
3245    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3247    28. Added a new feature that provides some of the functionality that Perl
3248    provides with (?{...}). The facility is termed a "callout". The way it is done
3249    in PCRE is for the caller to provide an optional function, by setting
3250    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3251    global variable. By default it is unset, which disables all calling out. To get
3252    the function called, the regex must include (?C) at appropriate points. This
3253    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3254    This provides a means of identifying different callout points. When PCRE
3255    reaches such a point in the regex, if pcre_callout has been set, the external
3256    function is called. It is provided with data in a structure called
3257    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3258    matching continues; if it returns a non-zero value, the match at the current
3259    point fails. However, backtracking will occur if possible. [This was changed
3260    later and other features added - see item 49 below.]
3262    29. pcretest is upgraded to test the callout functionality. It provides a
3263    callout function that displays information. By default, it shows the start of
3264    the match and the current position in the text. There are some new data escapes
3265    to vary what happens:
3267        \C+         in addition, show current contents of captured substrings
3268        \C-         do not supply a callout function
3269        \C!n        return 1 when callout number n is reached
3270        \C!n!m      return 1 when callout number n is reached for the mth time
3272    30. If pcregrep was called with the -l option and just a single file name, it
3273    output "<stdin>" if a match was found, instead of the file name.
3275    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3276    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3277    pcre_exec(). This saves a malloc/free per call. The default value of
3278    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3279    when configuring.
3281    32. The default maximum size of a compiled pattern is 64K. There have been a
3282    few cases of people hitting this limit. The code now uses macros to handle the
3283    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3284    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3285    configuring. Tests 2 and 5 work only with 2-byte links because they output
3286    debugging information about compiled patterns.
3288    33. Internal code re-arrangements:
3290    (a) Moved the debugging function for printing out a compiled regex into
3291        its own source file (printint.c) and used #include to pull it into
3292        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3293        separate copies.
3295    (b) Defined the list of op-code names for debugging as a macro in
3296        internal.h so that it is next to the definition of the opcodes.
3298    (c) Defined a table of op-code lengths for simpler skipping along compiled
3299        code. This is again a macro in internal.h so that it is next to the
3300        definition of the opcodes.
3302    34. Added support for recursive calls to individual subpatterns, along the
3303    lines of Robin Houston's patch (but implemented somewhat differently).
3305    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3306    allow it to read and process whole directories in Win32. This code was
3307    contributed by Lionel Fourquaux; it has not been tested by me.
3309    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3310    used to name a group. Names consist of alphanumerics and underscores, and must
3311    be unique. Back references use the syntax (?P=name) and recursive calls use
3312    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3313    numbers. The function pcre_fullinfo() can be used after compilation to extract
3314    a name/number map. There are three relevant calls:
3316      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3317      PCRE_INFO_NAMECOUNT            yields the number of entries
3318      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3320    The map is a vector of fixed-size entries. The size of each entry depends on
3321    the length of the longest name used. The first two bytes of each entry are the
3322    group number, most significant byte first. There follows the corresponding
3323    name, zero terminated. The names are in alphabetical order.
3325    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3326    case instead of 255. Making it the same both with and without UTF-8 support
3327    means that the same test output works with both.
3329    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3330    calling malloc() with a zero argument.
3332    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3333    optimization. I've improved things by keeping a bitmap of backreferences with
3334    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3335    fact referenced, the optimization can be applied. It is unlikely that a
3336    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3337    the match to follow \n) will appear inside brackets with a number greater than
3338    31, but if it does, any back reference > 31 suppresses the optimization.
3340    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3341    of disabling numbered capturing parentheses. Any opening parenthesis that is
3342    not followed by ? behaves as if it were followed by ?: but named parentheses
3343    can still be used for capturing (and they will acquire numbers in the usual
3344    way).
3346    41. Redesigned the return codes from the match() function into yes/no/error so
3347    that errors can be passed back from deep inside the nested calls. A malloc
3348    failure while inside a recursive subpattern call now causes the
3349    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3351    42. It is now possible to set a limit on the number of times the match()
3352    function is called in a call to pcre_exec(). This facility makes it possible to
3353    limit the amount of recursion and backtracking, though not in a directly
3354    obvious way, because the match() function is used in a number of different
3355    circumstances. The count starts from zero for each position in the subject
3356    string (for non-anchored patterns). The default limit is, for compatibility, a
3357    large number, namely 10 000 000. You can change this in two ways:
3359    (a) When configuring PCRE before making, you can use --with-match-limit=n
3360        to set a default value for the compiled library.
3362    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3363        a different value is set. See 45 below.
3365    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3367    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3368    of things that can be changed at compile time. The first argument specifies
3369    what is wanted and the second points to where the information is to be placed.
3370    The current list of available information is:
3372      PCRE_CONFIG_UTF8
3374    The output is an integer that is set to one if UTF-8 support is available;
3375    otherwise it is set to zero.
3379    The output is an integer that it set to the value of the code that is used for
3380    newline. It is either LF (10) or CR (13).
3384    The output is an integer that contains the number of bytes used for internal
3385    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3389    The output is an integer that contains the threshold above which the POSIX
3390    interface uses malloc() for output vectors. See item 31 above.
3394    The output is an unsigned integer that contains the default limit of the number
3395    of match() calls in a pcre_exec() execution. See 42 above.
3397    44. pcretest has been upgraded by the addition of the -C option. This causes it
3398    to extract all the available output from the new pcre_config() function, and to
3399    output it. The program then exits immediately.
3401    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3402    order to support additional features. One way would have been to define
3403    pcre_exec2() (for example) with extra arguments, but this would not have been
3404    extensible, and would also have required all calls to the original function to
3405    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3406    is used for passing in "extra" data from pcre_study().
3408    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3409    contains the following fields:
3411      flags         a bitmap indicating which of the following fields are set