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

Diff of /code/trunk/ChangeLog

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

revision 5 by nigel, Sat Feb 24 21:38:05 2007 UTC revision 953 by zherczeg, Thu Mar 29 17:41:57 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    9.  (*MARK) control verb is now supported by the JIT compiler.
46    
47    10. The command "./RunTest list" lists the available tests without actually
48        running any of them. (Because I keep forgetting what they all are.)
49    
50    11. Add PCRE_INFO_MAXLOOKBEHIND.
51    
52    12. Applied a (slightly modified) user-supplied patch that improves performance
53        when the heap is used for recursion (compiled with --disable-stack-for-
54        recursion). Instead of malloc and free for each heap frame each time a
55        logical recursion happens, frames are retained on a chain and re-used where
56        possible. This sometimes gives as much as 30% improvement.
57    
58    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59        call.
60    
61    14. As documented, (*COMMIT) is now confined to within a positive assertion.
62    
63    15. It is now possible to link pcretest with libedit as an alternative to
64        libreadline.
65    
66    16. (*COMMIT) control verb is now supported by the JIT compiler.
67    
68    17. The Unicode data tables have been updated to Unicode 6.1.0.
69    
70    18. Added --file-list option to pcregrep.
71    
72    19. Added binary file support to pcregrep, including the -a, --binary-files,
73        -I, and --text options.
74    
75    20. The madvise function is renamed for posix_madvise for QNX compatibility
76        reasons. Fixed by Giuseppe D'Angelo.
77    
78    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79    
80    
81    Version 8.30 04-February-2012
82    -----------------------------
83    
84    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
85        name is defined in ctype.h.
86    
87    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
88        only in quite long subpatterns.
89    
90    3.  Removed the function pcre_info(), which has been obsolete and deprecated
91        since it was replaced by pcre_fullinfo() in February 2000.
92    
93    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
94        match a (*MARK), and the match failed at the start of the subject, a
95        reference to memory before the start of the subject could occur. This bug
96        was introduced by fix 17 of release 8.21.
97    
98    5.  A reference to an unset group with zero minimum repetition was giving
99        totally wrong answers (in non-JavaScript-compatibility mode). For example,
100        /(another)?(\1?)test/ matched against "hello world test". This bug was
101        introduced in release 8.13.
102    
103    6.  Add support for 16-bit character strings (a large amount of work involving
104        many changes and refactorings).
105    
106    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
107        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
108        from a file.
109    
110    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
111        rounding is not applied in this particular case).
112    
113    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
114        if they appear, or are escaped, in patterns.
115    
116    10. Get rid of a number of -Wunused-but-set-variable warnings.
117    
118    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
119        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
120        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
121        also returns the mark "x". This bug applied to capturing parentheses,
122        non-capturing parentheses, and atomic parentheses. It also applied to some
123        assertions.
124    
125    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
126        information out of configure.ac instead of relying on pcre.h.generic, which
127        is not stored in the repository.
128    
129    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
130        -lreadline.
131    
132    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
133    
134    15. Applied Graycode's patch to put the top-level frame on the stack rather
135        than the heap when not using the stack for recursion. This gives a
136        performance improvement in many cases when recursion is not deep.
137    
138    16. Experimental code added to "pcretest -C" to output the stack frame size.
139    
140    
141    Version 8.21 12-Dec-2011
142    ------------------------
143    
144    1.  Updating the JIT compiler.
145    
146    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
147        are added as well.
148    
149    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
150        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
151        calling _pcre_jit_exec. Some extra comments are added.
152    
153    4.  (*MARK) settings inside atomic groups that do not contain any capturing
154        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
155        was introduced by change 18 for 8.20.
156    
157    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
158        ECMA-262 standard.
159    
160    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
161        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
162        This bug was probably introduced by change 9 of 8.13.
163    
164    7.  While fixing 6 above, I noticed that a number of other items were being
165        incorrectly rejected as "not fixed length". This arose partly because newer
166        opcodes had not been added to the fixed-length checking code. I have (a)
167        corrected the bug and added tests for these items, and (b) arranged for an
168        error to occur if an unknown opcode is encountered while checking for fixed
169        length instead of just assuming "not fixed length". The items that were
170        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
171        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
172        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
173    
174    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
175        being incorrectly compiled and would have given unpredicatble results.
176    
177    9.  A possessively repeated subpattern with minimum repeat count greater than
178        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
179        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
180        the first (A) could occur when it should not.
181    
182    10. Add a cast and remove a redundant test from the code.
183    
184    11. JIT should use pcre_malloc/pcre_free for allocation.
185    
186    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
187        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
188        is anything other than /usr, -L is still shown).
189    
190    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
191    
192    14. Perl does not support \N without a following name in a [] class; PCRE now
193        also gives an error.
194    
195    15. If a forward reference was repeated with an upper limit of around 2000,
196        it caused the error "internal error: overran compiling workspace". The
197        maximum number of forward references (including repeats) was limited by the
198        internal workspace, and dependent on the LINK_SIZE. The code has been
199        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
200        the default depends on LINK_SIZE. There is a new upper limit (for safety)
201        of around 200,000 forward references. While doing this, I also speeded up
202        the filling in of repeated forward references.
203    
204    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
205        incorrectly expecting the subject to contain another "a" after the start.
206    
207    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
208        in the match, the SKIP should be ignored. This was not happening; instead
209        the SKIP was being treated as NOMATCH. For patterns such as
210        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
211        tested.
212    
213    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
214        now much more compatible with Perl, in particular in cases where the result
215        is a non-match for a non-anchored pattern. For example, if
216        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
217        "m", where previously it did not return a name. A side effect of this
218        change is that for partial matches, the last encountered mark name is
219        returned, as for non matches. A number of tests that were previously not
220        Perl-compatible have been moved into the Perl-compatible test files. The
221        refactoring has had the pleasing side effect of removing one argument from
222        the match() function, thus reducing its stack requirements.
223    
224    19. If the /S+ option was used in pcretest to study a pattern using JIT,
225        subsequent uses of /S (without +) incorrectly behaved like /S+.
226    
227    21. Retrieve executable code size support for the JIT compiler and fixing
228        some warnings.
229    
230    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
231        not work when the shorter character appeared right at the end of the
232        subject string.
233    
234    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
235        systems.
236    
237    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
238        output it when the /M option is used in pcretest.
239    
240    25. The CheckMan script was not being included in the distribution. Also, added
241        an explicit "perl" to run Perl scripts from the PrepareRelease script
242        because this is reportedly needed in Windows.
243    
244    26. If study data was being save in a file and studying had not found a set of
245        "starts with" bytes for the pattern, the data written to the file (though
246        never used) was taken from uninitialized memory and so caused valgrind to
247        complain.
248    
249    27. Updated RunTest.bat as provided by Sheri Pierce.
250    
251    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
252    
253    29. Computation of memory usage for the table of capturing group names was
254        giving an unnecessarily large value.
255    
256    
257    Version 8.20 21-Oct-2011
258    ------------------------
259    
260    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
261        a POSIX class. After further experiments with Perl, which convinced me that
262        Perl has bugs and confusions, a closing square bracket is no longer allowed
263        in a POSIX name. This bug also affected patterns with classes that started
264        with full stops.
265    
266    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
267        captured substring, but while checking the failing first alternative,
268        substring 1 is temporarily captured. If the output vector supplied to
269        pcre_exec() was not big enough for this capture, the yield of the function
270        was still zero ("insufficient space for captured substrings"). This cannot
271        be totally fixed without adding another stack variable, which seems a lot
272        of expense for a edge case. However, I have improved the situation in cases
273        such as /(a)(b)x|abc/ matched against "abc", where the return code
274        indicates that fewer than the maximum number of slots in the ovector have
275        been set.
276    
277    3.  Related to (2) above: when there are more back references in a pattern than
278        slots in the output vector, pcre_exec() uses temporary memory during
279        matching, and copies in the captures as far as possible afterwards. It was
280        using the entire output vector, but this conflicts with the specification
281        that only 2/3 is used for passing back captured substrings. Now it uses
282        only the first 2/3, for compatibility. This is, of course, another edge
283        case.
284    
285    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
286        main code base, and can be used by building with --enable-jit. When this is
287        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
288        runtime --no-jit option is given.
289    
290    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
291        ovector, the return from the function was zero, implying that there were
292        other matches that did not fit. The correct "exactly full" value is now
293        returned.
294    
295    6.  If a subpattern that was called recursively or as a subroutine contained
296        (*PRUNE) or any other control that caused it to give a non-standard return,
297        invalid errors such as "Error -26 (nested recursion at the same subject
298        position)" or even infinite loops could occur.
299    
300    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
301        computing the minimum length on reaching *ACCEPT, and so ended up with the
302        wrong value of 1 rather than 0. Further investigation indicates that
303        computing a minimum subject length in the presence of *ACCEPT is difficult
304        (think back references, subroutine calls), and so I have changed the code
305        so that no minimum is registered for a pattern that contains *ACCEPT.
306    
307    8.  If (*THEN) was present in the first (true) branch of a conditional group,
308        it was not handled as intended. [But see 16 below.]
309    
310    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
311        Sheri Pierce.
312    
313    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
314        the first byte in a match must be "a".
315    
316    11. Change 17 for 8.13 increased the recursion depth for patterns like
317        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
318        pattern contains any instances of (*THEN). If it does not, the old
319        optimizations are restored. It would be nice to do this on a per-group
320        basis, but at the moment that is not feasible.
321    
322    12. In some environments, the output of pcretest -C is CRLF terminated. This
323        broke RunTest's code that checks for the link size. A single white space
324        character after the value is now allowed for.
325    
326    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
327        For "fr", it uses the Windows-specific input and output files.
328    
329    14. If (*THEN) appeared in a group that was called recursively or as a
330        subroutine, it did not work as intended. [But see next item.]
331    
332    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
333        pattern fragments (but not containing any | characters). If A and B are
334        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
335        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
336        D. In other words, Perl considers parentheses that do not contain any |
337        characters to be part of a surrounding alternative, whereas PCRE was
338        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
339        differently. PCRE now behaves in the same way as Perl, except in the case
340        of subroutine/recursion calls such as (?1) which have in any case always
341        been different (but PCRE had them first :-).
342    
343    16. Related to 15 above: Perl does not treat the | in a conditional group as
344        creating alternatives. Such a group is treated in the same way as an
345        ordinary group without any | characters when processing (*THEN). PCRE has
346        been changed to match Perl's behaviour.
347    
348    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
349        RunGrepTest script failed.
350    
351    18. Change 22 for version 13 caused atomic groups to use more stack. This is
352        inevitable for groups that contain captures, but it can lead to a lot of
353        stack use in large patterns. The old behaviour has been restored for atomic
354        groups that do not contain any capturing parentheses.
355    
356    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
357        suppress the check for a minimum subject length at run time. (If it was
358        given to pcre_exec() or pcre_dfa_exec() it did work.)
359    
360    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
361        fail to work when decoding hex characters in data strings in EBCDIC
362        environments.
363    
364    21. It appears that in at least one Mac OS environment, the isxdigit() function
365        is implemented as a macro that evaluates to its argument more than once,
366        contravening the C 90 Standard (I haven't checked a later standard). There
367        was an instance in pcretest which caused it to go wrong when processing
368        \x{...} escapes in subject strings. The has been rewritten to avoid using
369        things like p++ in the argument of isxdigit().
370    
371    
372    Version 8.13 16-Aug-2011
373    ------------------------
374    
375    1.  The Unicode data tables have been updated to Unicode 6.0.0.
376    
377    2.  Two minor typos in pcre_internal.h have been fixed.
378    
379    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
380        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
381        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
382    
383    4.  There were a number of related bugs in the code for matching backrefences
384        caselessly in UTF-8 mode when codes for the characters concerned were
385        different numbers of bytes. For example, U+023A and U+2C65 are an upper
386        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
387        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
388        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
389        2-byte code at the end of the subject (it thought there wasn't enough data
390        left).
391    
392    5.  Comprehensive information about what went wrong is now returned by
393        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
394        as the output vector has at least 2 elements. The offset of the start of
395        the failing character and a reason code are placed in the vector.
396    
397    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
398        now returned is for the first byte of the failing character, instead of the
399        last byte inspected. This is an incompatible change, but I hope it is small
400        enough not to be a problem. It makes the returned offset consistent with
401        pcre_exec() and pcre_dfa_exec().
402    
403    7.  pcretest now gives a text phrase as well as the error number when
404        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
405        failure, the offset and reason code are output.
406    
407    8.  When \R was used with a maximizing quantifier it failed to skip backwards
408        over a \r\n pair if the subsequent match failed. Instead, it just skipped
409        back over a single character (\n). This seems wrong (because it treated the
410        two characters as a single entity when going forwards), conflicts with the
411        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
412        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
413        has been changed.
414    
415    9.  Some internal refactoring has changed the processing so that the handling
416        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
417        time (the PCRE_DOTALL option was changed this way some time ago: version
418        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
419        which was always a bit of a fudge. It also means that there is one less
420        argument for the match() function, which reduces its stack requirements
421        slightly. This change also fixes an incompatibility with Perl: the pattern
422        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
423    
424    10. More internal refactoring has drastically reduced the number of recursive
425        calls to match() for possessively repeated groups such as (abc)++ when
426        using pcre_exec().
427    
428    11. While implementing 10, a number of bugs in the handling of groups were
429        discovered and fixed:
430    
431        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
432        (a|)*(?1) gave a compile-time internal error.
433        ((a|)+)+  did not notice that the outer group could match an empty string.
434        (^a|^)+   was not marked as anchored.
435        (.*a|.*)+ was not marked as matching at start or after a newline.
436    
437    12. Yet more internal refactoring has removed another argument from the match()
438        function. Special calls to this function are now indicated by setting a
439        value in a variable in the "match data" data block.
440    
441    13. Be more explicit in pcre_study() instead of relying on "default" for
442        opcodes that mean there is no starting character; this means that when new
443        ones are added and accidentally left out of pcre_study(), testing should
444        pick them up.
445    
446    14. The -s option of pcretest has been documented for ages as being an old
447        synonym of -m (show memory usage). I have changed it to mean "force study
448        for every regex", that is, assume /S for every regex. This is similar to -i
449        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
450        using it. It makes it easier to run collections of tests with and without
451        study enabled, and thereby test pcre_study() more easily. All the standard
452        tests are now run with and without -s (but some patterns can be marked as
453        "never study" - see 20 below).
454    
455    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
456        restoration of the capturing data to the outer values was not happening
457        correctly.
458    
459    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
460        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
461        pattern had matched an empty string, and so incorrectly returned a no
462        match.
463    
464    17. There was optimizing code for the last branch of non-capturing parentheses,
465        and also for the obeyed branch of a conditional subexpression, which used
466        tail recursion to cut down on stack usage. Unfortunately, now that there is
467        the possibility of (*THEN) occurring in these branches, tail recursion is
468        no longer possible because the return has to be checked for (*THEN). These
469        two optimizations have therefore been removed. [But see 8.20/11 above.]
470    
471    18. If a pattern containing \R was studied, it was assumed that \R always
472        matched two bytes, thus causing the minimum subject length to be
473        incorrectly computed because \R can also match just one byte.
474    
475    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
476        was incorrectly computed.
477    
478    20. If /S is present twice on a test pattern in pcretest input, it now
479        *disables* studying, thereby overriding the use of -s on the command line
480        (see 14 above). This is necessary for one or two tests to keep the output
481        identical in both cases.
482    
483    21. When (*ACCEPT) was used in an assertion that matched an empty string and
484        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
485    
486    22. When an atomic group that contained a capturing parenthesis was
487        successfully matched, but the branch in which it appeared failed, the
488        capturing was not being forgotten if a higher numbered group was later
489        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
490        group 1 to "a", when in fact it should be unset. This applied to multi-
491        branched capturing and non-capturing groups, repeated or not, and also to
492        positive assertions (capturing in negative assertions does not happen
493        in PCRE) and also to nested atomic groups.
494    
495    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
496        subject after a captured substring, to make it easier to tell which of a
497        number of identical substrings has been captured.
498    
499    24. The way atomic groups are processed by pcre_exec() has been changed so that
500        if they are repeated, backtracking one repetition now resets captured
501        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
502        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
503        "aaa". Previously, it would have been "a". As part of this code
504        refactoring, the way recursive calls are handled has also been changed.
505    
506    25. If an assertion condition captured any substrings, they were not passed
507        back unless some other capturing happened later. For example, if
508        (?(?=(a))a) was matched against "a", no capturing was returned.
509    
510    26. When studying a pattern that contained subroutine calls or assertions,
511        the code for finding the minimum length of a possible match was handling
512        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
513        group 1 called group 2 while simultaneously a separate group 2 called group
514        1). A stack overflow occurred in this case. I have fixed this by limiting
515        the recursion depth to 10.
516    
517    27. Updated RunTest.bat in the distribution to the version supplied by Tom
518        Fortmann. This supports explicit test numbers on the command line, and has
519        argument validation and error reporting.
520    
521    28. An instance of \X with an unlimited repeat could fail if at any point the
522        first character it looked at was a mark character.
523    
524    29. Some minor code refactoring concerning Unicode properties and scripts
525        should reduce the stack requirement of match() slightly.
526    
527    30. Added the '=' option to pcretest to check the setting of unused capturing
528        slots at the end of the pattern, which are documented as being -1, but are
529        not included in the return count.
530    
531    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
532        compiled something random. Now it gives a compile-time error (as does
533        Perl).
534    
535    32. A *MARK encountered during the processing of a positive assertion is now
536        recorded and passed back (compatible with Perl).
537    
538    33. If --only-matching or --colour was set on a pcregrep call whose pattern
539        had alternative anchored branches, the search for a second match in a line
540        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
541        matched the line "0102" twice. The same bug affected patterns that started
542        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
543        twice.
544    
545    34. Previously, PCRE did not allow quantification of assertions. However, Perl
546        does, and because of capturing effects, quantifying parenthesized
547        assertions may at times be useful. Quantifiers are now allowed for
548        parenthesized assertions.
549    
550    35. A minor code tidy in pcre_compile() when checking options for \R usage.
551    
552    36. \g was being checked for fancy things in a character class, when it should
553        just be a literal "g".
554    
555    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
556        appearance of a nested POSIX class supersedes an apparent external class.
557        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
558        unescaped square brackets may also appear as part of class names. For
559        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
560        more like Perl. (But see 8.20/1 above.)
561    
562    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
563        was because it thought it was \N{name}, which is not supported).
564    
565    39. Add minix to OS list not supporting the -S option in pcretest.
566    
567    40. PCRE tries to detect cases of infinite recursion at compile time, but it
568        cannot analyze patterns in sufficient detail to catch mutual recursions
569        such as ((?1))((?2)). There is now a runtime test that gives an error if a
570        subgroup is called recursively as a subpattern for a second time at the
571        same position in the subject string. In previous releases this might have
572        been caught by the recursion limit, or it might have run out of stack.
573    
574    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
575        happen only once. PCRE was, however incorrectly giving a compile time error
576        "recursive call could loop indefinitely" because it cannot analyze the
577        pattern in sufficient detail. The compile time test no longer happens when
578        PCRE is compiling a conditional subpattern, but actual runaway loops are
579        now caught at runtime (see 40 above).
580    
581    42. It seems that Perl allows any characters other than a closing parenthesis
582        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
583        has been changed to be the same.
584    
585    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
586        as not to get warnings when autogen.sh is called. Also changed
587        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
588    
589    44. To help people who use pcregrep to scan files containing exceedingly long
590        lines, the following changes have been made:
591    
592        (a) The default value of the buffer size parameter has been increased from
593            8K to 20K. (The actual buffer used is three times this size.)
594    
595        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
596            PCRE is built.
597    
598        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
599            to be set at run time.
600    
601        (d) Numerical values in pcregrep options can be followed by K or M, for
602            example --buffer-size=50K.
603    
604        (e) If a line being scanned overflows pcregrep's buffer, an error is now
605            given and the return code is set to 2.
606    
607    45. Add a pointer to the latest mark to the callout data block.
608    
609    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
610        partial match of an empty string instead of no match. This was specific to
611        the use of ".".
612    
613    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
614        complete match instead of a partial match. This bug was dependent on both
615        the PCRE_UTF8 and PCRE_DOTALL options being set.
616    
617    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
618        starting byte set, because \b was not being ignored.
619    
620    
621    Version 8.12 15-Jan-2011
622    ------------------------
623    
624    1.  Fixed some typos in the markup of the man pages, and wrote a script that
625        checks for such things as part of the documentation building process.
626    
627    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
628        --match-limit and --recursion-limit options (added for 8.11). In
629        particular, this made one of the standard tests fail. (The integer value
630        went into the wrong half of a long int.)
631    
632    3.  If the --colour option was given to pcregrep with -v (invert match), it
633        did strange things, either producing crazy output, or crashing. It should,
634        of course, ignore a request for colour when reporting lines that do not
635        match.
636    
637    4.  Another pcregrep bug caused similar problems if --colour was specified with
638        -M (multiline) and the pattern match finished with a line ending.
639    
640    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
641        matched in multiline mode, the following line was shown as part of the
642        match. This seems wrong, so I have changed it.
643    
644    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
645        the check for further matches in the same line (so they could be coloured)
646        to overrun the end of the current line. If another match was found, it was
647        incorrectly shown (and then shown again when found in the next line).
648    
649    7.  If pcregrep was compiled under Windows, there was a reference to the
650        function pcregrep_exit() before it was defined. I am assuming this was
651        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
652        reported by a user. I've moved the definition above the reference.
653    
654    
655    Version 8.11 10-Dec-2010
656    ------------------------
657    
658    1.  (*THEN) was not working properly if there were untried alternatives prior
659        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
660        backtracked to try for "b" instead of moving to the next alternative branch
661        at the same level (in this case, to look for "c"). The Perl documentation
662        is clear that when (*THEN) is backtracked onto, it goes to the "next
663        alternative in the innermost enclosing group".
664    
665    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
666        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
667        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
668        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
669        (*THEN).
670    
671    3.  If \s appeared in a character class, it removed the VT character from
672        the class, even if it had been included by some previous item, for example
673        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
674        of \s, but is part of the POSIX "space" class.)
675    
676    4.  A partial match never returns an empty string (because you can always
677        match an empty string at the end of the subject); however the checking for
678        an empty string was starting at the "start of match" point. This has been
679        changed to the "earliest inspected character" point, because the returned
680        data for a partial match starts at this character. This means that, for
681        example, /(?<=abc)def/ gives a partial match for the subject "abc"
682        (previously it gave "no match").
683    
684    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
685        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
686        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
687        has an implication that the given string is incomplete (because a partial
688        match is preferred over a full match). For this reason, these items now
689        give a partial match in this situation. [Aside: previously, the one case
690        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
691        match rather than a full match, which was wrong by the old rules, but is
692        now correct.]
693    
694    6.  There was a bug in the handling of #-introduced comments, recognized when
695        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
696        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
697        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
698        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
699        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
700        places in pcre_compile().
701    
702    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
703        comments when looking ahead for named forward references to subpatterns,
704        the only newline sequence it recognized was NL. It now handles newlines
705        according to the set newline convention.
706    
707    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
708        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
709        cater for a lack of strerror(). These oversights have been fixed.
710    
711    9.  Added --match-limit and --recursion-limit to pcregrep.
712    
713    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
714    
715    11. When the -o option was used, pcregrep was setting a return code of 1, even
716        when matches were found, and --line-buffered was not being honoured.
717    
718    12. Added an optional parentheses number to the -o and --only-matching options
719        of pcregrep.
720    
721    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
722        can match an empty string. The code to do it in pcretest and pcredemo
723        needed fixing:
724    
725        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
726            only one byte after an empty string match just before CRLF (this case
727            just got forgotten; "any" and "anycrlf" were OK).
728    
729        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
730            mode when an empty string match preceded an ASCII character followed by
731            a non-ASCII character. (The code for advancing by one character rather
732            than one byte was nonsense.)
733    
734        (c) The pcredemo.c sample program did not have any code at all to handle
735            the cases when CRLF is a valid newline sequence.
736    
737    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
738        as a starting offset was within the subject string. There is now a new
739        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
740        negative or greater than the length of the string. In order to test this,
741        pcretest is extended to allow the setting of negative starting offsets.
742    
743    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
744        starting offset points to the beginning of a UTF-8 character was
745        unnecessarily clumsy. I tidied it up.
746    
747    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
748        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
749    
750    17. Nobody had reported that the --include_dir option, which was added in
751        release 7.7 should have been called --include-dir (hyphen, not underscore)
752        for compatibility with GNU grep. I have changed it to --include-dir, but
753        left --include_dir as an undocumented synonym, and the same for
754        --exclude-dir, though that is not available in GNU grep, at least as of
755        release 2.5.4.
756    
757    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
758        characters from a string of bytes) have been redefined so as not to use
759        loops, in order to improve performance in some environments. At the same
760        time, I abstracted some of the common code into auxiliary macros to save
761        repetition (this should not affect the compiled code).
762    
763    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
764        compile-time error is now given if \c is not followed by an ASCII
765        character, that is, a byte less than 128. (In EBCDIC mode, the code is
766        different, and any byte value is allowed.)
767    
768    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
769        START_OPTIMIZE option, which is now allowed at compile time - but just
770        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
771        to pcregrep and other applications that have no direct access to PCRE
772        options. The new /Y option in pcretest sets this option when calling
773        pcre_compile().
774    
775    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
776        back references. Groups containing recursive back references were forced to
777        be atomic by that change, but in the case of named groups, the amount of
778        memory required was incorrectly computed, leading to "Failed: internal
779        error: code overflow". This has been fixed.
780    
781    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
782        pcretest.c, to avoid build problems in some Borland environments.
783    
784    
785    Version 8.10 25-Jun-2010
786    ------------------------
787    
788    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
789        THEN.
790    
791    2.  (*ACCEPT) was not working when inside an atomic group.
792    
793    3.  Inside a character class, \B is treated as a literal by default, but
794        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
795        causes the error). The code is unchanged, but I tidied the documentation.
796    
797    4.  Inside a character class, PCRE always treated \R and \X as literals,
798        whereas Perl faults them if its -w option is set. I have changed PCRE so
799        that it faults them when PCRE_EXTRA is set.
800    
801    5.  Added support for \N, which always matches any character other than
802        newline. (It is the same as "." when PCRE_DOTALL is not set.)
803    
804    6.  When compiling pcregrep with newer versions of gcc which may have
805        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
806        declared with attribute warn_unused_result" were given. Just casting the
807        result to (void) does not stop the warnings; a more elaborate fudge is
808        needed. I've used a macro to implement this.
809    
810    7.  Minor change to pcretest.c to avoid a compiler warning.
811    
812    8.  Added four artifical Unicode properties to help with an option to make
813        \s etc use properties (see next item). The new properties are: Xan
814        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
815    
816    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
817        use Unicode properties. (*UCP) at the start of a pattern can be used to set
818        this option. Modified pcretest to add /W to test this facility. Added
819        REG_UCP to make it available via the POSIX interface.
820    
821    10. Added --line-buffered to pcregrep.
822    
823    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
824        studied, and the match started with a letter with a code point greater than
825        127 whose first byte was different to the first byte of the other case of
826        the letter, the other case of this starting letter was not recognized
827        (#976).
828    
829    12. If a pattern that was studied started with a repeated Unicode property
830        test, for example, \p{Nd}+, there was the theoretical possibility of
831        setting up an incorrect bitmap of starting bytes, but fortunately it could
832        not have actually happened in practice until change 8 above was made (it
833        added property types that matched character-matching opcodes).
834    
835    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
836        possible starting bytes for non-anchored patterns.
837    
838    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
839        \R, and also a number of cases that involve Unicode properties, both
840        explicit and implicit when PCRE_UCP is set.
841    
842    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
843        input, it could crash or give wrong results if characters with values
844        greater than 0xc0 were present in the subject string. (Detail: it assumed
845        UTF-8 input when processing these items.)
846    
847    16. Added a lot of (int) casts to avoid compiler warnings in systems where
848        size_t is 64-bit (#991).
849    
850    17. Added a check for running out of memory when PCRE is compiled with
851        --disable-stack-for-recursion (#990).
852    
853    18. If the last data line in a file for pcretest does not have a newline on
854        the end, a newline was missing in the output.
855    
856    19. The default pcre_chartables.c file recognizes only ASCII characters (values
857        less than 128) in its various bitmaps. However, there is a facility for
858        generating tables according to the current locale when PCRE is compiled. It
859        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
860        characters, are recognized by isspace() and therefore were getting set in
861        these tables, and indeed these tables seem to approximate to ISO 8859. This
862        caused a problem in UTF-8 mode when pcre_study() was used to create a list
863        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
864        which of course cannot start UTF-8 characters. I have changed the code so
865        that only real ASCII characters (less than 128) and the correct starting
866        bytes for UTF-8 encodings are set for characters greater than 127 when in
867        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
868        altogether.)
869    
870    20. Added the /T option to pcretest so as to be able to run tests with non-
871        standard character tables, thus making it possible to include the tests
872        used for 19 above in the standard set of tests.
873    
874    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
875        reference to a subpattern the other side of a comment that contains an
876        opening parenthesis caused either an internal compiling error, or a
877        reference to the wrong subpattern.
878    
879    
880    Version 8.02 19-Mar-2010
881    ------------------------
882    
883    1.  The Unicode data tables have been updated to Unicode 5.2.0.
884    
885    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
886        configured.
887    
888    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
889        original author of that file, following a query about its status.
890    
891    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
892        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
893    
894    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
895        quantifier applied to a forward-referencing subroutine call, could compile
896        incorrect code or give the error "internal error: previously-checked
897        referenced subpattern not found".
898    
899    6.  Both MS Visual Studio and Symbian OS have problems with initializing
900        variables to point to external functions. For these systems, therefore,
901        pcre_malloc etc. are now initialized to local functions that call the
902        relevant global functions.
903    
904    7.  There were two entries missing in the vectors called coptable and poptable
905        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
906        I've fixed the data, and added a kludgy way of testing at compile time that
907        the lengths are correct (equal to the number of opcodes).
908    
909    8.  Following on from 7, I added a similar kludge to check the length of the
910        eint vector in pcreposix.c.
911    
912    9.  Error texts for pcre_compile() are held as one long string to avoid too
913        much relocation at load time. To find a text, the string is searched,
914        counting zeros. There was no check for running off the end of the string,
915        which could happen if a new error number was added without updating the
916        string.
917    
918    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
919    
920    11. \K was not working if it appeared in an atomic group or in a group that
921        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
922        \K is "not well defined" if used in an assertion. PCRE now accepts it if
923        the assertion is positive, but not if it is negative.
924    
925    12. Change 11 fortuitously reduced the size of the stack frame used in the
926        "match()" function of pcre_exec.c by one pointer. Forthcoming
927        implementation of support for (*MARK) will need an extra pointer on the
928        stack; I have reserved it now, so that the stack frame size does not
929        decrease.
930    
931    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
932        item in branch that calls a recursion is a subroutine call - as in the
933        second branch in the above example - was incorrectly given the compile-
934        time error "recursive call could loop indefinitely" because pcre_compile()
935        was not correctly checking the subroutine for matching a non-empty string.
936    
937    14. The checks for overrunning compiling workspace could trigger after an
938        overrun had occurred. This is a "should never occur" error, but it can be
939        triggered by pathological patterns such as hundreds of nested parentheses.
940        The checks now trigger 100 bytes before the end of the workspace.
941    
942    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
943    
944    
945    Version 8.01 19-Jan-2010
946    ------------------------
947    
948    1.  If a pattern contained a conditional subpattern with only one branch (in
949        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
950        computed the wrong minimum data length (which is of course zero for such
951        subpatterns). This could cause incorrect "no match" results.
952    
953    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
954        the pattern is reset in the first branch, pcre_compile() failed with
955        "internal error: code overflow at offset...". This happened only when
956        the reset was to the original external option setting. (An optimization
957        abstracts leading options settings into an external setting, which was the
958        cause of this.)
959    
960    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
961        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
962        assertion pattern did not match (meaning that the assertion was true), it
963        was incorrectly treated as false if the SKIP had been reached during the
964        matching. This also applied to assertions used as conditions.
965    
966    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
967        assertion subpattern, including such a pattern used as a condition,
968        unpredictable results occurred, instead of the error return
969        PCRE_ERROR_DFA_UITEM.
970    
971    5.  The C++ GlobalReplace function was not working like Perl for the special
972        situation when an empty string is matched. It now does the fancy magic
973        stuff that is necessary.
974    
975    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
976        removed. (These were left over from very, very early versions of PCRE.)
977    
978    7.  Some cosmetic changes to the code to make life easier when compiling it
979        as part of something else:
980    
981        (a) Change DEBUG to PCRE_DEBUG.
982    
983        (b) In pcre_compile(), rename the member of the "branch_chain" structure
984            called "current" as "current_branch", to prevent a collision with the
985            Linux macro when compiled as a kernel module.
986    
987        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
988            prevent a collision with the Linux macro when compiled as a kernel
989            module.
990    
991    8.  In pcre_compile() there are some checks for integer overflows that used to
992        cast potentially large values to (double). This has been changed to that
993        when building, a check for int64_t is made, and if it is found, it is used
994        instead, thus avoiding the use of floating point arithmetic. (There is no
995        other use of FP in PCRE.) If int64_t is not found, the fallback is to
996        double.
997    
998    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
999        2005 (difference between two addresses compared to an unsigned value).
1000    
1001    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1002        custom one, because of the following reported problem in Windows:
1003    
1004          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1005              under Win32.
1006          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1007              therefore missing the function definition.
1008          - The compiler thus generates a "C" signature for the test function.
1009          - The linker fails to find the "C" function.
1010          - PCRE fails to configure if asked to do so against libbz2.
1011    
1012    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1013        messages were output:
1014    
1015          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1016          rerunning libtoolize, to keep the correct libtool macros in-tree.
1017          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1018    
1019        I have done both of these things.
1020    
1021    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1022        most of the time, it *can* run out if it is given a pattern that contains a
1023        runaway infinite recursion. I updated the discussion in the pcrestack man
1024        page.
1025    
1026    13. Now that we have gone to the x.xx style of version numbers, the minor
1027        version may start with zero. Using 08 or 09 is a bad idea because users
1028        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1029        interpreted as invalid octal numbers. I've updated the previous comment in
1030        configure.ac, and also added a check that gives an error if 08 or 09 are
1031        used.
1032    
1033    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1034        causing partial matching to fail when the end of the subject matched \W
1035        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1036    
1037    15. There were some discrepancies between the declarations in pcre_internal.h
1038        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1039        their definitions. The declarations used "const uschar *" and the
1040        definitions used USPTR. Even though USPTR is normally defined as "const
1041        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1042        reported that: "This difference in casting confuses some C++ compilers, for
1043        example, SunCC recognizes above declarations as different functions and
1044        generates broken code for hbpcre." I have changed the declarations to use
1045        USPTR.
1046    
1047    16. GNU libtool is named differently on some systems. The autogen.sh script now
1048        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1049        (FreeBSD).
1050    
1051    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1052        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1053        comment: "Figure out how to create a longlong from a string: strtoll and
1054        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1055        instance, but it only takes 2 args instead of 3!"
1056    
1057    18. A subtle bug concerned with back references has been fixed by a change of
1058        specification, with a corresponding code fix. A pattern such as
1059        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1060        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1061        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1062        same bug. Such groups have to be quantified to be useful, or contained
1063        inside another quantified group. (If there's no repetition, the reference
1064        can never match.) The problem arises because, having left the group and
1065        moved on to the rest of the pattern, a later failure that backtracks into
1066        the group uses the captured value from the final iteration of the group
1067        rather than the correct earlier one. I have fixed this in PCRE by forcing
1068        any group that contains a reference to itself to be an atomic group; that
1069        is, there cannot be any backtracking into it once it has completed. This is
1070        similar to recursive and subroutine calls.
1071    
1072    
1073    Version 8.00 19-Oct-09
1074    ----------------------
1075    
1076    1.  The table for translating pcre_compile() error codes into POSIX error codes
1077        was out-of-date, and there was no check on the pcre_compile() error code
1078        being within the table. This could lead to an OK return being given in
1079        error.
1080    
1081    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1082        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1083        in a Windows environment.
1084    
1085    3.  The pcregrep --count option prints the count for each file even when it is
1086        zero, as does GNU grep. However, pcregrep was also printing all files when
1087        --files-with-matches was added. Now, when both options are given, it prints
1088        counts only for those files that have at least one match. (GNU grep just
1089        prints the file name in this circumstance, but including the count seems
1090        more useful - otherwise, why use --count?) Also ensured that the
1091        combination -clh just lists non-zero counts, with no names.
1092    
1093    4.  The long form of the pcregrep -F option was incorrectly implemented as
1094        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1095        but it seems right to fix it, and I didn't think it was worth preserving
1096        the old behaviour.
1097    
1098    5.  The command line items --regex=pattern and --regexp=pattern were not
1099        recognized by pcregrep, which required --regex pattern or --regexp pattern
1100        (with a space rather than an '='). The man page documented the '=' forms,
1101        which are compatible with GNU grep; these now work.
1102    
1103    6.  No libpcreposix.pc file was created for pkg-config; there was just
1104        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1105    
1106    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1107        when UCP support is not needed, by modifying the Python script that
1108        generates it from Unicode data files. This should not matter if the module
1109        is correctly used as a library, but I received one complaint about 50K of
1110        unwanted data. My guess is that the person linked everything into his
1111        program rather than using a library. Anyway, it does no harm.
1112    
1113    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1114        was a minimum greater than 1 for a wide character in a possessive
1115        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1116        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1117        character. Chaos in the form of incorrect output or a compiling loop could
1118        result.
1119    
1120    9.  The restrictions on what a pattern can contain when partial matching is
1121        requested for pcre_exec() have been removed. All patterns can now be
1122        partially matched by this function. In addition, if there are at least two
1123        slots in the offset vector, the offset of the earliest inspected character
1124        for the match and the offset of the end of the subject are set in them when
1125        PCRE_ERROR_PARTIAL is returned.
1126    
1127    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1128        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1129        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1130        and may be more useful for multi-segment matching.
1131    
1132    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1133        used to be given if ever the end of the subject was reached; now it is
1134        given only if matching could not proceed because another character was
1135        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1136        string "Z", which now yields "no match" instead of "partial match". In the
1137        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1138        final character ended with (*FAIL).
1139    
1140    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1141        if the pattern had a "must contain" character that was already found in the
1142        earlier partial match, unless partial matching was again requested. For
1143        example, with the pattern /dog.(body)?/, the "must contain" character is
1144        "g". If the first part-match was for the string "dog", restarting with
1145        "sbody" failed. This bug has been fixed.
1146    
1147    13. The string returned by pcre_dfa_exec() after a partial match has been
1148        changed so that it starts at the first inspected character rather than the
1149        first character of the match. This makes a difference only if the pattern
1150        starts with a lookbehind assertion or \b or \B (\K is not supported by
1151        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1152        matching functions compatible, and I think it's the right thing to do.
1153    
1154    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1155        so that the demonstration program is easily available in environments where
1156        PCRE has not been installed from source.
1157    
1158    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1159        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1160        library.
1161    
1162    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1163        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1164        is not the first non-POSIX option to be added. Clearly some people find
1165        these options useful.
1166    
1167    17. If a caller to the POSIX matching function regexec() passes a non-zero
1168        value for nmatch with a NULL value for pmatch, the value of
1169        nmatch is forced to zero.
1170    
1171    18. RunGrepTest did not have a test for the availability of the -u option of
1172        the diff command, as RunTest does. It now checks in the same way as
1173        RunTest, and also checks for the -b option.
1174    
1175    19. If an odd number of negated classes containing just a single character
1176        interposed, within parentheses, between a forward reference to a named
1177        subpattern and the definition of the subpattern, compilation crashed with
1178        an internal error, complaining that it could not find the referenced
1179        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1180        [The bug was that it was starting one character too far in when skipping
1181        over the character class, thus treating the ] as data rather than
1182        terminating the class. This meant it could skip too much.]
1183    
1184    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1185        /g option in pcretest when the pattern contains \K, which makes it possible
1186        to have an empty string match not at the start, even when the pattern is
1187        anchored. Updated pcretest and pcredemo to use this option.
1188    
1189    21. If the maximum number of capturing subpatterns in a recursion was greater
1190        than the maximum at the outer level, the higher number was returned, but
1191        with unset values at the outer level. The correct (outer level) value is
1192        now given.
1193    
1194    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1195        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1196        make it do so. The string so far is captured, making this feature
1197        compatible with Perl.
1198    
1199    23. The tests have been re-organized, adding tests 11 and 12, to make it
1200        possible to check the Perl 5.10 features against Perl 5.10.
1201    
1202    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1203        pattern matches a fixed length string. PCRE did not allow this; now it
1204        does. Neither allows recursion.
1205    
1206    25. I finally figured out how to implement a request to provide the minimum
1207        length of subject string that was needed in order to match a given pattern.
1208        (It was back references and recursion that I had previously got hung up
1209        on.) This code has now been added to pcre_study(); it finds a lower bound
1210        to the length of subject needed. It is not necessarily the greatest lower
1211        bound, but using it to avoid searching strings that are too short does give
1212        some useful speed-ups. The value is available to calling programs via
1213        pcre_fullinfo().
1214    
1215    26. While implementing 25, I discovered to my embarrassment that pcretest had
1216        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1217        study optimizations had never been tested with that matching function.
1218        Oops. What is worse, even when it was passed study data, there was a bug in
1219        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1220        were also very few tests of studied patterns with pcre_dfa_exec().
1221    
1222    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1223        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1224        on the other side of the coin, they are no longer allowed to have different
1225        names, because these cannot be distinguished in PCRE, and this has caused
1226        confusion. (This is a difference from Perl.)
1227    
1228    28. When duplicate subpattern names are present (necessarily with different
1229        numbers, as required by 27 above), and a test is made by name in a
1230        conditional pattern, either for a subpattern having been matched, or for
1231        recursion in such a pattern, all the associated numbered subpatterns are
1232        tested, and the overall condition is true if the condition is true for any
1233        one of them. This is the way Perl works, and is also more like the way
1234        testing by number works.
1235    
1236    
1237    Version 7.9 11-Apr-09
1238    ---------------------
1239    
1240    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1241        (pcretest), all targets were linked against these libraries. This included
1242        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1243        libraries. This caused unwanted dependencies to be created. This problem
1244        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1245        pcretest is linked with readline.
1246    
1247    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1248        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1249        moved outside it again, because FALSE and TRUE are already defined in AIX,
1250        but BOOL is not.
1251    
1252    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1253        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1254    
1255    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1256        hyphens) following file names and line numbers when outputting matching
1257        lines. This is not true; no spaces are inserted. I have also clarified the
1258        wording for the --colour (or --color) option.
1259    
1260    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1261        was not coloured; this is different to GNU grep, so I have changed it to be
1262        the same.
1263    
1264    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1265        each matching line was coloured. Now it goes on to look for further matches
1266        of any of the test patterns, which is the same behaviour as GNU grep.
1267    
1268    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1269        doesn't make sense to accept an empty string match in pcregrep, so I have
1270        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1271        seems to be how GNU grep behaves.
1272    
1273    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1274        start or after a newline", because the conditional assertion was not being
1275        correctly handled. The rule now is that both the assertion and what follows
1276        in the first alternative must satisfy the test.
1277    
1278    9.  If auto-callout was enabled in a pattern with a conditional group whose
1279        condition was an assertion, PCRE could crash during matching, both with
1280        pcre_exec() and pcre_dfa_exec().
1281    
1282    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1283        used for matching.
1284    
1285    11. Unicode property support in character classes was not working for
1286        characters (bytes) greater than 127 when not in UTF-8 mode.
1287    
1288    12. Added the -M command line option to pcretest.
1289    
1290    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1291    
1292    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1293    
1294    16. Added comments and documentation about mis-use of no_arg in the C++
1295        wrapper.
1296    
1297    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1298        from Martin Jerabek that uses macro names for all relevant character and
1299        string constants.
1300    
1301    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1302        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1303        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1304        these, but not everybody uses configure.
1305    
1306    19. A conditional group that had only one branch was not being correctly
1307        recognized as an item that could match an empty string. This meant that an
1308        enclosing group might also not be so recognized, causing infinite looping
1309        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1310        with the subject "ab", where knowledge that the repeated group can match
1311        nothing is needed in order to break the loop.
1312    
1313    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1314        exec(), but without supplying a callout function, matching went wrong.
1315    
1316    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1317        leak if the size of the offset vector was greater than 30. When the vector
1318        is smaller, the saved offsets during recursion go onto a local stack
1319        vector, but for larger vectors malloc() is used. It was failing to free
1320        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1321        error, in fact).
1322    
1323    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1324        heapframe that is used only when UTF-8 support is enabled. This caused no
1325        problem, but was untidy.
1326    
1327    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1328        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1329        included within another project.
1330    
1331    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1332        slightly modified by me:
1333    
1334          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1335              not building pcregrep.
1336    
1337          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1338              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1339    
1340    25. Forward references, both numeric and by name, in patterns that made use of
1341        duplicate group numbers, could behave incorrectly or give incorrect errors,
1342        because when scanning forward to find the reference group, PCRE was not
1343        taking into account the duplicate group numbers. A pattern such as
1344        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1345    
1346    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1347        the feature of a custom pointer more persuasive (as requested by a user).
1348    
1349    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1350        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1351        pre-defined.
1352    
1353    28. Added support for (*UTF8) at the start of a pattern.
1354    
1355    29. Arrange for flags added by the "release type" setting in CMake to be shown
1356        in the configuration summary.
1357    
1358    
1359    Version 7.8 05-Sep-08
1360    ---------------------
1361    
1362    1.  Replaced UCP searching code with optimized version as implemented for Ad
1363        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1364        stage table and inline lookup instead of a function, giving speed ups of 2
1365        to 5 times on some simple patterns that I tested. Permission was given to
1366        distribute the MultiStage2.py script that generates the tables (it's not in
1367        the tarball, but is in the Subversion repository).
1368    
1369    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1370        scripts.
1371    
1372    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1373        a group with a zero qualifier. The result of the study could be incorrect,
1374        or the function might crash, depending on the pattern.
1375    
1376    4.  Caseless matching was not working for non-ASCII characters in back
1377        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1378        It now works when Unicode Property Support is available.
1379    
1380    5.  In pcretest, an escape such as \x{de} in the data was always generating
1381        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1382        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1383        truncation.
1384    
1385    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1386    
1387    7.  Added two (int) casts to pcregrep when printing the difference of two
1388        pointers, in case they are 64-bit values.
1389    
1390    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1391        test 2 if it fails.
1392    
1393    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1394        and a #define of that name to empty if it is not externally set. This is to
1395        allow users of MSVC to set it if necessary.
1396    
1397    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1398        the convenience functions in the pcre_get.c source file.
1399    
1400    11. An option change at the start of a pattern that had top-level alternatives
1401        could cause overwriting and/or a crash. This command provoked a crash in
1402        some environments:
1403    
1404          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1405    
1406        This potential security problem was recorded as CVE-2008-2371.
1407    
1408    12. For a pattern where the match had to start at the beginning or immediately
1409        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1410        pcre_dfa_exec() could read past the end of the passed subject if there was
1411        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1412        pcretest so that it places the subject at the end of its malloc-ed buffer.
1413    
1414    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1415        exec() might read past the end of the data buffer in UTF-8 mode.
1416    
1417    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1418        the data contained the byte 0x85 as part of a UTF-8 character within its
1419        first line. This applied both to normal and DFA matching.
1420    
1421    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1422        /^[^d]*?$/8 failed to match "abc".
1423    
1424    16. Added a missing copyright notice to pcrecpp_internal.h.
1425    
1426    17. Make it more clear in the documentation that values returned from
1427        pcre_exec() in ovector are byte offsets, not character counts.
1428    
1429    18. Tidied a few places to stop certain compilers from issuing warnings.
1430    
1431    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1432        supplied by Stefan Weber. I made a further small update for 7.8 because
1433        there is a change of source arrangements: the pcre_searchfuncs.c module is
1434        replaced by pcre_ucd.c.
1435    
1436    
1437    Version 7.7 07-May-08
1438    ---------------------
1439    
1440    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1441        a string to a long long, pretend we don't even have a long long." This is
1442        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1443    
1444    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1445        pre-7.6 versions, which defined a global no_arg variable instead of putting
1446        it in the RE class. (See also #8 below.)
1447    
1448    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1449        Lopes.
1450    
1451    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1452    
1453        (1) The include/exclude patterns were being applied to the whole pathnames
1454            of files, instead of just to the final components.
1455    
1456        (2) If there was more than one level of directory, the subdirectories were
1457            skipped unless they satisfied the include/exclude conditions. This is
1458            inconsistent with GNU grep (and could even be seen as contrary to the
1459            pcregrep specification - which I improved to make it absolutely clear).
1460            The action now is always to scan all levels of directory, and just
1461            apply the include/exclude patterns to regular files.
1462    
1463    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1464        --exclude_dir in the tests to avoid scanning .svn directories.
1465    
1466    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1467        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1468        doesn't support NULs in patterns.
1469    
1470    7.  Added some missing "const"s to declarations of static tables in
1471        pcre_compile.c and pcre_dfa_exec.c.
1472    
1473    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1474        caused by fix #2  above. (Subsequently also a second patch to fix the
1475        first patch. And a third patch - this was a messy problem.)
1476    
1477    9.  Applied Craig's patch to remove the use of push_back().
1478    
1479    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1480        matching function regexec().
1481    
1482    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1483        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1484        references. PCRE supports relative numbers with this syntax (I don't think
1485        Oniguruma does).
1486    
1487    12. Previously, a group with a zero repeat such as (...){0} was completely
1488        omitted from the compiled regex. However, this means that if the group
1489        was called as a subroutine from elsewhere in the pattern, things went wrong
1490        (an internal error was given). Such groups are now left in the compiled
1491        pattern, with a new opcode that causes them to be skipped at execution
1492        time.
1493    
1494    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1495        to the way PCRE behaves:
1496    
1497        (a) A lone ] character is dis-allowed (Perl treats it as data).
1498    
1499        (b) A back reference to an unmatched subpattern matches an empty string
1500            (Perl fails the current match path).
1501    
1502        (c) A data ] in a character class must be notated as \] because if the
1503            first data character in a class is ], it defines an empty class. (In
1504            Perl it is not possible to have an empty class.) The empty class []
1505            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1506            The negative empty class [^] matches any one character, independently
1507            of the DOTALL setting.
1508    
1509    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1510        non-existent subpattern following a character class starting with ']' and
1511        containing () gave an internal compiling error instead of "reference to
1512        non-existent subpattern". Fortunately, when the pattern did exist, the
1513        compiled code was correct. (When scanning forwards to check for the
1514        existencd of the subpattern, it was treating the data ']' as terminating
1515        the class, so got the count wrong. When actually compiling, the reference
1516        was subsequently set up correctly.)
1517    
1518    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1519        it was being rejected as not supported by pcre_dfa_exec(), even though
1520        other assertions are supported. I have made pcre_dfa_exec() support
1521        (*FAIL).
1522    
1523    16. The implementation of 13c above involved the invention of a new opcode,
1524        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1525        cannot be changed at match time, I realized I could make a small
1526        improvement to matching performance by compiling OP_ALLANY instead of
1527        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1528        on the OP_ANY path.
1529    
1530    17. Compiling pcretest on Windows with readline support failed without the
1531        following two fixes: (1) Make the unistd.h include conditional on
1532        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1533    
1534    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1535        ncurses library to be included for pcretest when ReadLine support is
1536        requested, but also to allow for it to be overridden. This patch came from
1537        Daniel Bergström.
1538    
1539    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1540        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1541        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1542        spotting this.
1543    
1544    
1545    Version 7.6 28-Jan-08
1546    ---------------------
1547    
1548    1.  A character class containing a very large number of characters with
1549        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1550        overflow.
1551    
1552    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1553        HAVE_LONG_LONG is not defined.
1554    
1555    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1556        bring them up to date and include new features. This patch includes:
1557    
1558        - Fixed PH's badly added libz and libbz2 support.
1559        - Fixed a problem with static linking.
1560        - Added pcredemo. [But later removed - see 7 below.]
1561        - Fixed dftables problem and added an option.
1562        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1563            HAVE_LONG_LONG.
1564        - Added readline support for pcretest.
1565        - Added an listing of the option settings after cmake has run.
1566    
1567    4.  A user submitted a patch to Makefile that makes it easy to create
1568        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1569        Makefile.am that cause it to include this special target, without
1570        affecting anything else. Note that the same mingw target plus all
1571        the other distribution libraries and programs are now supported
1572        when configuring with CMake (see 6 below) instead of with
1573        Configure/Make.
1574    
1575    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1576        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1577        exported in the Windows port". It has not yet been confirmed that the patch
1578        solves the problem, but it does no harm.
1579    
1580    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1581        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1582        with CMake, and also correct the comment about stack recursion.
1583    
1584    7.  Remove the automatic building of pcredemo from the ./configure system and
1585        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1586        of a program that users should build themselves after PCRE is installed, so
1587        building it automatically is not really right. What is more, it gave
1588        trouble in some build environments.
1589    
1590    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1591    
1592    
1593    Version 7.5 10-Jan-08
1594    ---------------------
1595    
1596    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1597        values in parens when parsing an RE using the C++ wrapper."
1598    
1599    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1600        Characters greater than 255 were excluded from the class instead of being
1601        included.
1602    
1603    3.  The same bug as (2) above applied to negated POSIX classes such as
1604        [:^space:].
1605    
1606    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1607        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1608        I have changed it.
1609    
1610    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1611        first named subpattern) and a construct such as (?&a) would reference the
1612        first named subpattern whose name started with "a" (in other words, the
1613        length check was missing). Both these problems are fixed. "Subpattern name
1614        expected" is now given for (?&) (a zero-length name), and this patch also
1615        makes it give the same error for \k'' (previously it complained that that
1616        was a reference to a non-existent subpattern).
1617    
1618    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1619        this is right because (?- can be followed by option settings as well as by
1620        digits. I have, however, made the messages clearer.
1621    
1622    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1623        than the number used in the conditional) now cause a compile-time error.
1624        This is actually not compatible with Perl, which accepts such patterns, but
1625        treats the conditional as always being FALSE (as PCRE used to), but it
1626        seems to me that giving a diagnostic is better.
1627    
1628    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1629        and messages.
1630    
1631    9.  Fix two occurrences of "backslash" in comments that should have been
1632        "backspace".
1633    
1634    10. Remove two redundant lines of code that can never be obeyed (their function
1635        was moved elsewhere).
1636    
1637    11. The program that makes PCRE's Unicode character property table had a bug
1638        which caused it to generate incorrect table entries for sequences of
1639        characters that have the same character type, but are in different scripts.
1640        It amalgamated them into a single range, with the script of the first of
1641        them. In other words, some characters were in the wrong script. There were
1642        thirteen such cases, affecting characters in the following ranges:
1643    
1644          U+002b0 - U+002c1
1645          U+0060c - U+0060d
1646          U+0061e - U+00612
1647          U+0064b - U+0065e
1648          U+0074d - U+0076d
1649          U+01800 - U+01805
1650          U+01d00 - U+01d77
1651          U+01d9b - U+01dbf
1652          U+0200b - U+0200f
1653          U+030fc - U+030fe
1654          U+03260 - U+0327f
1655          U+0fb46 - U+0fbb1
1656          U+10450 - U+1049d
1657    
1658    12. The -o option (show only the matching part of a line) for pcregrep was not
1659        compatible with GNU grep in that, if there was more than one match in a
1660        line, it showed only the first of them. It now behaves in the same way as
1661        GNU grep.
1662    
1663    13. If the -o and -v options were combined for pcregrep, it printed a blank
1664        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1665        does the same. The return code can be used to tell if there were any
1666        non-matching lines.
1667    
1668    14. Added --file-offsets and --line-offsets to pcregrep.
1669    
1670    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1671        infinitely looping recursion. The bug was that positive lookaheads were not
1672        being skipped when checking for a possible empty match (negative lookaheads
1673        and both kinds of lookbehind were skipped).
1674    
1675    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1676        inclusion of <windows.h> to before rather than after the definition of
1677        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1678    
1679    17. Specifying a possessive quantifier with a specific limit for a Unicode
1680        character property caused pcre_compile() to compile bad code, which led at
1681        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1682        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1683        caused the error; without that there was no problem.
1684    
1685    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1686    
1687    19. Added --enable-pcretest-libreadline.
1688    
1689    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1690        RE::GlobalReplace(). As a result, the number of replacements returned was
1691        double what it should be. I removed one of the increments, but Craig sent a
1692        later patch that removed the other one (the right fix) and added unit tests
1693        that check the return values (which was not done before).
1694    
1695    21. Several CMake things:
1696    
1697        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1698            the names libpcre and libpcreposix, not just pcre and pcreposix.
1699    
1700        (2) The above change means that pcretest and pcregrep are now correctly
1701            linked with the newly-built libraries, not previously installed ones.
1702    
1703        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1704    
1705    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1706        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1707        UTF-8 newline character). The key issue is that the pattern starts .*;
1708        this means that the match must be either at the beginning, or after a
1709        newline. The bug was in the code for advancing after a failed match and
1710        checking that the new position followed a newline. It was not taking
1711        account of UTF-8 characters correctly.
1712    
1713    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1714        character classes. PCRE was not treating the sequence [:...:] as a
1715        character class unless the ... were all letters. Perl, however, seems to
1716        allow any characters between [: and :], though of course it rejects as
1717        unknown any "names" that contain non-letters, because all the known class
1718        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1719        for example, whereas PCRE did not - it did not recognize a POSIX character
1720        class. This seemed a bit dangerous, so the code has been changed to be
1721        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1722        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1723        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1724        Perl does, and where it didn't before.
1725    
1726    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1727        Windows environments %n is disabled by default.
1728    
1729    
1730    Version 7.4 21-Sep-07
1731    ---------------------
1732    
1733    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1734        means that a class such as [\s] counted as "explicit reference to CR or
1735        LF". That isn't really right - the whole point of the change was to try to
1736        help when there was an actual mention of one of the two characters. So now
1737        the change happens only if \r or \n (or a literal CR or LF) character is
1738        encountered.
1739    
1740    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1741        of both had grown to the point where there were only 3 bits left.
1742        Fortunately, there was spare space in the data structure, and so I have
1743        moved the internal flags into a new 16-bit field to free up more option
1744        bits.
1745    
1746    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1747        but did not set the internal JCHANGED flag - either of these is enough to
1748        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1749        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1750        start sets both bits.
1751    
1752    4.  Added options (at build time, compile time, exec time) to change \R from
1753        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1754    
1755    5.  doc/pcresyntax.html was missing from the distribution.
1756    
1757    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1758        compatibility, even though it is no longer used.
1759    
1760    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1761        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1762        windows.h file is present (where different names are used). [This was
1763        reversed later after testing - see 16 below.]
1764    
1765    8.  Changed all #include <config.h> to #include "config.h". There were also
1766        some further <pcre.h> cases that I changed to "pcre.h".
1767    
1768    9.  When pcregrep was used with the --colour option, it missed the line ending
1769        sequence off the lines that it output.
1770    
1771    10. It was pointed out to me that arrays of string pointers cause lots of
1772        relocations when a shared library is dynamically loaded. A technique of
1773        using a single long string with a table of offsets can drastically reduce
1774        these. I have refactored PCRE in four places to do this. The result is
1775        dramatic:
1776    
1777          Originally:                          290
1778          After changing UCP table:            187
1779          After changing error message table:   43
1780          After changing table of "verbs"       36
1781          After changing table of Posix names   22
1782    
1783        Thanks to the folks working on Gregex for glib for this insight.
1784    
1785    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1786        unicode-properties was also set.
1787    
1788    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1789    
1790    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1791        checked only for CRLF.
1792    
1793    14. Added casts to pcretest.c to avoid compiler warnings.
1794    
1795    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1796    
1797    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1798        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1799        entirely. This removes changes made in 7 above.
1800    
1801    17. The CMake files have been updated, and there is now more information about
1802        building with CMake in the NON-UNIX-USE document.
1803    
1804    
1805    Version 7.3 28-Aug-07
1806    ---------------------
1807    
1808     1. In the rejigging of the build system that eventually resulted in 7.1, the
1809        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1810        brackets there is not right, since it causes compilers to look for an
1811        installed pcre.h, not the version that is in the source that is being
1812        compiled (which of course may be different). I have changed it back to:
1813    
1814          #include "pcre.h"
1815    
1816        I have a vague recollection that the change was concerned with compiling in
1817        different directories, but in the new build system, that is taken care of
1818        by the VPATH setting the Makefile.
1819    
1820     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1821        when the subject happened to end in the byte 0x85 (e.g. if the last
1822        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1823        characters but of course it shouldn't be taken as a newline when it is part
1824        of another character. The bug was that, for an unlimited repeat of . in
1825        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1826        characters when looking for a newline.
1827    
1828     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1829    
1830     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1831        in debug output.
1832    
1833     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1834        long printing in the pcrecpp unittest when running under MinGW.
1835    
1836     6. ESC_K was left out of the EBCDIC table.
1837    
1838     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1839        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1840        limit also applies to "virtual nesting" when a pattern is recursive, and in
1841        this case 1000 isn't so big. I have been able to remove this limit at the
1842        expense of backing off one optimization in certain circumstances. Normally,
1843        when pcre_exec() would call its internal match() function recursively and
1844        immediately return the result unconditionally, it uses a "tail recursion"
1845        feature to save stack. However, when a subpattern that can match an empty
1846        string has an unlimited repetition quantifier, it no longer makes this
1847        optimization. That gives it a stack frame in which to save the data for
1848        checking that an empty string has been matched. Previously this was taken
1849        from the 1000-entry workspace that had been reserved. So now there is no
1850        explicit limit, but more stack is used.
1851    
1852     8. Applied Daniel's patches to solve problems with the import/export magic
1853        syntax that is required for Windows, and which was going wrong for the
1854        pcreposix and pcrecpp parts of the library. These were overlooked when this
1855        problem was solved for the main library.
1856    
1857     9. There were some crude static tests to avoid integer overflow when computing
1858        the size of patterns that contain repeated groups with explicit upper
1859        limits. As the maximum quantifier is 65535, the maximum group length was
1860        set at 30,000 so that the product of these two numbers did not overflow a
1861        32-bit integer. However, it turns out that people want to use groups that
1862        are longer than 30,000 bytes (though not repeat them that many times).
1863        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1864        made it possible to implement the integer overflow checks in a much more
1865        dynamic way, which I have now done. The artificial limitation on group
1866        length has been removed - we now have only the limit on the total length of
1867        the compiled pattern, which depends on the LINK_SIZE setting.
1868    
1869    10. Fixed a bug in the documentation for get/copy named substring when
1870        duplicate names are permitted. If none of the named substrings are set, the
1871        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1872        empty string.
1873    
1874    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1875        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1876        because the ] is interpreted as the first data character and the
1877        terminating ] is not found. PCRE has been made compatible with Perl in this
1878        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1879        cause memory overwriting.
1880    
1881    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1882        string has been matched (to stop an infinite loop). It was not recognizing
1883        a conditional subpattern that could match an empty string if that
1884        subpattern was within another subpattern. For example, it looped when
1885        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1886        condition was not nested. This bug has been fixed.
1887    
1888    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1889        past the start of the subject in the presence of bytes with the top bit
1890        set, for example "\x8aBCD".
1891    
1892    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1893        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1894    
1895    14. Optimized (?!) to (*FAIL).
1896    
1897    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1898        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1899        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1900        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1901        does: it's just the validity check that is more restrictive.
1902    
1903    16. Inserted checks for integer overflows during escape sequence (backslash)
1904        processing, and also fixed erroneous offset values for syntax errors during
1905        backslash processing.
1906    
1907    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1908        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1909    
1910    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1911        caused an overrun.
1912    
1913    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1914        something other than just ASCII characters) inside a group that had an
1915        unlimited repeat caused a loop at compile time (while checking to see
1916        whether the group could match an empty string).
1917    
1918    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1919        [\P{Any}] did so. (Error in the code for printing property names.)
1920    
1921    21. An orphan \E inside a character class could cause a crash.
1922    
1923    22. A repeated capturing bracket such as (A)? could cause a wild memory
1924        reference during compilation.
1925    
1926    23. There are several functions in pcre_compile() that scan along a compiled
1927        expression for various reasons (e.g. to see if it's fixed length for look
1928        behind). There were bugs in these functions when a repeated \p or \P was
1929        present in the pattern. These operators have additional parameters compared
1930        with \d, etc, and these were not being taken into account when moving along
1931        the compiled data. Specifically:
1932    
1933        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1934            length.
1935    
1936        (b) An item such as \pL+ within a repeated group could cause crashes or
1937            loops.
1938    
1939        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1940            "reference to non-existent subpattern" error.
1941    
1942        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1943    
1944    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1945        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1946    
1947    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1948    
1949    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1950        character were causing crashes (broken optimization).
1951    
1952    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1953        \p or \P) caused a compile-time loop.
1954    
1955    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1956        break. For example, the unstudied pattern [\r\n]A does not match the string
1957        "\r\nA" because change 7.0/46 below moves the current point on by two
1958        characters after failing to match at the start. However, the pattern \nA
1959        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1960        the same is true. There doesn't seem any very clean way out of this, but
1961        what I have chosen to do makes the common cases work: PCRE now takes note
1962        of whether there can be an explicit match for \r or \n anywhere in the
1963        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1964        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1965        pattern has explicit CR or LF references.
1966    
1967    29. Added (*CR) etc for changing newline setting at start of pattern.
1968    
1969    
1970    Version 7.2 19-Jun-07
1971    ---------------------
1972    
1973     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1974        which is apparently normally available under Windows.
1975    
1976     2. Re-jig the pcregrep tests with different newline settings in an attempt
1977        to make them independent of the local environment's newline setting.
1978    
1979     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1980    
1981     4. Some of the "internals" tests were previously cut out when the link size
1982        was not 2, because the output contained actual offsets. The recent new
1983        "Z" feature of pcretest means that these can be cut out, making the tests
1984        usable with all link sizes.
1985    
1986     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1987        stack recursion. This gives a massive performance boost under BSD, but just
1988        a small improvement under Linux. However, it saves one field in the frame
1989        in all cases.
1990    
1991     6. Added more features from the forthcoming Perl 5.10:
1992    
1993        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1994            recursion call. It refers to the nth most recently opened parentheses.
1995    
1996        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1997            to be opened parentheses.
1998    
1999        (c) Conditions that refer to capturing parentheses can be specified
2000            relatively, for example, (?(-2)... or (?(+3)...
2001    
2002        (d) \K resets the start of the current match so that everything before
2003            is not part of it.
2004    
2005        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2006    
2007        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2008            reference syntax.
2009    
2010        (g) (?| introduces a group in which the numbering of parentheses in each
2011            alternative starts with the same number.
2012    
2013        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2014    
2015     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2016        PCRE_INFO_JCHANGED.
2017    
2018     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2019        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2020        for detecting groups that can match an empty string.
2021    
2022     9. A pattern with a very large number of alternatives (more than several
2023        hundred) was running out of internal workspace during the pre-compile
2024        phase, where pcre_compile() figures out how much memory will be needed. A
2025        bit of new cunning has reduced the workspace needed for groups with
2026        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2027        workspace instead of running out of the 4096 that are available.
2028    
2029    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2030    
2031    11. Applied patch from Google to remove an optimization that didn't quite work.
2032        The report of the bug said:
2033    
2034          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2035          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2036          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2037    
2038    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2039        it matched the wrong number of bytes.
2040    
2041    
2042    Version 7.1 24-Apr-07
2043    ---------------------
2044    
2045     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2046        that is more "standard", making use of automake and other Autotools. There
2047        is some re-arrangement of the files and adjustment of comments consequent
2048        on this.
2049    
2050     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2051        for recursive directory scanning broke on some systems because the files
2052        are not scanned in any specific order and on different systems the order
2053        was different. A call to "sort" has been inserted into RunGrepTest for the
2054        approprate test as a short-term fix. In the longer term there may be an
2055        alternative.
2056    
2057     3. I had an email from Eric Raymond about problems translating some of PCRE's
2058        man pages to HTML (despite the fact that I distribute HTML pages, some
2059        people do their own conversions for various reasons). The problems
2060        concerned the use of low-level troff macros .br and .in. I have therefore
2061        removed all such uses from the man pages (some were redundant, some could
2062        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2063        HTML has been updated to handle .nf/.fi and to complain if it encounters
2064        .br or .in.
2065    
2066     4. Updated comments in configure.ac that get placed in config.h.in and also
2067        arranged for config.h to be included in the distribution, with the name
2068        config.h.generic, for the benefit of those who have to compile without
2069        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2070    
2071     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2072        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2073        makevp.bat for the current PCRE, using the additional files
2074        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2075    
2076     6. A Windows user reported a minor discrepancy with test 2, which turned out
2077        to be caused by a trailing space on an input line that had got lost in his
2078        copy. The trailing space was an accident, so I've just removed it.
2079    
2080     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2081        that is needed.
2082    
2083     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2084        as "const" (a) because they are and (b) because it helps the PHP
2085        maintainers who have recently made a script to detect big data structures
2086        in the php code that should be moved to the .rodata section. I remembered
2087        to update Builducptable as well, so it won't revert if ucptable.h is ever
2088        re-created.
2089    
2090     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2091        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2092        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2093        support is not required. This saves 1.5-2K of code, which is important in
2094        some applications.
2095    
2096        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2097        so as not to refer to the tables, even though these functions will never be
2098        called when UTF-8 support is disabled. Otherwise there are problems with a
2099        shared library.
2100    
2101    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2102    
2103        (a) It was defining its arguments as char * instead of void *.
2104    
2105        (b) It was assuming that all moves were upwards in memory; this was true
2106            a long time ago when I wrote it, but is no longer the case.
2107    
2108        The emulated memove() is provided for those environments that have neither
2109        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2110        is clearly not the case, as these two bugs were recently reported.
2111    
2112    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2113        and Detrail to create the HTML documentation, the .txt form of the man
2114        pages, and it removes trailing spaces from listed files. It also creates
2115        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2116        case, it wraps all the #defines with #ifndefs. This script should be run
2117        before "make dist".
2118    
2119    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2120        with Unicode property support.
2121    
2122        (a) For a maximizing quantifier, if the two different cases of the
2123            character were of different lengths in their UTF-8 codings (there are
2124            some cases like this - I found 11), and the matching function had to
2125            back up over a mixture of the two cases, it incorrectly assumed they
2126            were both the same length.
2127    
2128        (b) When PCRE was configured to use the heap rather than the stack for
2129            recursion during matching, it was not correctly preserving the data for
2130            the other case of a UTF-8 character when checking ahead for a match
2131            while processing a minimizing repeat. If the check also involved
2132            matching a wide character, but failed, corruption could cause an
2133            erroneous result when trying to check for a repeat of the original
2134            character.
2135    
2136    13. Some tidying changes to the testing mechanism:
2137    
2138        (a) The RunTest script now detects the internal link size and whether there
2139            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2140            values substituted by "configure". (The RunGrepTest script already did
2141            this for UTF-8.) The configure.ac script no longer substitutes the
2142            relevant variables.
2143    
2144        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2145            with length and offset values. This means that the output is different
2146            for different internal link sizes. Test 2 is skipped for link sizes
2147            other than 2 because of this, bypassing the problem. Unfortunately,
2148            there was also a test in test 3 (the locale tests) that used /B and
2149            failed for link sizes other than 2. Rather than cut the whole test out,
2150            I have added a new /Z option to pcretest that replaces the length and
2151            offset values with spaces. This is now used to make test 3 independent
2152            of link size. (Test 2 will be tidied up later.)
2153    
2154    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2155        segmentation fault instead of returning the appropriate error message.
2156    
2157    15. In multiline mode when the newline sequence was set to "any", the pattern
2158        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2159        This doesn't seem right; it now treats the CRLF combination as the line
2160        ending, and so does not match in that case. It's only a pattern such as ^$
2161        that would hit this one: something like ^ABC$ would have failed after \r
2162        and then tried again after \r\n.
2163    
2164    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2165        in an attempt to make files that differ only in their line terminators
2166        compare equal. This works on Linux.
2167    
2168    17. Under certain error circumstances pcregrep might try to free random memory
2169        as it exited. This is now fixed, thanks to valgrind.
2170    
2171    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2172        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2173        was because its rules for how to advance for /g after matching an empty
2174        string at the end of a line did not allow for this case. They now check for
2175        it specially.
2176    
2177    20. pcretest is supposed to handle patterns and data of any length, by
2178        extending its buffers when necessary. It was getting this wrong when the
2179        buffer for a data line had to be extended.
2180    
2181    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2182        CRLF as a newline sequence.
2183    
2184    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2185        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2186        I have nevertheless tidied it up.
2187    
2188    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2189    
2190    24. Added a man page for pcre-config.
2191    
2192    
2193    Version 7.0 19-Dec-06
2194    ---------------------
2195    
2196     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2197        moving to gcc 4.1.1.
2198    
2199     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2200        sys/time.h, which is documented as needed for this function. It doesn't
2201        seem to matter on Linux, but it showed up on some releases of OS X.
2202    
2203     3. It seems that there are systems where bytes whose values are greater than
2204        127 match isprint() in the "C" locale. The "C" locale should be the
2205        default when a C program starts up. In most systems, only ASCII printing
2206        characters match isprint(). This difference caused the output from pcretest
2207        to vary, making some of the tests fail. I have changed pcretest so that:
2208    
2209        (a) When it is outputting text in the compiled version of a pattern, bytes
2210            other than 32-126 are always shown as hex escapes.
2211    
2212        (b) When it is outputting text that is a matched part of a subject string,
2213            it does the same, unless a different locale has been set for the match
2214            (using the /L modifier). In this case, it uses isprint() to decide.
2215    
2216     4. Fixed a major bug that caused incorrect computation of the amount of memory
2217        required for a compiled pattern when options that changed within the
2218        pattern affected the logic of the preliminary scan that determines the
2219        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2220        that the computed length was too small. The symptoms of this bug were
2221        either the PCRE error "internal error: code overflow" from pcre_compile(),
2222        or a glibc crash with a message such as "pcretest: free(): invalid next
2223        size (fast)". Examples of patterns that provoked this bug (shown in
2224        pcretest format) are:
2225    
2226          /(?-x: )/x
2227          /(?x)(?-x: \s*#\s*)/
2228          /((?i)[\x{c0}])/8
2229          /(?i:[\x{c0}])/8
2230    
2231        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2232        is now done differently.
2233    
2234     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2235        wrapper classes; (b) implement a new function in the C++ scanner that is
2236        more efficient than the old way of doing things because it avoids levels of
2237        recursion in the regex matching; (c) add a paragraph to the documentation
2238        for the FullMatch() function.
2239    
2240     6. The escape sequence \n was being treated as whatever was defined as
2241        "newline". Not only was this contrary to the documentation, which states
2242        that \n is character 10 (hex 0A), but it also went horribly wrong when
2243        "newline" was defined as CRLF. This has been fixed.
2244    
2245     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2246        was being set to -1 for the "end of line" case (supposedly a value that no
2247        character can have). Though this value is never used (the check for end of
2248        line is "zero bytes in current character"), it caused compiler complaints.
2249        I've changed it to 0xffffffff.
2250    
2251     8. In pcre_version.c, the version string was being built by a sequence of
2252        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2253        string (as it is for production releases) called a macro with an empty
2254        argument. The C standard says the result of this is undefined. The gcc
2255        compiler treats it as an empty string (which was what was wanted) but it is
2256        reported that Visual C gives an error. The source has been hacked around to
2257        avoid this problem.
2258    
2259     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2260        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2261        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2262        of them did).
2263    
2264    10. Originally, pcretest opened its input and output without "b"; then I was
2265        told that "b" was needed in some environments, so it was added for release
2266        5.0 to both the input and output. (It makes no difference on Unix-like
2267        systems.) Later I was told that it is wrong for the input on Windows. I've
2268        now abstracted the modes into two macros, to make it easier to fiddle with
2269        them, and removed "b" from the input mode under Windows.
2270    
2271    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2272    
2273    12. Added -help and --help to pcretest as an official way of being reminded
2274        of the options.
2275    
2276    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2277        and pcrecpp.cc because they annoy compilers at high warning levels.
2278    
2279    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2280    
2281    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2282        scripts are not C programs :-) and which was not noticed because it works
2283        on Linux.
2284    
2285    16. pcretest is supposed to handle any length of pattern and data line (as one
2286        line or as a continued sequence of lines) by extending its input buffer if
2287        necessary. This feature was broken for very long pattern lines, leading to
2288        a string of junk being passed to pcre_compile() if the pattern was longer
2289        than about 50K.
2290    
2291    17. I have done a major re-factoring of the way pcre_compile() computes the
2292        amount of memory needed for a compiled pattern. Previously, there was code
2293        that made a preliminary scan of the pattern in order to do this. That was
2294        OK when PCRE was new, but as the facilities have expanded, it has become
2295        harder and harder to keep it in step with the real compile phase, and there
2296        have been a number of bugs (see for example, 4 above). I have now found a
2297        cunning way of running the real compile function in a "fake" mode that
2298        enables it to compute how much memory it would need, while actually only
2299        ever using a few hundred bytes of working memory and without too many
2300        tests of the mode. This should make future maintenance and development
2301        easier. A side effect of this work is that the limit of 200 on the nesting
2302        depth of parentheses has been removed (though this was never a serious
2303        limitation, I suspect). However, there is a downside: pcre_compile() now
2304        runs more slowly than before (30% or more, depending on the pattern). I
2305        hope this isn't a big issue. There is no effect on runtime performance.
2306    
2307    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2308        newline (only possible for the last line of a file) and it was a
2309        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2310    
2311    19. Added additional timing features to pcretest. (1) The -tm option now times
2312        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2313        separate command line item, by a number that specifies the number of
2314        repeats to use when timing. The default is 50000; this gives better
2315        precision, but takes uncomfortably long for very large patterns.
2316    
2317    20. Extended pcre_study() to be more clever in cases where a branch of a
2318        subpattern has no definite first character. For example, (a*|b*)[cd] would
2319        previously give no result from pcre_study(). Now it recognizes that the
2320        first character must be a, b, c, or d.
2321    
2322    21. There was an incorrect error "recursive call could loop indefinitely" if
2323        a subpattern (or the entire pattern) that was being tested for matching an
2324        empty string contained only one non-empty item after a nested subpattern.
2325        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2326        incorrectly, because the \d was being skipped in the check.
2327    
2328    22. The pcretest program now has a new pattern option /B and a command line
2329        option -b, which is equivalent to adding /B to every pattern. This causes
2330        it to show the compiled bytecode, without the additional information that
2331        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2332        is the same as /B/I).
2333    
2334    23. A new optimization is now able automatically to treat some sequences such
2335        as a*b as a*+b. More specifically, if something simple (such as a character
2336        or a simple class like \d) has an unlimited quantifier, and is followed by
2337        something that cannot possibly match the quantified thing, the quantifier
2338        is automatically "possessified".
2339    
2340    24. A recursive reference to a subpattern whose number was greater than 39
2341        went wrong under certain circumstances in UTF-8 mode. This bug could also
2342        have affected the operation of pcre_study().
2343    
2344    25. Realized that a little bit of performance could be had by replacing
2345        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2346    
2347    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2348    
2349    27. Possessive quantifiers such as a++ were previously implemented by turning
2350        them into atomic groups such as ($>a+). Now they have their own opcodes,
2351        which improves performance. This includes the automatically created ones
2352        from 23 above.
2353    
2354    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2355        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2356        the first matched character to be a colon. This applied both to named and
2357        numbered groups.
2358    
2359    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2360    
2361    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2362        building PCRE on Windows easier, so I have included it in the distribution.
2363    
2364    31. There is now a check in pcretest against a ridiculously large number being
2365        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2366        loop, the loop is abandoned.
2367    
2368    32. Forward references to subpatterns in conditions such as (?(2)...) where
2369        subpattern 2 is defined later cause pcre_compile() to search forwards in
2370        the pattern for the relevant set of parentheses. This search went wrong
2371        when there were unescaped parentheses in a character class, parentheses
2372        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2373    
2374    33. "Subroutine" calls and backreferences were previously restricted to
2375        referencing subpatterns earlier in the regex. This restriction has now
2376        been removed.
2377    
2378    34. Added a number of extra features that are going to be in Perl 5.10. On the
2379        whole, these are just syntactic alternatives for features that PCRE had
2380        previously implemented using the Python syntax or my own invention. The
2381        other formats are all retained for compatibility.
2382    
2383        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2384            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2385            also .NET compatible.
2386    
2387        (b) A recursion or subroutine call to a named group can now be defined as
2388            (?&name) as well as (?P>name).
2389    
2390        (c) A backreference to a named group can now be defined as \k<name> or
2391            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2392            5.10, are also .NET compatible.
2393    
2394        (d) A conditional reference to a named group can now use the syntax
2395            (?(<name>) or (?('name') as well as (?(name).
2396    
2397        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2398            groups (named and numbered) that are never evaluated inline, but can be
2399            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2400            is always false. There may be only one alternative in such a group.
2401    
2402        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2403            as the simple (?(R). The condition is true only if the most recent
2404            recursion is that of the given number or name. It does not search out
2405            through the entire recursion stack.
2406    
2407        (g) The escape \gN or \g{N} has been added, where N is a positive or
2408            negative number, specifying an absolute or relative reference.
2409    
2410    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2411        some "unreachable code" warnings.
2412    
2413    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2414        things, this adds five new scripts.
2415    
2416    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2417        There were also incompatibilities regarding the handling of \Q..\E inside
2418        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2419        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2420    
2421    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2422        matches an empty string, and forcibly breaks the loop. There were bugs in
2423        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2424        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2425        separate and independent bugs (that affected different cases) have been
2426        fixed.
2427    
2428    39. Refactored the code to abolish the use of different opcodes for small
2429        capturing bracket numbers. This is a tidy that I avoided doing when I
2430        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2431        The new approach is not only tidier, it makes it possible to reduce the
2432        memory needed to fix the previous bug (38).
2433    
2434    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2435        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2436        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2437        mode.
2438    
2439    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2440        report.
2441    
2442    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2443        copy construction and assignment in the C++ wrapper.
2444    
2445    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2446        couple of bugs that could have given wrong results in the "--newline=crlf"
2447        case.
2448    
2449    44. Added a number of casts and did some reorganization of signed/unsigned int
2450        variables following suggestions from Dair Grant. Also renamed the variable
2451        "this" as "item" because it is a C++ keyword.
2452    
2453    45. Arranged for dftables to add
2454    
2455          #include "pcre_internal.h"
2456    
2457        to pcre_chartables.c because without it, gcc 4.x may remove the array
2458        definition from the final binary if PCRE is built into a static library and
2459        dead code stripping is activated.
2460    
2461    46. For an unanchored pattern, if a match attempt fails at the start of a
2462        newline sequence, and the newline setting is CRLF or ANY, and the next two
2463        characters are CRLF, advance by two characters instead of one.
2464    
2465    
2466    Version 6.7 04-Jul-06
2467    ---------------------
2468    
2469     1. In order to handle tests when input lines are enormously long, pcretest has
2470        been re-factored so that it automatically extends its buffers when
2471        necessary. The code is crude, but this _is_ just a test program. The
2472        default size has been increased from 32K to 50K.
2473    
2474     2. The code in pcre_study() was using the value of the re argument before
2475        testing it for NULL. (Of course, in any sensible call of the function, it
2476        won't be NULL.)
2477    
2478     3. The memmove() emulation function in pcre_internal.h, which is used on
2479        systems that lack both memmove() and bcopy() - that is, hardly ever -
2480        was missing a "static" storage class specifier.
2481    
2482     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2483        containing an extended class (one that cannot be represented by a bitmap
2484        because it contains high-valued characters or Unicode property items, e.g.
2485        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2486        pattern, but PCRE should not loop if you do not (it no longer does).
2487        [Detail: two cases were found: (a) a repeated subpattern containing an
2488        extended class; (b) a recursive reference to a subpattern that followed a
2489        previous extended class. It wasn't skipping over the extended class
2490        correctly when UTF-8 mode was not set.]
2491    
2492     5. A negated single-character class was not being recognized as fixed-length
2493        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2494        compile error "lookbehind assertion is not fixed length".
2495    
2496     6. The RunPerlTest auxiliary script was showing an unexpected difference
2497        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2498        write a Perl script that can interpret lines of an input file either as
2499        byte characters or as UTF-8, which is what "perltest" was being required to
2500        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2501        can't do is switch easily at run time between having the "use utf8;" pragma
2502        or not. In the end, I fudged it by using the RunPerlTest script to insert
2503        "use utf8;" explicitly for the UTF-8 tests.
2504    
2505     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2506        the end of the subject string, contrary to the documentation and to what
2507        Perl does. This was true of both matching functions. Now it matches only at
2508        the start of the subject and immediately after *internal* newlines.
2509    
2510     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2511        a pointer to an int instead of a pointer to an unsigned long int. This
2512        caused problems on 64-bit systems.
2513    
2514     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2515        instance of the 'standard' template library not being so standard".
2516    
2517    10. There was no check on the number of named subpatterns nor the maximum
2518        length of a subpattern name. The product of these values is used to compute
2519        the size of the memory block for a compiled pattern. By supplying a very
2520        long subpattern name and a large number of named subpatterns, the size
2521        computation could be caused to overflow. This is now prevented by limiting
2522        the length of names to 32 characters, and the number of named subpatterns
2523        to 10,000.
2524    
2525    11. Subpatterns that are repeated with specific counts have to be replicated in
2526        the compiled pattern. The size of memory for this was computed from the
2527        length of the subpattern and the repeat count. The latter is limited to
2528        65535, but there was no limit on the former, meaning that integer overflow
2529        could in principle occur. The compiled length of a repeated subpattern is
2530        now limited to 30,000 bytes in order to prevent this.
2531    
2532    12. Added the optional facility to have named substrings with the same name.
2533    
2534    13. Added the ability to use a named substring as a condition, using the
2535        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2536        are numbers (not recommended). Forward references are permitted.
2537    
2538    14. Added forward references in named backreferences (if you see what I mean).
2539    
2540    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2541        pattern could run off the end of the subject. For example, the pattern
2542        "(?s)(.{1,5})"8 did this with the subject "ab".
2543    
2544    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2545        PCRE_CASELESS was set when matching characters that were quantified with ?
2546        or *.
2547    
2548    17. A character class other than a single negated character that had a minimum
2549        but no maximum quantifier - for example [ab]{6,} - was not handled
2550        correctly by pce_dfa_exec(). It would match only one character.
2551    
2552    18. A valid (though odd) pattern that looked like a POSIX character
2553        class but used an invalid character after [ (for example [[,abc,]]) caused
2554        pcre_compile() to give the error "Failed: internal error: code overflow" or
2555        in some cases to crash with a glibc free() error. This could even happen if
2556        the pattern terminated after [[ but there just happened to be a sequence of
2557        letters, a binary zero, and a closing ] in the memory that followed.
2558    
2559    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2560        over the years. Originally (before any Unicode support), just the bottom 8
2561        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2562        output from "man perlunicode" includes this:
2563    
2564          The regular expression compiler produces polymorphic opcodes.  That
2565          is, the pattern adapts to the data and automatically switches to
2566          the Unicode character scheme when presented with Unicode data--or
2567          instead uses a traditional byte scheme when presented with byte
2568          data.
2569    
2570        Sadly, a wide octal escape does not cause a switch, and in a string with
2571        no other multibyte characters, these octal escapes are treated as before.
2572        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2573        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2574        Unicode string.
2575    
2576        I have not perpetrated such confusion in PCRE. Up till now, it took just
2577        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2578        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2579        translate to the appropriate multibyte character.
2580    
2581    29. Applied some refactoring to reduce the number of warnings from Microsoft
2582        and Borland compilers. This has included removing the fudge introduced
2583        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2584        a warning about an unused variable.
2585    
2586    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2587        characters since release 4.0, because Perl (from release 5.004) does not.
2588        [Or at least, is documented not to: some releases seem to be in conflict
2589        with the documentation.] However, when a pattern was studied with
2590        pcre_study() and all its branches started with \s, PCRE still included VT
2591        as a possible starting character. Of course, this did no harm; it just
2592        caused an unnecessary match attempt.
2593    
2594    22. Removed a now-redundant internal flag bit that recorded the fact that case
2595        dependency changed within the pattern. This was once needed for "required
2596        byte" processing, but is no longer used. This recovers a now-scarce options
2597        bit. Also moved the least significant internal flag bit to the most-
2598        significant bit of the word, which was not previously used (hangover from
2599        the days when it was an int rather than a uint) to free up another bit for
2600        the future.
2601    
2602    23. Added support for CRLF line endings as well as CR and LF. As well as the
2603        default being selectable at build time, it can now be changed at runtime
2604        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2605        specify that it is scanning data with non-default line endings.
2606    
2607    24. Changed the definition of CXXLINK to make it agree with the definition of
2608        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2609    
2610    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2611        recursions. This makes a big different to stack usage for some patterns.
2612    
2613    26. If a subpattern containing a named recursion or subroutine reference such
2614        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2615        the space required for the compiled pattern went wrong and gave too small a
2616        value. Depending on the environment, this could lead to "Failed: internal
2617        error: code overflow at offset 49" or "glibc detected double free or
2618        corruption" errors.
2619    
2620    27. Applied patches from Google (a) to support the new newline modes and (b) to
2621        advance over multibyte UTF-8 characters in GlobalReplace.
2622    
2623    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2624        difference for some implementation of PCRE in some Windows version.
2625    
2626    29. Added some extra testing facilities to pcretest:
2627    
2628        \q<number>   in a data line sets the "match limit" value
2629        \Q<number>   in a data line sets the "match recursion limt" value
2630        -S <number>  sets the stack size, where <number> is in megabytes
2631    
2632        The -S option isn't available for Windows.
2633    
2634    
2635    Version 6.6 06-Feb-06
2636    ---------------------
2637    
2638     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2639        in pcreposix.h. I have copied the definition from pcre.h.
2640    
2641     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2642        because pcre.h is no longer a built file.
2643    
2644     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2645        not normally included in the compiled code.
2646    
2647    
2648    Version 6.5 01-Feb-06
2649    ---------------------
2650    
2651     1. When using the partial match feature with pcre_dfa_exec(), it was not
2652        anchoring the second and subsequent partial matches at the new starting
2653        point. This could lead to incorrect results. For example, with the pattern
2654        /1234/, partially matching against "123" and then "a4" gave a match.
2655    
2656     2. Changes to pcregrep:
2657    
2658        (a) All non-match returns from pcre_exec() were being treated as failures
2659            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2660            error message is output. Some extra information is given for the
2661            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2662            probably the only errors that are likely to be caused by users (by
2663            specifying a regex that has nested indefinite repeats, for instance).
2664            If there are more than 20 of these errors, pcregrep is abandoned.
2665    
2666        (b) A binary zero was treated as data while matching, but terminated the
2667            output line if it was written out. This has been fixed: binary zeroes
2668            are now no different to any other data bytes.
2669    
2670        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2671            used to set a locale for matching. The --locale=xxxx long option has
2672            been added (no short equivalent) to specify a locale explicitly on the
2673            pcregrep command, overriding the environment variables.
2674    
2675        (d) When -B was used with -n, some line numbers in the output were one less
2676            than they should have been.
2677    
2678        (e) Added the -o (--only-matching) option.
2679    
2680        (f) If -A or -C was used with -c (count only), some lines of context were
2681            accidentally printed for the final match.
2682    
2683        (g) Added the -H (--with-filename) option.
2684    
2685        (h) The combination of options -rh failed to suppress file names for files
2686            that were found from directory arguments.
2687    
2688        (i) Added the -D (--devices) and -d (--directories) options.
2689    
2690        (j) Added the -F (--fixed-strings) option.
2691    
2692        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2693    
2694        (l) Added the --colo(u)r option.
2695    
2696        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2697            is not present by default.
2698    
2699     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2700        items such as (?R) or (?1), when the recursion could match a number of
2701        alternatives. If it matched one of the alternatives, but subsequently,
2702        outside the recursion, there was a failure, the code tried to back up into
2703        the recursion. However, because of the way PCRE is implemented, this is not
2704        possible, and the result was an incorrect result from the match.
2705    
2706        In order to prevent this happening, the specification of recursion has
2707        been changed so that all such subpatterns are automatically treated as
2708        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2709    
2710     4. I had overlooked the fact that, in some locales, there are characters for
2711        which isalpha() is true but neither isupper() nor islower() are true. In
2712        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2713        and ordfeminine) are like this. This affected the treatment of \w and \W
2714        when they appeared in character classes, but not when they appeared outside
2715        a character class. The bit map for "word" characters is now created
2716        separately from the results of isalnum() instead of just taking it from the
2717        upper, lower, and digit maps. (Plus the underscore character, of course.)
2718    
2719     5. The above bug also affected the handling of POSIX character classes such as
2720        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2721        permanent tables. Instead, the bit maps for such a class were previously
2722        created as the appropriate unions of the upper, lower, and digit bitmaps.
2723        Now they are created by subtraction from the [[:word:]] class, which has
2724        its own bitmap.
2725    
2726     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2727        It is created by subtracting the vertical space characters (\x09, \x0a,
2728        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2729        subtraction was done in the overall bitmap for a character class, meaning
2730        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2731        be recognized. This bug has been fixed.
2732    
2733     7. Patches from the folks at Google:
2734    
2735          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2736          real life, but is still worth protecting against".
2737    
2738          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2739          regular expressions".
2740    
2741          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2742          have it.
2743    
2744          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2745          "configure" and the latter not, in order to fix a problem somebody had
2746          with compiling the Arg class on HP-UX.
2747    
2748          (e) Improve the error-handling of the C++ wrapper a little bit.
2749    
2750          (f) New tests for checking recursion limiting.
2751    
2752     8. The pcre_memmove() function, which is used only if the environment does not
2753        have a standard memmove() function (and is therefore rarely compiled),
2754        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2755        returning a result (though PCRE never actually uses the result).
2756    
2757     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2758        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2759        returned instead of calling malloc() with an overflowing number that would
2760        most likely cause subsequent chaos.
2761    
2762    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2763    
2764    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2765        with this option is matched, the nmatch and pmatch options of regexec() are
2766        ignored.
2767    
2768    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2769        provided in case anyone wants to the the POSIX interface with UTF-8
2770        strings.
2771    
2772    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2773        C++ linking (needed for some HP-UX environments).
2774    
2775    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2776        (unused parameter) and in the pcre_printint() function (omitted "default"
2777        switch label when the default is to do nothing).
2778    
2779    15. Added some code to make it possible, when PCRE is compiled as a C++
2780        library, to replace subject pointers for pcre_exec() with a smart pointer
2781        class, thus making it possible to process discontinuous strings.
2782    
2783    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2784        much the same function. They were added by different people who were trying
2785        to make PCRE easy to compile on non-Unix systems. It has been suggested
2786        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2787        for compiling on Windows systems. I have therefore replaced it with
2788        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2789        defaults to "extern" for C or "extern C" for C++, which works fine on
2790        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2791        SCOPE with something explicit in config.h. In addition:
2792    
2793        (a) pcreposix.h still had just "extern" instead of either of these macros;
2794            I have replaced it with PCRE_DATA_SCOPE.
2795    
2796        (b) Functions such as _pcre_xclass(), which are internal to the library,
2797            but external in the C sense, all had PCRE_EXPORT in their definitions.
2798            This is apparently wrong for the Windows case, so I have removed it.
2799            (It makes no difference on Unix-like systems.)
2800    
2801    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2802        of recursive calls to match(). This is different to MATCH_LIMIT because
2803        that limits the total number of calls to match(), not all of which increase
2804        the depth of recursion. Limiting the recursion depth limits the amount of
2805        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2806        when PCRE is compiled, and changed at run time. A patch from Google adds
2807        this functionality to the C++ interface.
2808    
2809    18. Changes to the handling of Unicode character properties:
2810    
2811        (a) Updated the table to Unicode 4.1.0.
2812    
2813        (b) Recognize characters that are not in the table as "Cn" (undefined).
2814    
2815        (c) I revised the way the table is implemented to a much improved format
2816            which includes recognition of ranges. It now supports the ranges that
2817            are defined in UnicodeData.txt, and it also amalgamates other
2818            characters into ranges. This has reduced the number of entries in the
2819            table from around 16,000 to around 3,000, thus reducing its size
2820            considerably. I realized I did not need to use a tree structure after
2821            all - a binary chop search is just as efficient. Having reduced the
2822            number of entries, I extended their size from 6 bytes to 8 bytes to
2823            allow for more data.
2824    
2825        (d) Added support for Unicode script names via properties such as \p{Han}.
2826    
2827    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2828        matching that character.
2829    
2830    20. When matching a repeated Unicode property with a minimum greater than zero,
2831        (for example \pL{2,}), PCRE could look past the end of the subject if it
2832        reached it while seeking the minimum number of characters. This could
2833        happen only if some of the characters were more than one byte long, because
2834        there is a check for at least the minimum number of bytes.
2835    
2836    21. Refactored the implementation of \p and \P so as to be more general, to
2837        allow for more different types of property in future. This has changed the
2838        compiled form incompatibly. Anybody with saved compiled patterns that use
2839        \p or \P will have to recompile them.
2840    
2841    22. Added "Any" and "L&" to the supported property types.
2842    
2843    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2844        but give a compile time error if the value is greater than 0xff.
2845    
2846    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2847        accidentally not being installed or uninstalled.
2848    
2849    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2850        made were to insert the current release number. This seemed silly, because
2851        it made things harder for people building PCRE on systems that don't run
2852        "configure". I have turned pcre.h into a distributed file, no longer built
2853        by "configure", with the version identification directly included. There is
2854        no longer a pcre.h.in file.
2855    
2856        However, this change necessitated a change to the pcre-config script as
2857        well. It is built from pcre-config.in, and one of the substitutions was the
2858        release number. I have updated configure.ac so that ./configure now finds
2859        the release number by grepping pcre.h.
2860    
2861    26. Added the ability to run the tests under valgrind.
2862    
2863    
2864    Version 6.4 05-Sep-05
2865    ---------------------
2866    
2867     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2868        "--" to be printed when multiple files were scanned, even when none of the
2869        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2870        consider it to be a bug, and have restored the previous behaviour.
2871    
2872     2. A couple of code tidies to get rid of compiler warnings.
2873    
2874     3. The pcretest program used to cheat by referring to symbols in the library
2875        whose names begin with _pcre_. These are internal symbols that are not
2876        really supposed to be visible externally, and in some environments it is
2877        possible to suppress them. The cheating is now confined to including
2878        certain files from the library's source, which is a bit cleaner.
2879    
2880     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2881        file's purpose clearer.
2882    
2883     5. Reorganized pcre_ucp_findchar().
2884    
2885    
2886    Version 6.3 15-Aug-05
2887    ---------------------
2888    
2889     1. The file libpcre.pc.in did not have general read permission in the tarball.
2890    
2891     2. There were some problems when building without C++ support:
2892    
2893        (a) If C++ support was not built, "make install" and "make test" still
2894            tried to test it.
2895    
2896        (b) There were problems when the value of CXX was explicitly set. Some
2897            changes have been made to try to fix these, and ...
2898    
2899        (c) --disable-cpp can now be used to explicitly disable C++ support.
2900    
2901        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2902            backslash in a target when C++ was disabled. This confuses some
2903            versions of "make", apparently. Using an intermediate variable solves
2904            this. (Same for CPP_LOBJ.)
2905    
2906     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2907        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2908        necessary on certain architectures.
2909    
2910     4. Added a setting of -export-symbols-regex to the link command to remove
2911        those symbols that are exported in the C sense, but actually are local
2912        within the library, and not documented. Their names all begin with
2913        "_pcre_". This is not a perfect job, because (a) we have to except some
2914        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2915        available (and never for static libraries). I have made a note to try to
2916        find a way round (a) in the future.
2917    
2918    
2919    Version 6.2 01-Aug-05
2920    ---------------------
2921    
2922     1. There was no test for integer overflow of quantifier values. A construction
2923        such as {1111111111111111} would give undefined results. What is worse, if
2924        a minimum quantifier for a parenthesized subpattern overflowed and became
2925        negative, the calculation of the memory size went wrong. This could have
2926        led to memory overwriting.
2927    
2928     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2929    
2930     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2931        operating environments where this matters.
2932    
2933     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2934        PCRE options from within the C++ wrapper.
2935    
2936     5. Named capturing subpatterns were not being correctly counted when a pattern
2937        was compiled. This caused two problems: (a) If there were more than 100
2938        such subpatterns, the calculation of the memory needed for the whole
2939        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2940        back references of the form \12, where the number was greater than 9, were
2941        not recognized as back references, even though there were sufficient
2942        previous subpatterns.
2943    
2944     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2945        versions of gcc, e.g. 2.95.4.
2946    
2947    
2948    Version 6.1 21-Jun-05
2949    ---------------------
2950    
2951     1. There was one reference to the variable "posix" in pcretest.c that was not
2952        surrounded by "#if !defined NOPOSIX".
2953    
2954     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2955        the cross-check on the old pcre_info() function, for the benefit of the
2956        cut-down version of PCRE that is currently imported into Exim.
2957    
2958     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2959        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2960        patterns that start (?i) but I don't think that matters. The use of (?i) is
2961        just an example; this all applies to the other options as well.
2962    
2963     4. Since libtool seems to echo the compile commands it is issuing, the output
2964        from "make" can be reduced a bit by putting "@" in front of each libtool
2965        compile command.
2966    
2967     5. Patch from the folks at Google for configure.in to be a bit more thorough
2968        in checking for a suitable C++ installation before trying to compile the
2969        C++ stuff. This should fix a reported problem when a compiler was present,
2970        but no suitable headers.
2971    
2972     6. The man pages all had just "PCRE" as their title. I have changed them to
2973        be the relevant file name. I have also arranged that these names are
2974        retained in the file doc/pcre.txt, which is a concatenation in text format
2975        of all the man pages except the little individual ones for each function.
2976    
2977     7. The NON-UNIX-USE file had not been updated for the different set of source
2978        files that come with release 6. I also added a few comments about the C++
2979        wrapper.
2980    
2981    
2982    Version 6.0 07-Jun-05
2983    ---------------------
2984    
2985     1. Some minor internal re-organization to help with my DFA experiments.
2986    
2987     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2988        didn't matter for the library itself when fully configured, but did matter
2989        when compiling without UCP support, or within Exim, where the ucp files are
2990        not imported.
2991    
2992     3. Refactoring of the library code to split up the various functions into
2993        different source modules. The addition of the new DFA matching code (see
2994        below) to a single monolithic source would have made it really too
2995        unwieldy, quite apart from causing all the code to be include in a
2996        statically linked application, when only some functions are used. This is
2997        relevant even without the DFA addition now that patterns can be compiled in
2998        one application and matched in another.
2999    
3000        The downside of splitting up is that there have to be some external
3001        functions and data tables that are used internally in different modules of
3002        the library but which are not part of the API. These have all had their
3003        names changed to start with "_pcre_" so that they are unlikely to clash
3004        with other external names.
3005    
3006     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3007        a different (DFA) algorithm. Although it is slower than the original
3008        function, it does have some advantages for certain types of matching
3009        problem.
3010    
3011     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3012        including restarting after a partial match.
3013    
3014     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3015        defined when compiling for Windows was sent to me. I have put it into the
3016        code, though I have no means of testing or verifying it.
3017    
3018     7. Added the pcre_refcount() auxiliary function.
3019    
3020     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3021        match before or at the first newline in the subject string. In pcretest,
3022        the /f option on a pattern can be used to set this.
3023    
3024     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3025        would behave wrongly. This has been present in PCRE since release 4.0.
3026    
3027    10. A number of changes to the pcregrep command:
3028    
3029        (a) Refactored how -x works; insert ^(...)$ instead of setting
3030            PCRE_ANCHORED and checking the length, in preparation for adding
3031            something similar for -w.
3032    
3033        (b) Added the -w (match as a word) option.
3034    
3035        (c) Refactored the way lines are read and buffered so as to have more
3036            than one at a time available.
3037    
3038        (d) Implemented a pcregrep test script.
3039    
3040        (e) Added the -M (multiline match) option. This allows patterns to match
3041            over several lines of the subject. The buffering ensures that at least
3042            8K, or the rest of the document (whichever is the shorter) is available
3043            for matching (and similarly the previous 8K for lookbehind assertions).
3044    
3045        (f) Changed the --help output so that it now says
3046    
3047              -w, --word-regex(p)
3048    
3049            instead of two lines, one with "regex" and the other with "regexp"
3050            because that confused at least one person since the short forms are the
3051            same. (This required a bit of code, as the output is generated
3052            automatically from a table. It wasn't just a text change.)
3053    
3054        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3055            option but starts with a hyphen. Could be a pattern or a path name
3056            starting with a hyphen, for instance.
3057    
3058        (h) "-" can be given as a file name to represent stdin.
3059    
3060        (i) When file names are being printed, "(standard input)" is used for
3061            the standard input, for compatibility with GNU grep. Previously
3062            "<stdin>" was used.
3063    
3064        (j) The option --label=xxx can be used to supply a name to be used for
3065            stdin when file names are being printed. There is no short form.
3066    
3067        (k) Re-factored the options decoding logic because we are going to add
3068            two more options that take data. Such options can now be given in four
3069            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3070    
3071        (l) Added the -A, -B, and -C options for requesting that lines of context
3072            around matches be printed.
3073    
3074        (m) Added the -L option to print the names of files that do not contain
3075            any matching lines, that is, the complement of -l.
3076    
3077        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3078            continue to scan other files.
3079    
3080        (o) The -s option was incorrectly implemented. For compatibility with other
3081            greps, it now suppresses the error message for a non-existent or non-
3082            accessible file (but not the return code). There is a new option called
3083            -q that suppresses the output of matching lines, which was what -s was
3084            previously doing.
3085    
3086        (p) Added --include and --exclude options to specify files for inclusion
3087            and exclusion when recursing.
3088    
3089    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3090        Hopefully, it now does.
3091    
3092    12. Missing cast in pcre_study().
3093    
3094    13. Added an "uninstall" target to the makefile.
3095    
3096    14. Replaced "extern" in the function prototypes in Makefile.in with
3097        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3098        world, but is set differently for Windows.
3099    
3100    15. Added a second compiling function called pcre_compile2(). The only
3101        difference is that it has an extra argument, which is a pointer to an
3102        integer error code. When there is a compile-time failure, this is set
3103        non-zero, in addition to the error test pointer being set to point to an
3104        error message. The new argument may be NULL if no error number is required
3105        (but then you may as well call pcre_compile(), which is now just a
3106        wrapper). This facility is provided because some applications need a
3107        numeric error indication, but it has also enabled me to tidy up the way
3108        compile-time errors are handled in the POSIX wrapper.
3109    
3110    16. Added VPATH=.libs to the makefile; this should help when building with one
3111        prefix path and installing with another. (Or so I'm told by someone who
3112        knows more about this stuff than I do.)
3113    
3114    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3115        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3116        match everything, including newlines. This is not POSIX-compatible, but
3117        somebody wanted the feature. From pcretest it can be activated by using
3118        both the P and the s flags.
3119    
3120    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3121    
3122    19. libpcre.pc was being incorrectly installed as executable.
3123    
3124    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3125        it now also looks for '\r' so that it will work unmodified on Windows.
3126    
3127    21. Added Google's contributed C++ wrapper to the distribution.
3128    
3129    22. Added some untidy missing memory free() calls in pcretest, to keep
3130        Electric Fence happy when testing.
3131    
3132    
3133    
3134    Version 5.0 13-Sep-04
3135    ---------------------
3136    
3137     1. Internal change: literal characters are no longer packed up into items
3138        containing multiple characters in a single byte-string. Each character
3139        is now matched using a separate opcode. However, there may be more than one
3140        byte in the character in UTF-8 mode.
3141    
3142     2. The pcre_callout_block structure has two new fields: pattern_position and
3143        next_item_length. These contain the offset in the pattern to the next match
3144        item, and its length, respectively.
3145    
3146     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3147        insertion of callouts before each pattern item. Added the /C option to
3148        pcretest to make use of this.
3149    
3150     4. On the advice of a Windows user, the lines
3151    
3152          #if defined(_WIN32) || defined(WIN32)
3153          _setmode( _fileno( stdout ), 0x8000 );
3154          #endif  /* defined(_WIN32) || defined(WIN32) */
3155    
3156        have been added to the source of pcretest. This apparently does useful
3157        magic in relation to line terminators.
3158    
3159     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3160        for the benefit of those environments where the "b" makes a difference.
3161    
3162     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3163        to know about it. I have put a hack into configure.in that adds in code
3164        to set GCC=yes if CC=icc. This seems to end up at a point in the
3165        generated configure script that is early enough to affect the setting of
3166        compiler options, which is what is needed, but I have no means of testing
3167        whether it really works. (The user who reported this had patched the
3168        generated configure script, which of course I cannot do.)
3169    
3170        LATER: After change 22 below (new libtool files), the configure script
3171        seems to know about icc (and also ecc). Therefore, I have commented out
3172        this hack in configure.in.
3173    
3174     7. Added support for pkg-config (2 patches were sent in).
3175    
3176     8. Negated POSIX character classes that used a combination of internal tables
3177        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3178        [[:^ascii]]. Typically, they would match almost any characters. The other
3179        POSIX classes were not broken in this way.
3180    
3181     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3182        to find the match, as PCRE was deluded into thinking that the match had to
3183        start at the start point or following a newline. The same bug applied to
3184        patterns with negative forward assertions or any backward assertions
3185        preceding ".*" at the start, unless the pattern required a fixed first
3186        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3187    
3188    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3189        starting at the last subject character, bytes beyond the end of the subject
3190        string were read.
3191    
3192    11. Renamed the variable "class" as "classbits" to make life easier for C++
3193        users. (Previously there was a macro definition, but it apparently wasn't
3194        enough.)
3195    
3196    12. Added the new field "tables" to the extra data so that tables can be passed
3197        in at exec time, or the internal tables can be re-selected. This allows
3198        a compiled regex to be saved and re-used at a later time by a different
3199        program that might have everything at different addresses.
3200    
3201    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3202        -R library as well as a -L library.
3203    
3204    14. The debugging options of pcretest (-d on the command line or D on a
3205        pattern) showed incorrect output for anything following an extended class
3206        that contained multibyte characters and which was followed by a quantifier.
3207    
3208    15. Added optional support for general category Unicode character properties
3209        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3210        support. It adds about 90K to the size of the library. The meanings of the
3211        inbuilt class escapes such as \d and \s have NOT been changed.
3212    
3213    16. Updated pcredemo.c to include calls to free() to release the memory for the
3214        compiled pattern.
3215    
3216    17. The generated file chartables.c was being created in the source directory
3217        instead of in the building directory. This caused the build to fail if the
3218        source directory was different from the building directory, and was
3219        read-only.
3220    
3221    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3222        file. No doubt somebody will tell me if they don't make sense... Also added
3223        Dan Mooney's comments about building on OpenVMS.
3224    
3225    19. Added support for partial matching via the PCRE_PARTIAL option for
3226        pcre_exec() and the \P data escape in pcretest.
3227    
3228    20. Extended pcretest with 3 new pattern features:
3229    
3230        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3231              write the compiled pattern to the file whose name is "rest-of-line".
3232              This is a straight binary dump of the data, with the saved pointer to
3233              the character tables forced to be NULL. The study data, if any, is
3234              written too. After writing, pcretest reads a new pattern.
3235    
3236        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3237              compiled pattern from the given file. There must not be any
3238              occurrences of "<" in the file name (pretty unlikely); if there are,
3239              pcretest will instead treat the initial "<" as a pattern delimiter.
3240              After reading in the pattern, pcretest goes on to read data lines as
3241              usual.
3242    
3243        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3244              and 16-bit fields in a compiled pattern, to simulate a pattern that
3245              was compiled on a host of opposite endianness.
3246    
3247    21. The pcre-exec() function can now cope with patterns that were compiled on
3248        hosts of opposite endianness, with this restriction:
3249    
3250          As for any compiled expression that is saved and used later, the tables
3251          pointer field cannot be preserved; the extra_data field in the arguments
3252          to pcre_exec() should be used to pass in a tables address if a value
3253          other than the default internal tables were used at compile time.
3254    
3255    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3256        now diagnosed as an error. Previously, most of the time, a negative number
3257        would have been treated as zero, but if in addition "ovector" was passed as
3258        NULL, a crash could occur.
3259    
3260    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3261        new versions from the libtool 1.5 distribution (the last one is a copy of
3262        a file called libtool.m4). This seems to have fixed the need to patch
3263        "configure" to support Darwin 1.3 (which I used to do). However, I still
3264        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3265        workstation).
3266    
3267    24. Changed the PCRE licence to be the more standard "BSD" licence.
3268    
3269    
3270    Version 4.5 01-Dec-03
3271    ---------------------
3272    
3273     1. There has been some re-arrangement of the code for the match() function so
3274        that it can be compiled in a version that does not call itself recursively.
3275        Instead, it keeps those local variables that need separate instances for
3276        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3277        needs to "recurse". Keeping track of where control must go is done by means
3278        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3279        hide most of the details from the main code, and operates only if
3280        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3281        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3282        operating.
3283    
3284        To make it easier for callers to provide specially tailored get/free
3285        functions for this usage, two new functions, pcre_stack_malloc, and
3286        pcre_stack_free, are used. They are always called in strict stacking order,
3287        and the size of block requested is always the same.
3288    
3289        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3290        PCRE has been compiled to use the stack or the heap for recursion. The
3291        -C option of pcretest uses this to show which version is compiled.
3292    
3293        A new data escape \S, is added to pcretest; it causes the amounts of store
3294        obtained and freed by both kinds of malloc/free at match time to be added
3295        to the output.
3296    
3297     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3298        what's available on my current Linux desktop machine.
3299    
3300     3. When matching a UTF-8 string, the test for a valid string at the start has
3301        been extended. If start_offset is not zero, PCRE now checks that it points
3302        to a byte that is the start of a UTF-8 character. If not, it returns
3303        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3304        this is necessary because there may be backward assertions in the pattern.
3305        When matching the same subject several times, it may save resources to use
3306        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3307    
3308     4. The code for checking the validity of UTF-8 strings has been tightened so
3309        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3310        containing "overlong sequences".
3311    
3312     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3313        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3314        should have been "&", but it just so happened that all the cases this let
3315        through by mistake were picked up later in the function.
3316    
3317     6. I had used a variable called "isblank" - this is a C99 function, causing
3318        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3319    
3320     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3321        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3322        so the version is shown; (c) stop "make test" echoing "./RunTest".
3323    
3324     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3325    
3326     9. The prototype for memmove() for systems that don't have it was using
3327        size_t, but the inclusion of the header that defines size_t was later. I've
3328        moved the #includes for the C headers earlier to avoid this.
3329    
3330    10. Added some adjustments to the code to make it easier to compiler on certain
3331        special systems:
3332    
3333          (a) Some "const" qualifiers were missing.
3334          (b) Added the macro EXPORT before all exported functions; by default this
3335              is defined to be empty.
3336          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3337              that it reads its output file name as an argument instead of writing
3338              to the standard output and assuming this can be redirected.
3339    
3340    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3341        class containing characters with values greater than 255, PCRE compilation
3342        went into a loop.
3343    
3344    12. A recursive reference to a subpattern that was within another subpattern
3345        that had a minimum quantifier of zero caused PCRE to crash. For example,
3346        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3347        recursion. If the recursively-called subpattern itself had a zero repeat,
3348        that was OK.
3349    
3350    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3351        buffer into which it was copied (for escape processing) was still set at
3352        1024, so long lines caused crashes.
3353    
3354    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3355        "internal error: code overflow...". This applied to any character class
3356        that was followed by a possessive quantifier.
3357    
3358    15. Modified the Makefile to add libpcre.la as a prerequisite for
3359        libpcreposix.la because I was told this is needed for a parallel build to
3360        work.
3361    
3362    16. If a pattern that contained .* following optional items at the start was
3363        studied, the wrong optimizing data was generated, leading to matching
3364        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3365        matching string must start with a or b or c. The correct conclusion for
3366        this pattern is that a match can start with any character.
3367    
3368    
3369    Version 4.4 13-Aug-03
3370    ---------------------
3371    
3372     1. In UTF-8 mode, a character class containing characters with values between
3373        127 and 255 was not handled correctly if the compiled pattern was studied.
3374        In fixing this, I have also improved the studying algorithm for such
3375        classes (slightly).
3376    
3377     2. Three internal functions had redundant arguments passed to them. Removal
3378        might give a very teeny performance improvement.
3379    
3380     3. Documentation bug: the value of the capture_top field in a callout is *one
3381        more than* the number of the hightest numbered captured substring.
3382    
3383     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3384        in incorrectly linking with a previously installed version. They now link
3385        explicitly with libpcre.la.
3386    
3387     5. configure.in no longer needs to recognize Cygwin specially.
3388    
3389     6. A problem in pcre.in for Windows platforms is fixed.
3390    
3391     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3392        pcretest, it used to include the size of the study block as part of its
3393        output. Unfortunately, the structure contains a field that has a different
3394        size on different hardware architectures. This meant that the tests that
3395        showed this size failed. As the block is currently always of a fixed size,
3396        this information isn't actually particularly useful in pcretest output, so
3397        I have just removed it.
3398    
3399     8. Three pre-processor statements accidentally did not start in column 1.
3400        Sadly, there are *still* compilers around that complain, even though
3401        standard C has not required this for well over a decade. Sigh.
3402    
3403     9. In pcretest, the code for checking callouts passed small integers in the
3404        callout_data field, which is a void * field. However, some picky compilers
3405        complained about the casts involved for this on 64-bit systems. Now
3406        pcretest passes the address of the small integer instead, which should get
3407        rid of the warnings.
3408    
3409    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3410        both compile and run time, and gives an error if an invalid UTF-8 sequence
3411        is found. There is a option for disabling this check in cases where the
3412        string is known to be correct and/or the maximum performance is wanted.
3413    
3414    11. In response to a bug report, I changed one line in Makefile.in from
3415    
3416            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3417        to
3418            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3419    
3420        to look similar to other lines, but I have no way of telling whether this
3421        is the right thing to do, as I do not use Windows. No doubt I'll get told
3422        if it's wrong...
3423    
3424    
3425    Version 4.3 21-May-03
3426    ---------------------
3427    
3428    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3429       Makefile.
3430    
3431    2. Some refactoring to improve the quality of the code:
3432    
3433       (i)   The utf8_table... variables are now declared "const".
3434    
3435       (ii)  The code for \cx, which used the "case flipping" table to upper case
3436             lower case letters, now just substracts 32. This is ASCII-specific,
3437             but the whole concept of \cx is ASCII-specific, so it seems
3438             reasonable.
3439    
3440       (iii) PCRE was using its character types table to recognize decimal and
3441             hexadecimal digits in the pattern. This is silly, because it handles
3442             only 0-9, a-f, and A-F, but the character types table is locale-
3443             specific, which means strange things might happen. A private
3444             table is now used for this - though it costs 256 bytes, a table is
3445             much faster than multiple explicit tests. Of course, the standard
3446             character types table is still used for matching digits in subject
3447             strings against \d.
3448    
3449       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3450             ending in _t are). So I've renamed it as ESC_tee.
3451    
3452    3. The first argument for regexec() in the POSIX wrapper should have been
3453       defined as "const".
3454    
3455    4. Changed pcretest to use malloc() for its buffers so that they can be
3456       Electric Fenced for debugging.
3457    
3458    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3459       to read one or more bytes before the start of the subject string. Often this
3460       had no effect on PCRE's behaviour, but in some circumstances it could
3461       provoke a segmentation fault.
3462    
3463    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3464       to try to read one or more bytes before the start of the subject string.
3465    
3466    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3467       UTF-8 support could misbehave in various ways if the subject string
3468       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3469       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3470       back over UTF-8 characters.)
3471    
3472    
3473    Version 4.2 14-Apr-03
3474    ---------------------
3475    
3476    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3477    
3478    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3479         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3480         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3481         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3482         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3483           and BUILD_EXEEXT
3484         Note: automatic setting of the BUILD variables is not yet working
3485         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3486           compile-time but not at link-time
3487         [LINK]: use for linking executables only
3488         make different versions for Windows and non-Windows
3489         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3490           libraries
3491         [LINK_FOR_BUILD]: new variable
3492         [OBJEXT]: use throughout
3493         [EXEEXT]: use throughout
3494         <winshared>: new target
3495         <wininstall>: new target
3496         <dftables.o>: use native compiler
3497         <dftables>: use native linker
3498         <install>: handle Windows platform correctly
3499         <clean>: ditto
3500         <check>: ditto
3501         copy DLL to top builddir before testing
3502    
3503       As part of these changes, -no-undefined was removed again. This was reported
3504       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3505       in any case.
3506    
3507    3. Some tidies to get rid of compiler warnings:
3508    
3509       . In the match_data structure, match_limit was an unsigned long int, whereas
3510         match_call_count was an int. I've made them both unsigned long ints.
3511    
3512       . In pcretest the fact that a const uschar * doesn't automatically cast to
3513         a void * provoked a warning.
3514    
3515       . Turning on some more compiler warnings threw up some "shadow" variables
3516         and a few more missing casts.
3517    
3518    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3519       option, a class that contained a single character with a value between 128
3520       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3521    
3522    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3523       option, a class that contained several characters, but with at least one
3524       whose value was between 128 and 255 caused PCRE to crash.
3525    
3526    
3527    Version 4.1 12-Mar-03
3528    ---------------------
3529    
3530    1. Compiling with gcc -pedantic found a couple of places where casts were
3531    needed, and a string in dftables.c that was longer than standard compilers are
3532    required to support.
3533    
3534    2. Compiling with Sun's compiler found a few more places where the code could
3535    be tidied up in order to avoid warnings.
3536    
3537    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3538    first of these names is deprecated in the latest Autoconf in favour of the name
3539    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3540    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3541    analogy I have changed it to CFLAGS_FOR_BUILD.
3542    
3543    4. Added -no-undefined to the linking command in the Makefile, because this is
3544    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3545    linking step for the pcreposix library.
3546    
3547    5. PCRE was failing to diagnose the case of two named groups with the same
3548    name.
3549    
3550    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3551    literal character that is needed in the subject for a match, and scans along to
3552    ensure that it is present before embarking on the full matching process. This
3553    saves time in cases of nested unlimited repeats that are never going to match.
3554    Problem: the scan can take a lot of time if the subject is very long (e.g.
3555    megabytes), thus penalizing straightforward matches. It is now done only if the
3556    amount of subject to be scanned is less than 1000 bytes.
3557    
3558    7. A lesser problem with the same optimization is that it was recording the
3559    first character of an anchored pattern as "needed", thus provoking a search
3560    right along the subject, even when the first match of the pattern was going to
3561    fail. The "needed" character is now not set for anchored patterns, unless it
3562    follows something in the pattern that is of non-fixed length. Thus, it still
3563    fulfils its original purpose of finding quick non-matches in cases of nested
3564    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3565    
3566    
3567    Version 4.0 17-Feb-03
3568    ---------------------
3569    
3570    1. If a comment in an extended regex that started immediately after a meta-item
3571    extended to the end of string, PCRE compiled incorrect data. This could lead to
3572    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3573    
3574    2. Moved to autoconf 2.53 and libtool 1.4.2.
3575    
3576    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3577    the special perltest8 script is no longer needed - all the tests can be run
3578    from a single perltest script.
3579    
3580    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3581    by \s. It has now been removed in PCRE. This means it isn't recognized as
3582    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3583    class [:space:] *does* include VT, thereby creating a mess.
3584    
3585    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3586    space and tab.
3587    
3588    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3589    its new features into the main test script, reducing the number of scripts.
3590    
3591    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3592    were backward compatible, and made the (?i) apply to the whole pattern, as if
3593    /i were given. Now it behaves more logically, and applies the option setting
3594    only to what follows. PCRE has been changed to follow suit. However, if it
3595    finds options settings right at the start of the pattern, it extracts them into
3596    the global options, as before. Thus, they show up in the info data.
3597    
3598    8. Added support for the \Q...\E escape sequence. Characters in between are
3599    treated as literals. This is slightly different from Perl in that $ and @ are
3600    also handled as literals inside the quotes. In Perl, they will cause variable
3601    interpolation. Note the following examples:
3602    
3603        Pattern            PCRE matches      Perl matches
3604    
3605        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3606        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3607        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3608    
3609    For compatibility with Perl, \Q...\E sequences are recognized inside character
3610    classes as well as outside them.
3611    
3612    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3613    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3614    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3615    signed/unsigned warnings.
3616    
3617    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3618    option for pcretest, so I've replaced it by a simple function that does just
3619    that job.
3620    
3621    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3622    "pcregrep -".
3623    
3624    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3625    Java package. This provides some syntactic sugar for simple cases of what my
3626    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3627    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3628    item, you can use this simplified notation. Note that only makes sense with
3629    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3630    greediness, whatever the setting of the PCRE_UNGREEDY option.
3631    
3632    13. A change of greediness default within a pattern was not taking effect at
3633    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3634    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3635    was abstracted outside.
3636    
3637    14. PCRE now supports the \G assertion. It is true when the current matching
3638    position is at the start point of the match. This differs from \A when the
3639    starting offset is non-zero. Used with the /g option of pcretest (or similar
3640    code), it works in the same way as it does for Perl's /g option. If all
3641    alternatives of a regex begin with \G, the expression is anchored to the start
3642    match position, and the "anchored" flag is set in the compiled expression.
3643    
3644    15. Some bugs concerning the handling of certain option changes within patterns
3645    have been fixed. These applied to options other than (?ims). For example,
3646    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3647    been the other way round. Some of this was related to change 7 above.
3648    
3649    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3650    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3651    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3652    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3653    
3654    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3655    mode. Unlike ".", it always matches newline, whatever the setting of
3656    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3657    assertions. Perl allows it, but it doesn't (in general) work because it can't
3658    calculate the length of the lookbehind. At least, that's the case for Perl
3659    5.8.0 - I've been told they are going to document that it doesn't work in
3660    future.
3661    
3662    18. Added an error diagnosis for escapes that PCRE does not support: these are
3663    \L, \l, \N, \P, \p, \U, \u, and \X.
3664    
3665    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3666    reading past the end of the pattern in cases such as /[abcd/.
3667    
3668    20. PCRE was getting more memory than necessary for patterns with classes that
3669    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3670    
3671    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3672    compiling PCRE for use with Virtual Pascal.
3673    
3674    22. Small fix to the Makefile to make it work properly if the build is done
3675    outside the source tree.
3676    
3677    23. Added a new extension: a condition to go with recursion. If a conditional
3678    subpattern starts with (?(R) the "true" branch is used if recursion has
3679    happened, whereas the "false" branch is used only at the top level.
3680    
3681    24. When there was a very long string of literal characters (over 255 bytes
3682    without UTF support, over 250 bytes with UTF support), the computation of how
3683    much memory was required could be incorrect, leading to segfaults or other
3684    strange effects.
3685    
3686    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3687    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3688    there was a subsequent back reference to those brackets. This meant that, for
3689    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3690    possible to check for precisely this case. All we can do is abandon the
3691    optimization if .* occurs inside capturing brackets when there are any back
3692    references whatsoever. (See below for a better fix that came later.)
3693    
3694    26. The handling of the optimization for finding the first character of a
3695    non-anchored pattern, and for finding a character that is required later in the
3696    match were failing in some cases. This didn't break the matching; it just
3697    failed to optimize when it could. The way this is done has been re-implemented.
3698    
3699    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3700    
3701    28. Added a new feature that provides some of the functionality that Perl
3702    provides with (?{...}). The facility is termed a "callout". The way it is done
3703    in PCRE is for the caller to provide an optional function, by setting
3704    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3705    global variable. By default it is unset, which disables all calling out. To get
3706    the function called, the regex must include (?C) at appropriate points. This
3707    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3708    This provides a means of identifying different callout points. When PCRE
3709    reaches such a point in the regex, if pcre_callout has been set, the external
3710    function is called. It is provided with data in a structure called
3711    pcre_callout_block, which is