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

Diff of /code/trunk/ChangeLog

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

revision 23 by nigel, Sat Feb 24 21:38:41 2007 UTC revision 961 by ph10, Fri Apr 20 11:49:13 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    22. Changed the option for creating symbolic links for 16-bit man pages from
81        -s to -sf so that re-installing does not cause issues.
82    
83    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84    
85    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86        matches in certain environments (the workspace was not being correctly
87        retained). Also added to pcre_dfa_exec() a simple plausibility check on
88        some of the workspace data at the beginning of a restart.
89    
90    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91        was not doing so when it should - probably a typo introduced by SVN 528
92        (change 8.10/14).
93    
94    
95    Version 8.30 04-February-2012
96    -----------------------------
97    
98    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
99        name is defined in ctype.h.
100    
101    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
102        only in quite long subpatterns.
103    
104    3.  Removed the function pcre_info(), which has been obsolete and deprecated
105        since it was replaced by pcre_fullinfo() in February 2000.
106    
107    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
108        match a (*MARK), and the match failed at the start of the subject, a
109        reference to memory before the start of the subject could occur. This bug
110        was introduced by fix 17 of release 8.21.
111    
112    5.  A reference to an unset group with zero minimum repetition was giving
113        totally wrong answers (in non-JavaScript-compatibility mode). For example,
114        /(another)?(\1?)test/ matched against "hello world test". This bug was
115        introduced in release 8.13.
116    
117    6.  Add support for 16-bit character strings (a large amount of work involving
118        many changes and refactorings).
119    
120    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
121        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
122        from a file.
123    
124    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
125        rounding is not applied in this particular case).
126    
127    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
128        if they appear, or are escaped, in patterns.
129    
130    10. Get rid of a number of -Wunused-but-set-variable warnings.
131    
132    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
133        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
134        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
135        also returns the mark "x". This bug applied to capturing parentheses,
136        non-capturing parentheses, and atomic parentheses. It also applied to some
137        assertions.
138    
139    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
140        information out of configure.ac instead of relying on pcre.h.generic, which
141        is not stored in the repository.
142    
143    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
144        -lreadline.
145    
146    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
147    
148    15. Applied Graycode's patch to put the top-level frame on the stack rather
149        than the heap when not using the stack for recursion. This gives a
150        performance improvement in many cases when recursion is not deep.
151    
152    16. Experimental code added to "pcretest -C" to output the stack frame size.
153    
154    
155    Version 8.21 12-Dec-2011
156    ------------------------
157    
158    1.  Updating the JIT compiler.
159    
160    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
161        are added as well.
162    
163    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
164        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
165        calling _pcre_jit_exec. Some extra comments are added.
166    
167    4.  (*MARK) settings inside atomic groups that do not contain any capturing
168        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
169        was introduced by change 18 for 8.20.
170    
171    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
172        ECMA-262 standard.
173    
174    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
175        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
176        This bug was probably introduced by change 9 of 8.13.
177    
178    7.  While fixing 6 above, I noticed that a number of other items were being
179        incorrectly rejected as "not fixed length". This arose partly because newer
180        opcodes had not been added to the fixed-length checking code. I have (a)
181        corrected the bug and added tests for these items, and (b) arranged for an
182        error to occur if an unknown opcode is encountered while checking for fixed
183        length instead of just assuming "not fixed length". The items that were
184        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
185        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
186        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
187    
188    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
189        being incorrectly compiled and would have given unpredicatble results.
190    
191    9.  A possessively repeated subpattern with minimum repeat count greater than
192        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
193        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
194        the first (A) could occur when it should not.
195    
196    10. Add a cast and remove a redundant test from the code.
197    
198    11. JIT should use pcre_malloc/pcre_free for allocation.
199    
200    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
201        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
202        is anything other than /usr, -L is still shown).
203    
204    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
205    
206    14. Perl does not support \N without a following name in a [] class; PCRE now
207        also gives an error.
208    
209    15. If a forward reference was repeated with an upper limit of around 2000,
210        it caused the error "internal error: overran compiling workspace". The
211        maximum number of forward references (including repeats) was limited by the
212        internal workspace, and dependent on the LINK_SIZE. The code has been
213        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
214        the default depends on LINK_SIZE. There is a new upper limit (for safety)
215        of around 200,000 forward references. While doing this, I also speeded up
216        the filling in of repeated forward references.
217    
218    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
219        incorrectly expecting the subject to contain another "a" after the start.
220    
221    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
222        in the match, the SKIP should be ignored. This was not happening; instead
223        the SKIP was being treated as NOMATCH. For patterns such as
224        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
225        tested.
226    
227    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
228        now much more compatible with Perl, in particular in cases where the result
229        is a non-match for a non-anchored pattern. For example, if
230        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
231        "m", where previously it did not return a name. A side effect of this
232        change is that for partial matches, the last encountered mark name is
233        returned, as for non matches. A number of tests that were previously not
234        Perl-compatible have been moved into the Perl-compatible test files. The
235        refactoring has had the pleasing side effect of removing one argument from
236        the match() function, thus reducing its stack requirements.
237    
238    19. If the /S+ option was used in pcretest to study a pattern using JIT,
239        subsequent uses of /S (without +) incorrectly behaved like /S+.
240    
241    21. Retrieve executable code size support for the JIT compiler and fixing
242        some warnings.
243    
244    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
245        not work when the shorter character appeared right at the end of the
246        subject string.
247    
248    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
249        systems.
250    
251    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
252        output it when the /M option is used in pcretest.
253    
254    25. The CheckMan script was not being included in the distribution. Also, added
255        an explicit "perl" to run Perl scripts from the PrepareRelease script
256        because this is reportedly needed in Windows.
257    
258    26. If study data was being save in a file and studying had not found a set of
259        "starts with" bytes for the pattern, the data written to the file (though
260        never used) was taken from uninitialized memory and so caused valgrind to
261        complain.
262    
263    27. Updated RunTest.bat as provided by Sheri Pierce.
264    
265    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
266    
267    29. Computation of memory usage for the table of capturing group names was
268        giving an unnecessarily large value.
269    
270    
271    Version 8.20 21-Oct-2011
272    ------------------------
273    
274    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
275        a POSIX class. After further experiments with Perl, which convinced me that
276        Perl has bugs and confusions, a closing square bracket is no longer allowed
277        in a POSIX name. This bug also affected patterns with classes that started
278        with full stops.
279    
280    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
281        captured substring, but while checking the failing first alternative,
282        substring 1 is temporarily captured. If the output vector supplied to
283        pcre_exec() was not big enough for this capture, the yield of the function
284        was still zero ("insufficient space for captured substrings"). This cannot
285        be totally fixed without adding another stack variable, which seems a lot
286        of expense for a edge case. However, I have improved the situation in cases
287        such as /(a)(b)x|abc/ matched against "abc", where the return code
288        indicates that fewer than the maximum number of slots in the ovector have
289        been set.
290    
291    3.  Related to (2) above: when there are more back references in a pattern than
292        slots in the output vector, pcre_exec() uses temporary memory during
293        matching, and copies in the captures as far as possible afterwards. It was
294        using the entire output vector, but this conflicts with the specification
295        that only 2/3 is used for passing back captured substrings. Now it uses
296        only the first 2/3, for compatibility. This is, of course, another edge
297        case.
298    
299    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
300        main code base, and can be used by building with --enable-jit. When this is
301        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
302        runtime --no-jit option is given.
303    
304    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
305        ovector, the return from the function was zero, implying that there were
306        other matches that did not fit. The correct "exactly full" value is now
307        returned.
308    
309    6.  If a subpattern that was called recursively or as a subroutine contained
310        (*PRUNE) or any other control that caused it to give a non-standard return,
311        invalid errors such as "Error -26 (nested recursion at the same subject
312        position)" or even infinite loops could occur.
313    
314    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
315        computing the minimum length on reaching *ACCEPT, and so ended up with the
316        wrong value of 1 rather than 0. Further investigation indicates that
317        computing a minimum subject length in the presence of *ACCEPT is difficult
318        (think back references, subroutine calls), and so I have changed the code
319        so that no minimum is registered for a pattern that contains *ACCEPT.
320    
321    8.  If (*THEN) was present in the first (true) branch of a conditional group,
322        it was not handled as intended. [But see 16 below.]
323    
324    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
325        Sheri Pierce.
326    
327    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
328        the first byte in a match must be "a".
329    
330    11. Change 17 for 8.13 increased the recursion depth for patterns like
331        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
332        pattern contains any instances of (*THEN). If it does not, the old
333        optimizations are restored. It would be nice to do this on a per-group
334        basis, but at the moment that is not feasible.
335    
336    12. In some environments, the output of pcretest -C is CRLF terminated. This
337        broke RunTest's code that checks for the link size. A single white space
338        character after the value is now allowed for.
339    
340    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
341        For "fr", it uses the Windows-specific input and output files.
342    
343    14. If (*THEN) appeared in a group that was called recursively or as a
344        subroutine, it did not work as intended. [But see next item.]
345    
346    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
347        pattern fragments (but not containing any | characters). If A and B are
348        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
349        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
350        D. In other words, Perl considers parentheses that do not contain any |
351        characters to be part of a surrounding alternative, whereas PCRE was
352        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
353        differently. PCRE now behaves in the same way as Perl, except in the case
354        of subroutine/recursion calls such as (?1) which have in any case always
355        been different (but PCRE had them first :-).
356    
357    16. Related to 15 above: Perl does not treat the | in a conditional group as
358        creating alternatives. Such a group is treated in the same way as an
359        ordinary group without any | characters when processing (*THEN). PCRE has
360        been changed to match Perl's behaviour.
361    
362    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
363        RunGrepTest script failed.
364    
365    18. Change 22 for version 13 caused atomic groups to use more stack. This is
366        inevitable for groups that contain captures, but it can lead to a lot of
367        stack use in large patterns. The old behaviour has been restored for atomic
368        groups that do not contain any capturing parentheses.
369    
370    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
371        suppress the check for a minimum subject length at run time. (If it was
372        given to pcre_exec() or pcre_dfa_exec() it did work.)
373    
374    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
375        fail to work when decoding hex characters in data strings in EBCDIC
376        environments.
377    
378    21. It appears that in at least one Mac OS environment, the isxdigit() function
379        is implemented as a macro that evaluates to its argument more than once,
380        contravening the C 90 Standard (I haven't checked a later standard). There
381        was an instance in pcretest which caused it to go wrong when processing
382        \x{...} escapes in subject strings. The has been rewritten to avoid using
383        things like p++ in the argument of isxdigit().
384    
385    
386    Version 8.13 16-Aug-2011
387    ------------------------
388    
389    1.  The Unicode data tables have been updated to Unicode 6.0.0.
390    
391    2.  Two minor typos in pcre_internal.h have been fixed.
392    
393    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
394        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
395        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
396    
397    4.  There were a number of related bugs in the code for matching backrefences
398        caselessly in UTF-8 mode when codes for the characters concerned were
399        different numbers of bytes. For example, U+023A and U+2C65 are an upper
400        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
401        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
402        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
403        2-byte code at the end of the subject (it thought there wasn't enough data
404        left).
405    
406    5.  Comprehensive information about what went wrong is now returned by
407        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
408        as the output vector has at least 2 elements. The offset of the start of
409        the failing character and a reason code are placed in the vector.
410    
411    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
412        now returned is for the first byte of the failing character, instead of the
413        last byte inspected. This is an incompatible change, but I hope it is small
414        enough not to be a problem. It makes the returned offset consistent with
415        pcre_exec() and pcre_dfa_exec().
416    
417    7.  pcretest now gives a text phrase as well as the error number when
418        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
419        failure, the offset and reason code are output.
420    
421    8.  When \R was used with a maximizing quantifier it failed to skip backwards
422        over a \r\n pair if the subsequent match failed. Instead, it just skipped
423        back over a single character (\n). This seems wrong (because it treated the
424        two characters as a single entity when going forwards), conflicts with the
425        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
426        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
427        has been changed.
428    
429    9.  Some internal refactoring has changed the processing so that the handling
430        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
431        time (the PCRE_DOTALL option was changed this way some time ago: version
432        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
433        which was always a bit of a fudge. It also means that there is one less
434        argument for the match() function, which reduces its stack requirements
435        slightly. This change also fixes an incompatibility with Perl: the pattern
436        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
437    
438    10. More internal refactoring has drastically reduced the number of recursive
439        calls to match() for possessively repeated groups such as (abc)++ when
440        using pcre_exec().
441    
442    11. While implementing 10, a number of bugs in the handling of groups were
443        discovered and fixed:
444    
445        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
446        (a|)*(?1) gave a compile-time internal error.
447        ((a|)+)+  did not notice that the outer group could match an empty string.
448        (^a|^)+   was not marked as anchored.
449        (.*a|.*)+ was not marked as matching at start or after a newline.
450    
451    12. Yet more internal refactoring has removed another argument from the match()
452        function. Special calls to this function are now indicated by setting a
453        value in a variable in the "match data" data block.
454    
455    13. Be more explicit in pcre_study() instead of relying on "default" for
456        opcodes that mean there is no starting character; this means that when new
457        ones are added and accidentally left out of pcre_study(), testing should
458        pick them up.
459    
460    14. The -s option of pcretest has been documented for ages as being an old
461        synonym of -m (show memory usage). I have changed it to mean "force study
462        for every regex", that is, assume /S for every regex. This is similar to -i
463        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
464        using it. It makes it easier to run collections of tests with and without
465        study enabled, and thereby test pcre_study() more easily. All the standard
466        tests are now run with and without -s (but some patterns can be marked as
467        "never study" - see 20 below).
468    
469    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
470        restoration of the capturing data to the outer values was not happening
471        correctly.
472    
473    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
474        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
475        pattern had matched an empty string, and so incorrectly returned a no
476        match.
477    
478    17. There was optimizing code for the last branch of non-capturing parentheses,
479        and also for the obeyed branch of a conditional subexpression, which used
480        tail recursion to cut down on stack usage. Unfortunately, now that there is
481        the possibility of (*THEN) occurring in these branches, tail recursion is
482        no longer possible because the return has to be checked for (*THEN). These
483        two optimizations have therefore been removed. [But see 8.20/11 above.]
484    
485    18. If a pattern containing \R was studied, it was assumed that \R always
486        matched two bytes, thus causing the minimum subject length to be
487        incorrectly computed because \R can also match just one byte.
488    
489    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
490        was incorrectly computed.
491    
492    20. If /S is present twice on a test pattern in pcretest input, it now
493        *disables* studying, thereby overriding the use of -s on the command line
494        (see 14 above). This is necessary for one or two tests to keep the output
495        identical in both cases.
496    
497    21. When (*ACCEPT) was used in an assertion that matched an empty string and
498        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
499    
500    22. When an atomic group that contained a capturing parenthesis was
501        successfully matched, but the branch in which it appeared failed, the
502        capturing was not being forgotten if a higher numbered group was later
503        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
504        group 1 to "a", when in fact it should be unset. This applied to multi-
505        branched capturing and non-capturing groups, repeated or not, and also to
506        positive assertions (capturing in negative assertions does not happen
507        in PCRE) and also to nested atomic groups.
508    
509    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
510        subject after a captured substring, to make it easier to tell which of a
511        number of identical substrings has been captured.
512    
513    24. The way atomic groups are processed by pcre_exec() has been changed so that
514        if they are repeated, backtracking one repetition now resets captured
515        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
516        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
517        "aaa". Previously, it would have been "a". As part of this code
518        refactoring, the way recursive calls are handled has also been changed.
519    
520    25. If an assertion condition captured any substrings, they were not passed
521        back unless some other capturing happened later. For example, if
522        (?(?=(a))a) was matched against "a", no capturing was returned.
523    
524    26. When studying a pattern that contained subroutine calls or assertions,
525        the code for finding the minimum length of a possible match was handling
526        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
527        group 1 called group 2 while simultaneously a separate group 2 called group
528        1). A stack overflow occurred in this case. I have fixed this by limiting
529        the recursion depth to 10.
530    
531    27. Updated RunTest.bat in the distribution to the version supplied by Tom
532        Fortmann. This supports explicit test numbers on the command line, and has
533        argument validation and error reporting.
534    
535    28. An instance of \X with an unlimited repeat could fail if at any point the
536        first character it looked at was a mark character.
537    
538    29. Some minor code refactoring concerning Unicode properties and scripts
539        should reduce the stack requirement of match() slightly.
540    
541    30. Added the '=' option to pcretest to check the setting of unused capturing
542        slots at the end of the pattern, which are documented as being -1, but are
543        not included in the return count.
544    
545    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
546        compiled something random. Now it gives a compile-time error (as does
547        Perl).
548    
549    32. A *MARK encountered during the processing of a positive assertion is now
550        recorded and passed back (compatible with Perl).
551    
552    33. If --only-matching or --colour was set on a pcregrep call whose pattern
553        had alternative anchored branches, the search for a second match in a line
554        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
555        matched the line "0102" twice. The same bug affected patterns that started
556        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
557        twice.
558    
559    34. Previously, PCRE did not allow quantification of assertions. However, Perl
560        does, and because of capturing effects, quantifying parenthesized
561        assertions may at times be useful. Quantifiers are now allowed for
562        parenthesized assertions.
563    
564    35. A minor code tidy in pcre_compile() when checking options for \R usage.
565    
566    36. \g was being checked for fancy things in a character class, when it should
567        just be a literal "g".
568    
569    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
570        appearance of a nested POSIX class supersedes an apparent external class.
571        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
572        unescaped square brackets may also appear as part of class names. For
573        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
574        more like Perl. (But see 8.20/1 above.)
575    
576    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
577        was because it thought it was \N{name}, which is not supported).
578    
579    39. Add minix to OS list not supporting the -S option in pcretest.
580    
581    40. PCRE tries to detect cases of infinite recursion at compile time, but it
582        cannot analyze patterns in sufficient detail to catch mutual recursions
583        such as ((?1))((?2)). There is now a runtime test that gives an error if a
584        subgroup is called recursively as a subpattern for a second time at the
585        same position in the subject string. In previous releases this might have
586        been caught by the recursion limit, or it might have run out of stack.
587    
588    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
589        happen only once. PCRE was, however incorrectly giving a compile time error
590        "recursive call could loop indefinitely" because it cannot analyze the
591        pattern in sufficient detail. The compile time test no longer happens when
592        PCRE is compiling a conditional subpattern, but actual runaway loops are
593        now caught at runtime (see 40 above).
594    
595    42. It seems that Perl allows any characters other than a closing parenthesis
596        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
597        has been changed to be the same.
598    
599    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
600        as not to get warnings when autogen.sh is called. Also changed
601        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
602    
603    44. To help people who use pcregrep to scan files containing exceedingly long
604        lines, the following changes have been made:
605    
606        (a) The default value of the buffer size parameter has been increased from
607            8K to 20K. (The actual buffer used is three times this size.)
608    
609        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
610            PCRE is built.
611    
612        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
613            to be set at run time.
614    
615        (d) Numerical values in pcregrep options can be followed by K or M, for
616            example --buffer-size=50K.
617    
618        (e) If a line being scanned overflows pcregrep's buffer, an error is now
619            given and the return code is set to 2.
620    
621    45. Add a pointer to the latest mark to the callout data block.
622    
623    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
624        partial match of an empty string instead of no match. This was specific to
625        the use of ".".
626    
627    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
628        complete match instead of a partial match. This bug was dependent on both
629        the PCRE_UTF8 and PCRE_DOTALL options being set.
630    
631    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
632        starting byte set, because \b was not being ignored.
633    
634    
635    Version 8.12 15-Jan-2011
636    ------------------------
637    
638    1.  Fixed some typos in the markup of the man pages, and wrote a script that
639        checks for such things as part of the documentation building process.
640    
641    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
642        --match-limit and --recursion-limit options (added for 8.11). In
643        particular, this made one of the standard tests fail. (The integer value
644        went into the wrong half of a long int.)
645    
646    3.  If the --colour option was given to pcregrep with -v (invert match), it
647        did strange things, either producing crazy output, or crashing. It should,
648        of course, ignore a request for colour when reporting lines that do not
649        match.
650    
651    4.  Another pcregrep bug caused similar problems if --colour was specified with
652        -M (multiline) and the pattern match finished with a line ending.
653    
654    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
655        matched in multiline mode, the following line was shown as part of the
656        match. This seems wrong, so I have changed it.
657    
658    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
659        the check for further matches in the same line (so they could be coloured)
660        to overrun the end of the current line. If another match was found, it was
661        incorrectly shown (and then shown again when found in the next line).
662    
663    7.  If pcregrep was compiled under Windows, there was a reference to the
664        function pcregrep_exit() before it was defined. I am assuming this was
665        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
666        reported by a user. I've moved the definition above the reference.
667    
668    
669    Version 8.11 10-Dec-2010
670    ------------------------
671    
672    1.  (*THEN) was not working properly if there were untried alternatives prior
673        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
674        backtracked to try for "b" instead of moving to the next alternative branch
675        at the same level (in this case, to look for "c"). The Perl documentation
676        is clear that when (*THEN) is backtracked onto, it goes to the "next
677        alternative in the innermost enclosing group".
678    
679    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
680        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
681        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
682        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
683        (*THEN).
684    
685    3.  If \s appeared in a character class, it removed the VT character from
686        the class, even if it had been included by some previous item, for example
687        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
688        of \s, but is part of the POSIX "space" class.)
689    
690    4.  A partial match never returns an empty string (because you can always
691        match an empty string at the end of the subject); however the checking for
692        an empty string was starting at the "start of match" point. This has been
693        changed to the "earliest inspected character" point, because the returned
694        data for a partial match starts at this character. This means that, for
695        example, /(?<=abc)def/ gives a partial match for the subject "abc"
696        (previously it gave "no match").
697    
698    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
699        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
700        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
701        has an implication that the given string is incomplete (because a partial
702        match is preferred over a full match). For this reason, these items now
703        give a partial match in this situation. [Aside: previously, the one case
704        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
705        match rather than a full match, which was wrong by the old rules, but is
706        now correct.]
707    
708    6.  There was a bug in the handling of #-introduced comments, recognized when
709        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
710        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
711        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
712        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
713        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
714        places in pcre_compile().
715    
716    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
717        comments when looking ahead for named forward references to subpatterns,
718        the only newline sequence it recognized was NL. It now handles newlines
719        according to the set newline convention.
720    
721    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
722        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
723        cater for a lack of strerror(). These oversights have been fixed.
724    
725    9.  Added --match-limit and --recursion-limit to pcregrep.
726    
727    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
728    
729    11. When the -o option was used, pcregrep was setting a return code of 1, even
730        when matches were found, and --line-buffered was not being honoured.
731    
732    12. Added an optional parentheses number to the -o and --only-matching options
733        of pcregrep.
734    
735    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
736        can match an empty string. The code to do it in pcretest and pcredemo
737        needed fixing:
738    
739        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
740            only one byte after an empty string match just before CRLF (this case
741            just got forgotten; "any" and "anycrlf" were OK).
742    
743        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
744            mode when an empty string match preceded an ASCII character followed by
745            a non-ASCII character. (The code for advancing by one character rather
746            than one byte was nonsense.)
747    
748        (c) The pcredemo.c sample program did not have any code at all to handle
749            the cases when CRLF is a valid newline sequence.
750    
751    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
752        as a starting offset was within the subject string. There is now a new
753        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
754        negative or greater than the length of the string. In order to test this,
755        pcretest is extended to allow the setting of negative starting offsets.
756    
757    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
758        starting offset points to the beginning of a UTF-8 character was
759        unnecessarily clumsy. I tidied it up.
760    
761    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
762        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
763    
764    17. Nobody had reported that the --include_dir option, which was added in
765        release 7.7 should have been called --include-dir (hyphen, not underscore)
766        for compatibility with GNU grep. I have changed it to --include-dir, but
767        left --include_dir as an undocumented synonym, and the same for
768        --exclude-dir, though that is not available in GNU grep, at least as of
769        release 2.5.4.
770    
771    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
772        characters from a string of bytes) have been redefined so as not to use
773        loops, in order to improve performance in some environments. At the same
774        time, I abstracted some of the common code into auxiliary macros to save
775        repetition (this should not affect the compiled code).
776    
777    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
778        compile-time error is now given if \c is not followed by an ASCII
779        character, that is, a byte less than 128. (In EBCDIC mode, the code is
780        different, and any byte value is allowed.)
781    
782    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
783        START_OPTIMIZE option, which is now allowed at compile time - but just
784        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
785        to pcregrep and other applications that have no direct access to PCRE
786        options. The new /Y option in pcretest sets this option when calling
787        pcre_compile().
788    
789    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
790        back references. Groups containing recursive back references were forced to
791        be atomic by that change, but in the case of named groups, the amount of
792        memory required was incorrectly computed, leading to "Failed: internal
793        error: code overflow". This has been fixed.
794    
795    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
796        pcretest.c, to avoid build problems in some Borland environments.
797    
798    
799    Version 8.10 25-Jun-2010
800    ------------------------
801    
802    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
803        THEN.
804    
805    2.  (*ACCEPT) was not working when inside an atomic group.
806    
807    3.  Inside a character class, \B is treated as a literal by default, but
808        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
809        causes the error). The code is unchanged, but I tidied the documentation.
810    
811    4.  Inside a character class, PCRE always treated \R and \X as literals,
812        whereas Perl faults them if its -w option is set. I have changed PCRE so
813        that it faults them when PCRE_EXTRA is set.
814    
815    5.  Added support for \N, which always matches any character other than
816        newline. (It is the same as "." when PCRE_DOTALL is not set.)
817    
818    6.  When compiling pcregrep with newer versions of gcc which may have
819        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
820        declared with attribute warn_unused_result" were given. Just casting the
821        result to (void) does not stop the warnings; a more elaborate fudge is
822        needed. I've used a macro to implement this.
823    
824    7.  Minor change to pcretest.c to avoid a compiler warning.
825    
826    8.  Added four artifical Unicode properties to help with an option to make
827        \s etc use properties (see next item). The new properties are: Xan
828        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
829    
830    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
831        use Unicode properties. (*UCP) at the start of a pattern can be used to set
832        this option. Modified pcretest to add /W to test this facility. Added
833        REG_UCP to make it available via the POSIX interface.
834    
835    10. Added --line-buffered to pcregrep.
836    
837    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
838        studied, and the match started with a letter with a code point greater than
839        127 whose first byte was different to the first byte of the other case of
840        the letter, the other case of this starting letter was not recognized
841        (#976).
842    
843    12. If a pattern that was studied started with a repeated Unicode property
844        test, for example, \p{Nd}+, there was the theoretical possibility of
845        setting up an incorrect bitmap of starting bytes, but fortunately it could
846        not have actually happened in practice until change 8 above was made (it
847        added property types that matched character-matching opcodes).
848    
849    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
850        possible starting bytes for non-anchored patterns.
851    
852    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
853        \R, and also a number of cases that involve Unicode properties, both
854        explicit and implicit when PCRE_UCP is set.
855    
856    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
857        input, it could crash or give wrong results if characters with values
858        greater than 0xc0 were present in the subject string. (Detail: it assumed
859        UTF-8 input when processing these items.)
860    
861    16. Added a lot of (int) casts to avoid compiler warnings in systems where
862        size_t is 64-bit (#991).
863    
864    17. Added a check for running out of memory when PCRE is compiled with
865        --disable-stack-for-recursion (#990).
866    
867    18. If the last data line in a file for pcretest does not have a newline on
868        the end, a newline was missing in the output.
869    
870    19. The default pcre_chartables.c file recognizes only ASCII characters (values
871        less than 128) in its various bitmaps. However, there is a facility for
872        generating tables according to the current locale when PCRE is compiled. It
873        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
874        characters, are recognized by isspace() and therefore were getting set in
875        these tables, and indeed these tables seem to approximate to ISO 8859. This
876        caused a problem in UTF-8 mode when pcre_study() was used to create a list
877        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
878        which of course cannot start UTF-8 characters. I have changed the code so
879        that only real ASCII characters (less than 128) and the correct starting
880        bytes for UTF-8 encodings are set for characters greater than 127 when in
881        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
882        altogether.)
883    
884    20. Added the /T option to pcretest so as to be able to run tests with non-
885        standard character tables, thus making it possible to include the tests
886        used for 19 above in the standard set of tests.
887    
888    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
889        reference to a subpattern the other side of a comment that contains an
890        opening parenthesis caused either an internal compiling error, or a
891        reference to the wrong subpattern.
892    
893    
894    Version 8.02 19-Mar-2010
895    ------------------------
896    
897    1.  The Unicode data tables have been updated to Unicode 5.2.0.
898    
899    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
900        configured.
901    
902    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
903        original author of that file, following a query about its status.
904    
905    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
906        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
907    
908    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
909        quantifier applied to a forward-referencing subroutine call, could compile
910        incorrect code or give the error "internal error: previously-checked
911        referenced subpattern not found".
912    
913    6.  Both MS Visual Studio and Symbian OS have problems with initializing
914        variables to point to external functions. For these systems, therefore,
915        pcre_malloc etc. are now initialized to local functions that call the
916        relevant global functions.
917    
918    7.  There were two entries missing in the vectors called coptable and poptable
919        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
920        I've fixed the data, and added a kludgy way of testing at compile time that
921        the lengths are correct (equal to the number of opcodes).
922    
923    8.  Following on from 7, I added a similar kludge to check the length of the
924        eint vector in pcreposix.c.
925    
926    9.  Error texts for pcre_compile() are held as one long string to avoid too
927        much relocation at load time. To find a text, the string is searched,
928        counting zeros. There was no check for running off the end of the string,
929        which could happen if a new error number was added without updating the
930        string.
931    
932    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
933    
934    11. \K was not working if it appeared in an atomic group or in a group that
935        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
936        \K is "not well defined" if used in an assertion. PCRE now accepts it if
937        the assertion is positive, but not if it is negative.
938    
939    12. Change 11 fortuitously reduced the size of the stack frame used in the
940        "match()" function of pcre_exec.c by one pointer. Forthcoming
941        implementation of support for (*MARK) will need an extra pointer on the
942        stack; I have reserved it now, so that the stack frame size does not
943        decrease.
944    
945    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
946        item in branch that calls a recursion is a subroutine call - as in the
947        second branch in the above example - was incorrectly given the compile-
948        time error "recursive call could loop indefinitely" because pcre_compile()
949        was not correctly checking the subroutine for matching a non-empty string.
950    
951    14. The checks for overrunning compiling workspace could trigger after an
952        overrun had occurred. This is a "should never occur" error, but it can be
953        triggered by pathological patterns such as hundreds of nested parentheses.
954        The checks now trigger 100 bytes before the end of the workspace.
955    
956    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
957    
958    
959    Version 8.01 19-Jan-2010
960    ------------------------
961    
962    1.  If a pattern contained a conditional subpattern with only one branch (in
963        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
964        computed the wrong minimum data length (which is of course zero for such
965        subpatterns). This could cause incorrect "no match" results.
966    
967    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
968        the pattern is reset in the first branch, pcre_compile() failed with
969        "internal error: code overflow at offset...". This happened only when
970        the reset was to the original external option setting. (An optimization
971        abstracts leading options settings into an external setting, which was the
972        cause of this.)
973    
974    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
975        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
976        assertion pattern did not match (meaning that the assertion was true), it
977        was incorrectly treated as false if the SKIP had been reached during the
978        matching. This also applied to assertions used as conditions.
979    
980    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
981        assertion subpattern, including such a pattern used as a condition,
982        unpredictable results occurred, instead of the error return
983        PCRE_ERROR_DFA_UITEM.
984    
985    5.  The C++ GlobalReplace function was not working like Perl for the special
986        situation when an empty string is matched. It now does the fancy magic
987        stuff that is necessary.
988    
989    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
990        removed. (These were left over from very, very early versions of PCRE.)
991    
992    7.  Some cosmetic changes to the code to make life easier when compiling it
993        as part of something else:
994    
995        (a) Change DEBUG to PCRE_DEBUG.
996    
997        (b) In pcre_compile(), rename the member of the "branch_chain" structure
998            called "current" as "current_branch", to prevent a collision with the
999            Linux macro when compiled as a kernel module.
1000    
1001        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1002            prevent a collision with the Linux macro when compiled as a kernel
1003            module.
1004    
1005    8.  In pcre_compile() there are some checks for integer overflows that used to
1006        cast potentially large values to (double). This has been changed to that
1007        when building, a check for int64_t is made, and if it is found, it is used
1008        instead, thus avoiding the use of floating point arithmetic. (There is no
1009        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1010        double.
1011    
1012    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1013        2005 (difference between two addresses compared to an unsigned value).
1014    
1015    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1016        custom one, because of the following reported problem in Windows:
1017    
1018          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1019              under Win32.
1020          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1021              therefore missing the function definition.
1022          - The compiler thus generates a "C" signature for the test function.
1023          - The linker fails to find the "C" function.
1024          - PCRE fails to configure if asked to do so against libbz2.
1025    
1026    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1027        messages were output:
1028    
1029          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1030          rerunning libtoolize, to keep the correct libtool macros in-tree.
1031          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1032    
1033        I have done both of these things.
1034    
1035    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1036        most of the time, it *can* run out if it is given a pattern that contains a
1037        runaway infinite recursion. I updated the discussion in the pcrestack man
1038        page.
1039    
1040    13. Now that we have gone to the x.xx style of version numbers, the minor
1041        version may start with zero. Using 08 or 09 is a bad idea because users
1042        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1043        interpreted as invalid octal numbers. I've updated the previous comment in
1044        configure.ac, and also added a check that gives an error if 08 or 09 are
1045        used.
1046    
1047    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1048        causing partial matching to fail when the end of the subject matched \W
1049        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1050    
1051    15. There were some discrepancies between the declarations in pcre_internal.h
1052        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1053        their definitions. The declarations used "const uschar *" and the
1054        definitions used USPTR. Even though USPTR is normally defined as "const
1055        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1056        reported that: "This difference in casting confuses some C++ compilers, for
1057        example, SunCC recognizes above declarations as different functions and
1058        generates broken code for hbpcre." I have changed the declarations to use
1059        USPTR.
1060    
1061    16. GNU libtool is named differently on some systems. The autogen.sh script now
1062        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1063        (FreeBSD).
1064    
1065    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1066        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1067        comment: "Figure out how to create a longlong from a string: strtoll and
1068        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1069        instance, but it only takes 2 args instead of 3!"
1070    
1071    18. A subtle bug concerned with back references has been fixed by a change of
1072        specification, with a corresponding code fix. A pattern such as
1073        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1074        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1075        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1076        same bug. Such groups have to be quantified to be useful, or contained
1077        inside another quantified group. (If there's no repetition, the reference
1078        can never match.) The problem arises because, having left the group and
1079        moved on to the rest of the pattern, a later failure that backtracks into
1080        the group uses the captured value from the final iteration of the group
1081        rather than the correct earlier one. I have fixed this in PCRE by forcing
1082        any group that contains a reference to itself to be an atomic group; that
1083        is, there cannot be any backtracking into it once it has completed. This is
1084        similar to recursive and subroutine calls.
1085    
1086    
1087    Version 8.00 19-Oct-09
1088    ----------------------
1089    
1090    1.  The table for translating pcre_compile() error codes into POSIX error codes
1091        was out-of-date, and there was no check on the pcre_compile() error code
1092        being within the table. This could lead to an OK return being given in
1093        error.
1094    
1095    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1096        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1097        in a Windows environment.
1098    
1099    3.  The pcregrep --count option prints the count for each file even when it is
1100        zero, as does GNU grep. However, pcregrep was also printing all files when
1101        --files-with-matches was added. Now, when both options are given, it prints
1102        counts only for those files that have at least one match. (GNU grep just
1103        prints the file name in this circumstance, but including the count seems
1104        more useful - otherwise, why use --count?) Also ensured that the
1105        combination -clh just lists non-zero counts, with no names.
1106    
1107    4.  The long form of the pcregrep -F option was incorrectly implemented as
1108        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1109        but it seems right to fix it, and I didn't think it was worth preserving
1110        the old behaviour.
1111    
1112    5.  The command line items --regex=pattern and --regexp=pattern were not
1113        recognized by pcregrep, which required --regex pattern or --regexp pattern
1114        (with a space rather than an '='). The man page documented the '=' forms,
1115        which are compatible with GNU grep; these now work.
1116    
1117    6.  No libpcreposix.pc file was created for pkg-config; there was just
1118        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1119    
1120    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1121        when UCP support is not needed, by modifying the Python script that
1122        generates it from Unicode data files. This should not matter if the module
1123        is correctly used as a library, but I received one complaint about 50K of
1124        unwanted data. My guess is that the person linked everything into his
1125        program rather than using a library. Anyway, it does no harm.
1126    
1127    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1128        was a minimum greater than 1 for a wide character in a possessive
1129        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1130        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1131        character. Chaos in the form of incorrect output or a compiling loop could
1132        result.
1133    
1134    9.  The restrictions on what a pattern can contain when partial matching is
1135        requested for pcre_exec() have been removed. All patterns can now be
1136        partially matched by this function. In addition, if there are at least two
1137        slots in the offset vector, the offset of the earliest inspected character
1138        for the match and the offset of the end of the subject are set in them when
1139        PCRE_ERROR_PARTIAL is returned.
1140    
1141    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1142        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1143        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1144        and may be more useful for multi-segment matching.
1145    
1146    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1147        used to be given if ever the end of the subject was reached; now it is
1148        given only if matching could not proceed because another character was
1149        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1150        string "Z", which now yields "no match" instead of "partial match". In the
1151        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1152        final character ended with (*FAIL).
1153    
1154    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1155        if the pattern had a "must contain" character that was already found in the
1156        earlier partial match, unless partial matching was again requested. For
1157        example, with the pattern /dog.(body)?/, the "must contain" character is
1158        "g". If the first part-match was for the string "dog", restarting with
1159        "sbody" failed. This bug has been fixed.
1160    
1161    13. The string returned by pcre_dfa_exec() after a partial match has been
1162        changed so that it starts at the first inspected character rather than the
1163        first character of the match. This makes a difference only if the pattern
1164        starts with a lookbehind assertion or \b or \B (\K is not supported by
1165        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1166        matching functions compatible, and I think it's the right thing to do.
1167    
1168    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1169        so that the demonstration program is easily available in environments where
1170        PCRE has not been installed from source.
1171    
1172    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1173        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1174        library.
1175    
1176    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1177        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1178        is not the first non-POSIX option to be added. Clearly some people find
1179        these options useful.
1180    
1181    17. If a caller to the POSIX matching function regexec() passes a non-zero
1182        value for nmatch with a NULL value for pmatch, the value of
1183        nmatch is forced to zero.
1184    
1185    18. RunGrepTest did not have a test for the availability of the -u option of
1186        the diff command, as RunTest does. It now checks in the same way as
1187        RunTest, and also checks for the -b option.
1188    
1189    19. If an odd number of negated classes containing just a single character
1190        interposed, within parentheses, between a forward reference to a named
1191        subpattern and the definition of the subpattern, compilation crashed with
1192        an internal error, complaining that it could not find the referenced
1193        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1194        [The bug was that it was starting one character too far in when skipping
1195        over the character class, thus treating the ] as data rather than
1196        terminating the class. This meant it could skip too much.]
1197    
1198    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1199        /g option in pcretest when the pattern contains \K, which makes it possible
1200        to have an empty string match not at the start, even when the pattern is
1201        anchored. Updated pcretest and pcredemo to use this option.
1202    
1203    21. If the maximum number of capturing subpatterns in a recursion was greater
1204        than the maximum at the outer level, the higher number was returned, but
1205        with unset values at the outer level. The correct (outer level) value is
1206        now given.
1207    
1208    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1209        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1210        make it do so. The string so far is captured, making this feature
1211        compatible with Perl.
1212    
1213    23. The tests have been re-organized, adding tests 11 and 12, to make it
1214        possible to check the Perl 5.10 features against Perl 5.10.
1215    
1216    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1217        pattern matches a fixed length string. PCRE did not allow this; now it
1218        does. Neither allows recursion.
1219    
1220    25. I finally figured out how to implement a request to provide the minimum
1221        length of subject string that was needed in order to match a given pattern.
1222        (It was back references and recursion that I had previously got hung up
1223        on.) This code has now been added to pcre_study(); it finds a lower bound
1224        to the length of subject needed. It is not necessarily the greatest lower
1225        bound, but using it to avoid searching strings that are too short does give
1226        some useful speed-ups. The value is available to calling programs via
1227        pcre_fullinfo().
1228    
1229    26. While implementing 25, I discovered to my embarrassment that pcretest had
1230        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1231        study optimizations had never been tested with that matching function.
1232        Oops. What is worse, even when it was passed study data, there was a bug in
1233        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1234        were also very few tests of studied patterns with pcre_dfa_exec().
1235    
1236    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1237        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1238        on the other side of the coin, they are no longer allowed to have different
1239        names, because these cannot be distinguished in PCRE, and this has caused
1240        confusion. (This is a difference from Perl.)
1241    
1242    28. When duplicate subpattern names are present (necessarily with different
1243        numbers, as required by 27 above), and a test is made by name in a
1244        conditional pattern, either for a subpattern having been matched, or for
1245        recursion in such a pattern, all the associated numbered subpatterns are
1246        tested, and the overall condition is true if the condition is true for any
1247        one of them. This is the way Perl works, and is also more like the way
1248        testing by number works.
1249    
1250    
1251    Version 7.9 11-Apr-09
1252    ---------------------
1253    
1254    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1255        (pcretest), all targets were linked against these libraries. This included
1256        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1257        libraries. This caused unwanted dependencies to be created. This problem
1258        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1259        pcretest is linked with readline.
1260    
1261    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1262        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1263        moved outside it again, because FALSE and TRUE are already defined in AIX,
1264        but BOOL is not.
1265    
1266    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1267        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1268    
1269    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1270        hyphens) following file names and line numbers when outputting matching
1271        lines. This is not true; no spaces are inserted. I have also clarified the
1272        wording for the --colour (or --color) option.
1273    
1274    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1275        was not coloured; this is different to GNU grep, so I have changed it to be
1276        the same.
1277    
1278    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1279        each matching line was coloured. Now it goes on to look for further matches
1280        of any of the test patterns, which is the same behaviour as GNU grep.
1281    
1282    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1283        doesn't make sense to accept an empty string match in pcregrep, so I have
1284        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1285        seems to be how GNU grep behaves.
1286    
1287    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1288        start or after a newline", because the conditional assertion was not being
1289        correctly handled. The rule now is that both the assertion and what follows
1290        in the first alternative must satisfy the test.
1291    
1292    9.  If auto-callout was enabled in a pattern with a conditional group whose
1293        condition was an assertion, PCRE could crash during matching, both with
1294        pcre_exec() and pcre_dfa_exec().
1295    
1296    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1297        used for matching.
1298    
1299    11. Unicode property support in character classes was not working for
1300        characters (bytes) greater than 127 when not in UTF-8 mode.
1301    
1302    12. Added the -M command line option to pcretest.
1303    
1304    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1305    
1306    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1307    
1308    16. Added comments and documentation about mis-use of no_arg in the C++
1309        wrapper.
1310    
1311    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1312        from Martin Jerabek that uses macro names for all relevant character and
1313        string constants.
1314    
1315    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1316        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1317        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1318        these, but not everybody uses configure.
1319    
1320    19. A conditional group that had only one branch was not being correctly
1321        recognized as an item that could match an empty string. This meant that an
1322        enclosing group might also not be so recognized, causing infinite looping
1323        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1324        with the subject "ab", where knowledge that the repeated group can match
1325        nothing is needed in order to break the loop.
1326    
1327    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1328        exec(), but without supplying a callout function, matching went wrong.
1329    
1330    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1331        leak if the size of the offset vector was greater than 30. When the vector
1332        is smaller, the saved offsets during recursion go onto a local stack
1333        vector, but for larger vectors malloc() is used. It was failing to free
1334        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1335        error, in fact).
1336    
1337    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1338        heapframe that is used only when UTF-8 support is enabled. This caused no
1339        problem, but was untidy.
1340    
1341    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1342        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1343        included within another project.
1344    
1345    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1346        slightly modified by me:
1347    
1348          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1349              not building pcregrep.
1350    
1351          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1352              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1353    
1354    25. Forward references, both numeric and by name, in patterns that made use of
1355        duplicate group numbers, could behave incorrectly or give incorrect errors,
1356        because when scanning forward to find the reference group, PCRE was not
1357        taking into account the duplicate group numbers. A pattern such as
1358        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1359    
1360    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1361        the feature of a custom pointer more persuasive (as requested by a user).
1362    
1363    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1364        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1365        pre-defined.
1366    
1367    28. Added support for (*UTF8) at the start of a pattern.
1368    
1369    29. Arrange for flags added by the "release type" setting in CMake to be shown
1370        in the configuration summary.
1371    
1372    
1373    Version 7.8 05-Sep-08
1374    ---------------------
1375    
1376    1.  Replaced UCP searching code with optimized version as implemented for Ad
1377        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1378        stage table and inline lookup instead of a function, giving speed ups of 2
1379        to 5 times on some simple patterns that I tested. Permission was given to
1380        distribute the MultiStage2.py script that generates the tables (it's not in
1381        the tarball, but is in the Subversion repository).
1382    
1383    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1384        scripts.
1385    
1386    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1387        a group with a zero qualifier. The result of the study could be incorrect,
1388        or the function might crash, depending on the pattern.
1389    
1390    4.  Caseless matching was not working for non-ASCII characters in back
1391        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1392        It now works when Unicode Property Support is available.
1393    
1394    5.  In pcretest, an escape such as \x{de} in the data was always generating
1395        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1396        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1397        truncation.
1398    
1399    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1400    
1401    7.  Added two (int) casts to pcregrep when printing the difference of two
1402        pointers, in case they are 64-bit values.
1403    
1404    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1405        test 2 if it fails.
1406    
1407    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1408        and a #define of that name to empty if it is not externally set. This is to
1409        allow users of MSVC to set it if necessary.
1410    
1411    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1412        the convenience functions in the pcre_get.c source file.
1413    
1414    11. An option change at the start of a pattern that had top-level alternatives
1415        could cause overwriting and/or a crash. This command provoked a crash in
1416        some environments:
1417    
1418          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1419    
1420        This potential security problem was recorded as CVE-2008-2371.
1421    
1422    12. For a pattern where the match had to start at the beginning or immediately
1423        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1424        pcre_dfa_exec() could read past the end of the passed subject if there was
1425        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1426        pcretest so that it places the subject at the end of its malloc-ed buffer.
1427    
1428    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1429        exec() might read past the end of the data buffer in UTF-8 mode.
1430    
1431    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1432        the data contained the byte 0x85 as part of a UTF-8 character within its
1433        first line. This applied both to normal and DFA matching.
1434    
1435    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1436        /^[^d]*?$/8 failed to match "abc".
1437    
1438    16. Added a missing copyright notice to pcrecpp_internal.h.
1439    
1440    17. Make it more clear in the documentation that values returned from
1441        pcre_exec() in ovector are byte offsets, not character counts.
1442    
1443    18. Tidied a few places to stop certain compilers from issuing warnings.
1444    
1445    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1446        supplied by Stefan Weber. I made a further small update for 7.8 because
1447        there is a change of source arrangements: the pcre_searchfuncs.c module is
1448        replaced by pcre_ucd.c.
1449    
1450    
1451    Version 7.7 07-May-08
1452    ---------------------
1453    
1454    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1455        a string to a long long, pretend we don't even have a long long." This is
1456        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1457    
1458    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1459        pre-7.6 versions, which defined a global no_arg variable instead of putting
1460        it in the RE class. (See also #8 below.)
1461    
1462    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1463        Lopes.
1464    
1465    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1466    
1467        (1) The include/exclude patterns were being applied to the whole pathnames
1468            of files, instead of just to the final components.
1469    
1470        (2) If there was more than one level of directory, the subdirectories were
1471            skipped unless they satisfied the include/exclude conditions. This is
1472            inconsistent with GNU grep (and could even be seen as contrary to the
1473            pcregrep specification - which I improved to make it absolutely clear).
1474            The action now is always to scan all levels of directory, and just
1475            apply the include/exclude patterns to regular files.
1476    
1477    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1478        --exclude_dir in the tests to avoid scanning .svn directories.
1479    
1480    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1481        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1482        doesn't support NULs in patterns.
1483    
1484    7.  Added some missing "const"s to declarations of static tables in
1485        pcre_compile.c and pcre_dfa_exec.c.
1486    
1487    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1488        caused by fix #2  above. (Subsequently also a second patch to fix the
1489        first patch. And a third patch - this was a messy problem.)
1490    
1491    9.  Applied Craig's patch to remove the use of push_back().
1492    
1493    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1494        matching function regexec().
1495    
1496    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1497        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1498        references. PCRE supports relative numbers with this syntax (I don't think
1499        Oniguruma does).
1500    
1501    12. Previously, a group with a zero repeat such as (...){0} was completely
1502        omitted from the compiled regex. However, this means that if the group
1503        was called as a subroutine from elsewhere in the pattern, things went wrong
1504        (an internal error was given). Such groups are now left in the compiled
1505        pattern, with a new opcode that causes them to be skipped at execution
1506        time.
1507    
1508    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1509        to the way PCRE behaves:
1510    
1511        (a) A lone ] character is dis-allowed (Perl treats it as data).
1512    
1513        (b) A back reference to an unmatched subpattern matches an empty string
1514            (Perl fails the current match path).
1515    
1516        (c) A data ] in a character class must be notated as \] because if the
1517            first data character in a class is ], it defines an empty class. (In
1518            Perl it is not possible to have an empty class.) The empty class []
1519            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1520            The negative empty class [^] matches any one character, independently
1521            of the DOTALL setting.
1522    
1523    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1524        non-existent subpattern following a character class starting with ']' and
1525        containing () gave an internal compiling error instead of "reference to
1526        non-existent subpattern". Fortunately, when the pattern did exist, the
1527        compiled code was correct. (When scanning forwards to check for the
1528        existencd of the subpattern, it was treating the data ']' as terminating
1529        the class, so got the count wrong. When actually compiling, the reference
1530        was subsequently set up correctly.)
1531    
1532    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1533        it was being rejected as not supported by pcre_dfa_exec(), even though
1534        other assertions are supported. I have made pcre_dfa_exec() support
1535        (*FAIL).
1536    
1537    16. The implementation of 13c above involved the invention of a new opcode,
1538        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1539        cannot be changed at match time, I realized I could make a small
1540        improvement to matching performance by compiling OP_ALLANY instead of
1541        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1542        on the OP_ANY path.
1543    
1544    17. Compiling pcretest on Windows with readline support failed without the
1545        following two fixes: (1) Make the unistd.h include conditional on
1546        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1547    
1548    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1549        ncurses library to be included for pcretest when ReadLine support is
1550        requested, but also to allow for it to be overridden. This patch came from
1551        Daniel Bergström.
1552    
1553    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1554        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1555        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1556        spotting this.
1557    
1558    
1559    Version 7.6 28-Jan-08
1560    ---------------------
1561    
1562    1.  A character class containing a very large number of characters with
1563        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1564        overflow.
1565    
1566    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1567        HAVE_LONG_LONG is not defined.
1568    
1569    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1570        bring them up to date and include new features. This patch includes:
1571    
1572        - Fixed PH's badly added libz and libbz2 support.
1573        - Fixed a problem with static linking.
1574        - Added pcredemo. [But later removed - see 7 below.]
1575        - Fixed dftables problem and added an option.
1576        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1577            HAVE_LONG_LONG.
1578        - Added readline support for pcretest.
1579        - Added an listing of the option settings after cmake has run.
1580    
1581    4.  A user submitted a patch to Makefile that makes it easy to create
1582        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1583        Makefile.am that cause it to include this special target, without
1584        affecting anything else. Note that the same mingw target plus all
1585        the other distribution libraries and programs are now supported
1586        when configuring with CMake (see 6 below) instead of with
1587        Configure/Make.
1588    
1589    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1590        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1591        exported in the Windows port". It has not yet been confirmed that the patch
1592        solves the problem, but it does no harm.
1593    
1594    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1595        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1596        with CMake, and also correct the comment about stack recursion.
1597    
1598    7.  Remove the automatic building of pcredemo from the ./configure system and
1599        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1600        of a program that users should build themselves after PCRE is installed, so
1601        building it automatically is not really right. What is more, it gave
1602        trouble in some build environments.
1603    
1604    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1605    
1606    
1607    Version 7.5 10-Jan-08
1608    ---------------------
1609    
1610    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1611        values in parens when parsing an RE using the C++ wrapper."
1612    
1613    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1614        Characters greater than 255 were excluded from the class instead of being
1615        included.
1616    
1617    3.  The same bug as (2) above applied to negated POSIX classes such as
1618        [:^space:].
1619    
1620    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1621        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1622        I have changed it.
1623    
1624    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1625        first named subpattern) and a construct such as (?&a) would reference the
1626        first named subpattern whose name started with "a" (in other words, the
1627        length check was missing). Both these problems are fixed. "Subpattern name
1628        expected" is now given for (?&) (a zero-length name), and this patch also
1629        makes it give the same error for \k'' (previously it complained that that
1630        was a reference to a non-existent subpattern).
1631    
1632    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1633        this is right because (?- can be followed by option settings as well as by
1634        digits. I have, however, made the messages clearer.
1635    
1636    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1637        than the number used in the conditional) now cause a compile-time error.
1638        This is actually not compatible with Perl, which accepts such patterns, but
1639        treats the conditional as always being FALSE (as PCRE used to), but it
1640        seems to me that giving a diagnostic is better.
1641    
1642    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1643        and messages.
1644    
1645    9.  Fix two occurrences of "backslash" in comments that should have been
1646        "backspace".
1647    
1648    10. Remove two redundant lines of code that can never be obeyed (their function
1649        was moved elsewhere).
1650    
1651    11. The program that makes PCRE's Unicode character property table had a bug
1652        which caused it to generate incorrect table entries for sequences of
1653        characters that have the same character type, but are in different scripts.
1654        It amalgamated them into a single range, with the script of the first of
1655        them. In other words, some characters were in the wrong script. There were
1656        thirteen such cases, affecting characters in the following ranges:
1657    
1658          U+002b0 - U+002c1
1659          U+0060c - U+0060d
1660          U+0061e - U+00612
1661          U+0064b - U+0065e
1662          U+0074d - U+0076d
1663          U+01800 - U+01805
1664          U+01d00 - U+01d77
1665          U+01d9b - U+01dbf
1666          U+0200b - U+0200f
1667          U+030fc - U+030fe
1668          U+03260 - U+0327f
1669          U+0fb46 - U+0fbb1
1670          U+10450 - U+1049d
1671    
1672    12. The -o option (show only the matching part of a line) for pcregrep was not
1673        compatible with GNU grep in that, if there was more than one match in a
1674        line, it showed only the first of them. It now behaves in the same way as
1675        GNU grep.
1676    
1677    13. If the -o and -v options were combined for pcregrep, it printed a blank
1678        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1679        does the same. The return code can be used to tell if there were any
1680        non-matching lines.
1681    
1682    14. Added --file-offsets and --line-offsets to pcregrep.
1683    
1684    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1685        infinitely looping recursion. The bug was that positive lookaheads were not
1686        being skipped when checking for a possible empty match (negative lookaheads
1687        and both kinds of lookbehind were skipped).
1688    
1689    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1690        inclusion of <windows.h> to before rather than after the definition of
1691        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1692    
1693    17. Specifying a possessive quantifier with a specific limit for a Unicode
1694        character property caused pcre_compile() to compile bad code, which led at
1695        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1696        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1697        caused the error; without that there was no problem.
1698    
1699    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1700    
1701    19. Added --enable-pcretest-libreadline.
1702    
1703    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1704        RE::GlobalReplace(). As a result, the number of replacements returned was
1705        double what it should be. I removed one of the increments, but Craig sent a
1706        later patch that removed the other one (the right fix) and added unit tests
1707        that check the return values (which was not done before).
1708    
1709    21. Several CMake things:
1710    
1711        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1712            the names libpcre and libpcreposix, not just pcre and pcreposix.
1713    
1714        (2) The above change means that pcretest and pcregrep are now correctly
1715            linked with the newly-built libraries, not previously installed ones.
1716    
1717        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1718    
1719    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1720        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1721        UTF-8 newline character). The key issue is that the pattern starts .*;
1722        this means that the match must be either at the beginning, or after a
1723        newline. The bug was in the code for advancing after a failed match and
1724        checking that the new position followed a newline. It was not taking
1725        account of UTF-8 characters correctly.
1726    
1727    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1728        character classes. PCRE was not treating the sequence [:...:] as a
1729        character class unless the ... were all letters. Perl, however, seems to
1730        allow any characters between [: and :], though of course it rejects as
1731        unknown any "names" that contain non-letters, because all the known class
1732        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1733        for example, whereas PCRE did not - it did not recognize a POSIX character
1734        class. This seemed a bit dangerous, so the code has been changed to be
1735        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1736        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1737        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1738        Perl does, and where it didn't before.
1739    
1740    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1741        Windows environments %n is disabled by default.
1742    
1743    
1744    Version 7.4 21-Sep-07
1745    ---------------------
1746    
1747    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1748        means that a class such as [\s] counted as "explicit reference to CR or
1749        LF". That isn't really right - the whole point of the change was to try to
1750        help when there was an actual mention of one of the two characters. So now
1751        the change happens only if \r or \n (or a literal CR or LF) character is
1752        encountered.
1753    
1754    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1755        of both had grown to the point where there were only 3 bits left.
1756        Fortunately, there was spare space in the data structure, and so I have
1757        moved the internal flags into a new 16-bit field to free up more option
1758        bits.
1759    
1760    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1761        but did not set the internal JCHANGED flag - either of these is enough to
1762        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1763        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1764        start sets both bits.
1765    
1766    4.  Added options (at build time, compile time, exec time) to change \R from
1767        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1768    
1769    5.  doc/pcresyntax.html was missing from the distribution.
1770    
1771    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1772        compatibility, even though it is no longer used.
1773    
1774    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1775        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1776        windows.h file is present (where different names are used). [This was
1777        reversed later after testing - see 16 below.]
1778    
1779    8.  Changed all #include <config.h> to #include "config.h". There were also
1780        some further <pcre.h> cases that I changed to "pcre.h".
1781    
1782    9.  When pcregrep was used with the --colour option, it missed the line ending
1783        sequence off the lines that it output.
1784    
1785    10. It was pointed out to me that arrays of string pointers cause lots of
1786        relocations when a shared library is dynamically loaded. A technique of
1787        using a single long string with a table of offsets can drastically reduce
1788        these. I have refactored PCRE in four places to do this. The result is
1789        dramatic:
1790    
1791          Originally:                          290
1792          After changing UCP table:            187
1793          After changing error message table:   43
1794          After changing table of "verbs"       36
1795          After changing table of Posix names   22
1796    
1797        Thanks to the folks working on Gregex for glib for this insight.
1798    
1799    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1800        unicode-properties was also set.
1801    
1802    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1803    
1804    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1805        checked only for CRLF.
1806    
1807    14. Added casts to pcretest.c to avoid compiler warnings.
1808    
1809    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1810    
1811    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1812        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1813        entirely. This removes changes made in 7 above.
1814    
1815    17. The CMake files have been updated, and there is now more information about
1816        building with CMake in the NON-UNIX-USE document.
1817    
1818    
1819    Version 7.3 28-Aug-07
1820    ---------------------
1821    
1822     1. In the rejigging of the build system that eventually resulted in 7.1, the
1823        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1824        brackets there is not right, since it causes compilers to look for an
1825        installed pcre.h, not the version that is in the source that is being
1826        compiled (which of course may be different). I have changed it back to:
1827    
1828          #include "pcre.h"
1829    
1830        I have a vague recollection that the change was concerned with compiling in
1831        different directories, but in the new build system, that is taken care of
1832        by the VPATH setting the Makefile.
1833    
1834     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1835        when the subject happened to end in the byte 0x85 (e.g. if the last
1836        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1837        characters but of course it shouldn't be taken as a newline when it is part
1838        of another character. The bug was that, for an unlimited repeat of . in
1839        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1840        characters when looking for a newline.
1841    
1842     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1843    
1844     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1845        in debug output.
1846    
1847     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1848        long printing in the pcrecpp unittest when running under MinGW.
1849    
1850     6. ESC_K was left out of the EBCDIC table.
1851    
1852     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1853        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1854        limit also applies to "virtual nesting" when a pattern is recursive, and in
1855        this case 1000 isn't so big. I have been able to remove this limit at the
1856        expense of backing off one optimization in certain circumstances. Normally,
1857        when pcre_exec() would call its internal match() function recursively and
1858        immediately return the result unconditionally, it uses a "tail recursion"
1859        feature to save stack. However, when a subpattern that can match an empty
1860        string has an unlimited repetition quantifier, it no longer makes this
1861        optimization. That gives it a stack frame in which to save the data for
1862        checking that an empty string has been matched. Previously this was taken
1863        from the 1000-entry workspace that had been reserved. So now there is no
1864        explicit limit, but more stack is used.
1865    
1866     8. Applied Daniel's patches to solve problems with the import/export magic
1867        syntax that is required for Windows, and which was going wrong for the
1868        pcreposix and pcrecpp parts of the library. These were overlooked when this
1869        problem was solved for the main library.
1870    
1871     9. There were some crude static tests to avoid integer overflow when computing
1872        the size of patterns that contain repeated groups with explicit upper
1873        limits. As the maximum quantifier is 65535, the maximum group length was
1874        set at 30,000 so that the product of these two numbers did not overflow a
1875        32-bit integer. However, it turns out that people want to use groups that
1876        are longer than 30,000 bytes (though not repeat them that many times).
1877        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1878        made it possible to implement the integer overflow checks in a much more
1879        dynamic way, which I have now done. The artificial limitation on group
1880        length has been removed - we now have only the limit on the total length of
1881        the compiled pattern, which depends on the LINK_SIZE setting.
1882    
1883    10. Fixed a bug in the documentation for get/copy named substring when
1884        duplicate names are permitted. If none of the named substrings are set, the
1885        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1886        empty string.
1887    
1888    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1889        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1890        because the ] is interpreted as the first data character and the
1891        terminating ] is not found. PCRE has been made compatible with Perl in this
1892        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1893        cause memory overwriting.
1894    
1895    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1896        string has been matched (to stop an infinite loop). It was not recognizing
1897        a conditional subpattern that could match an empty string if that
1898        subpattern was within another subpattern. For example, it looped when
1899        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1900        condition was not nested. This bug has been fixed.
1901    
1902    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1903        past the start of the subject in the presence of bytes with the top bit
1904        set, for example "\x8aBCD".
1905    
1906    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1907        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1908    
1909    14. Optimized (?!) to (*FAIL).
1910    
1911    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1912        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1913        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1914        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1915        does: it's just the validity check that is more restrictive.
1916    
1917    16. Inserted checks for integer overflows during escape sequence (backslash)
1918        processing, and also fixed erroneous offset values for syntax errors during
1919        backslash processing.
1920    
1921    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1922        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1923    
1924    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1925        caused an overrun.
1926    
1927    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1928        something other than just ASCII characters) inside a group that had an
1929        unlimited repeat caused a loop at compile time (while checking to see
1930        whether the group could match an empty string).
1931    
1932    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1933        [\P{Any}] did so. (Error in the code for printing property names.)
1934    
1935    21. An orphan \E inside a character class could cause a crash.
1936    
1937    22. A repeated capturing bracket such as (A)? could cause a wild memory
1938        reference during compilation.
1939    
1940    23. There are several functions in pcre_compile() that scan along a compiled
1941        expression for various reasons (e.g. to see if it's fixed length for look
1942        behind). There were bugs in these functions when a repeated \p or \P was
1943        present in the pattern. These operators have additional parameters compared
1944        with \d, etc, and these were not being taken into account when moving along
1945        the compiled data. Specifically:
1946    
1947        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1948            length.
1949    
1950        (b) An item such as \pL+ within a repeated group could cause crashes or
1951            loops.
1952    
1953        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1954            "reference to non-existent subpattern" error.
1955    
1956        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1957    
1958    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1959        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1960    
1961    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1962    
1963    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1964        character were causing crashes (broken optimization).
1965    
1966    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1967        \p or \P) caused a compile-time loop.
1968    
1969    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1970        break. For example, the unstudied pattern [\r\n]A does not match the string
1971        "\r\nA" because change 7.0/46 below moves the current point on by two
1972        characters after failing to match at the start. However, the pattern \nA
1973        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1974        the same is true. There doesn't seem any very clean way out of this, but
1975        what I have chosen to do makes the common cases work: PCRE now takes note
1976        of whether there can be an explicit match for \r or \n anywhere in the
1977        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1978        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1979        pattern has explicit CR or LF references.
1980    
1981    29. Added (*CR) etc for changing newline setting at start of pattern.
1982    
1983    
1984    Version 7.2 19-Jun-07
1985    ---------------------
1986    
1987     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1988        which is apparently normally available under Windows.
1989    
1990     2. Re-jig the pcregrep tests with different newline settings in an attempt
1991        to make them independent of the local environment's newline setting.
1992    
1993     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1994    
1995     4. Some of the "internals" tests were previously cut out when the link size
1996        was not 2, because the output contained actual offsets. The recent new
1997        "Z" feature of pcretest means that these can be cut out, making the tests
1998        usable with all link sizes.
1999    
2000     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2001        stack recursion. This gives a massive performance boost under BSD, but just
2002        a small improvement under Linux. However, it saves one field in the frame
2003        in all cases.
2004    
2005     6. Added more features from the forthcoming Perl 5.10:
2006    
2007        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2008            recursion call. It refers to the nth most recently opened parentheses.
2009    
2010        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2011            to be opened parentheses.
2012    
2013        (c) Conditions that refer to capturing parentheses can be specified
2014            relatively, for example, (?(-2)... or (?(+3)...
2015    
2016        (d) \K resets the start of the current match so that everything before
2017            is not part of it.
2018    
2019        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2020    
2021        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2022            reference syntax.
2023    
2024        (g) (?| introduces a group in which the numbering of parentheses in each
2025            alternative starts with the same number.
2026    
2027        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2028    
2029     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2030        PCRE_INFO_JCHANGED.
2031    
2032     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2033        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2034        for detecting groups that can match an empty string.
2035    
2036     9. A pattern with a very large number of alternatives (more than several
2037        hundred) was running out of internal workspace during the pre-compile
2038        phase, where pcre_compile() figures out how much memory will be needed. A
2039        bit of new cunning has reduced the workspace needed for groups with
2040        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2041        workspace instead of running out of the 4096 that are available.
2042    
2043    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2044    
2045    11. Applied patch from Google to remove an optimization that didn't quite work.
2046        The report of the bug said:
2047    
2048          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2049          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2050          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2051    
2052    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2053        it matched the wrong number of bytes.
2054    
2055    
2056    Version 7.1 24-Apr-07
2057    ---------------------
2058    
2059     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2060        that is more "standard", making use of automake and other Autotools. There
2061        is some re-arrangement of the files and adjustment of comments consequent
2062        on this.
2063    
2064     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2065        for recursive directory scanning broke on some systems because the files
2066        are not scanned in any specific order and on different systems the order
2067        was different. A call to "sort" has been inserted into RunGrepTest for the
2068        approprate test as a short-term fix. In the longer term there may be an
2069        alternative.
2070    
2071     3. I had an email from Eric Raymond about problems translating some of PCRE's
2072        man pages to HTML (despite the fact that I distribute HTML pages, some
2073        people do their own conversions for various reasons). The problems
2074        concerned the use of low-level troff macros .br and .in. I have therefore
2075        removed all such uses from the man pages (some were redundant, some could
2076        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2077        HTML has been updated to handle .nf/.fi and to complain if it encounters
2078        .br or .in.
2079    
2080     4. Updated comments in configure.ac that get placed in config.h.in and also
2081        arranged for config.h to be included in the distribution, with the name
2082        config.h.generic, for the benefit of those who have to compile without
2083        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2084    
2085     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2086        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2087        makevp.bat for the current PCRE, using the additional files
2088        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2089    
2090     6. A Windows user reported a minor discrepancy with test 2, which turned out
2091        to be caused by a trailing space on an input line that had got lost in his
2092        copy. The trailing space was an accident, so I've just removed it.
2093    
2094     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2095        that is needed.
2096    
2097     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2098        as "const" (a) because they are and (b) because it helps the PHP
2099        maintainers who have recently made a script to detect big data structures
2100        in the php code that should be moved to the .rodata section. I remembered
2101        to update Builducptable as well, so it won't revert if ucptable.h is ever
2102        re-created.
2103    
2104     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2105        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2106        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2107        support is not required. This saves 1.5-2K of code, which is important in
2108        some applications.
2109    
2110        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2111        so as not to refer to the tables, even though these functions will never be
2112        called when UTF-8 support is disabled. Otherwise there are problems with a
2113        shared library.
2114    
2115    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2116    
2117        (a) It was defining its arguments as char * instead of void *.
2118    
2119        (b) It was assuming that all moves were upwards in memory; this was true
2120            a long time ago when I wrote it, but is no longer the case.
2121    
2122        The emulated memove() is provided for those environments that have neither
2123        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2124        is clearly not the case, as these two bugs were recently reported.
2125    
2126    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2127        and Detrail to create the HTML documentation, the .txt form of the man
2128        pages, and it removes trailing spaces from listed files. It also creates
2129        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2130        case, it wraps all the #defines with #ifndefs. This script should be run
2131        before "make dist".
2132    
2133    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2134        with Unicode property support.
2135    
2136        (a) For a maximizing quantifier, if the two different cases of the
2137            character were of different lengths in their UTF-8 codings (there are
2138            some cases like this - I found 11), and the matching function had to
2139            back up over a mixture of the two cases, it incorrectly assumed they
2140            were both the same length.
2141    
2142        (b) When PCRE was configured to use the heap rather than the stack for
2143            recursion during matching, it was not correctly preserving the data for
2144            the other case of a UTF-8 character when checking ahead for a match
2145            while processing a minimizing repeat. If the check also involved
2146            matching a wide character, but failed, corruption could cause an
2147            erroneous result when trying to check for a repeat of the original
2148            character.
2149    
2150    13. Some tidying changes to the testing mechanism:
2151    
2152        (a) The RunTest script now detects the internal link size and whether there
2153            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2154            values substituted by "configure". (The RunGrepTest script already did
2155            this for UTF-8.) The configure.ac script no longer substitutes the
2156            relevant variables.
2157    
2158        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2159            with length and offset values. This means that the output is different
2160            for different internal link sizes. Test 2 is skipped for link sizes
2161            other than 2 because of this, bypassing the problem. Unfortunately,
2162            there was also a test in test 3 (the locale tests) that used /B and
2163            failed for link sizes other than 2. Rather than cut the whole test out,
2164            I have added a new /Z option to pcretest that replaces the length and
2165            offset values with spaces. This is now used to make test 3 independent
2166            of link size. (Test 2 will be tidied up later.)
2167    
2168    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2169        segmentation fault instead of returning the appropriate error message.
2170    
2171    15. In multiline mode when the newline sequence was set to "any", the pattern
2172        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2173        This doesn't seem right; it now treats the CRLF combination as the line
2174        ending, and so does not match in that case. It's only a pattern such as ^$
2175        that would hit this one: something like ^ABC$ would have failed after \r
2176        and then tried again after \r\n.
2177    
2178    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2179        in an attempt to make files that differ only in their line terminators
2180        compare equal. This works on Linux.
2181    
2182    17. Under certain error circumstances pcregrep might try to free random memory
2183        as it exited. This is now fixed, thanks to valgrind.
2184    
2185    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2186        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2187        was because its rules for how to advance for /g after matching an empty
2188        string at the end of a line did not allow for this case. They now check for
2189        it specially.
2190    
2191    20. pcretest is supposed to handle patterns and data of any length, by
2192        extending its buffers when necessary. It was getting this wrong when the
2193        buffer for a data line had to be extended.
2194    
2195    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2196        CRLF as a newline sequence.
2197    
2198    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2199        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2200        I have nevertheless tidied it up.
2201    
2202    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2203    
2204    24. Added a man page for pcre-config.
2205    
2206    
2207    Version 7.0 19-Dec-06
2208    ---------------------
2209    
2210     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2211        moving to gcc 4.1.1.
2212    
2213     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2214        sys/time.h, which is documented as needed for this function. It doesn't
2215        seem to matter on Linux, but it showed up on some releases of OS X.
2216    
2217     3. It seems that there are systems where bytes whose values are greater than
2218        127 match isprint() in the "C" locale. The "C" locale should be the
2219        default when a C program starts up. In most systems, only ASCII printing
2220        characters match isprint(). This difference caused the output from pcretest
2221        to vary, making some of the tests fail. I have changed pcretest so that:
2222    
2223        (a) When it is outputting text in the compiled version of a pattern, bytes
2224            other than 32-126 are always shown as hex escapes.
2225    
2226        (b) When it is outputting text that is a matched part of a subject string,
2227            it does the same, unless a different locale has been set for the match
2228            (using the /L modifier). In this case, it uses isprint() to decide.
2229    
2230     4. Fixed a major bug that caused incorrect computation of the amount of memory
2231        required for a compiled pattern when options that changed within the
2232        pattern affected the logic of the preliminary scan that determines the
2233        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2234        that the computed length was too small. The symptoms of this bug were
2235        either the PCRE error "internal error: code overflow" from pcre_compile(),
2236        or a glibc crash with a message such as "pcretest: free(): invalid next
2237        size (fast)". Examples of patterns that provoked this bug (shown in
2238        pcretest format) are:
2239    
2240          /(?-x: )/x
2241          /(?x)(?-x: \s*#\s*)/
2242          /((?i)[\x{c0}])/8
2243          /(?i:[\x{c0}])/8
2244    
2245        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2246        is now done differently.
2247    
2248     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2249        wrapper classes; (b) implement a new function in the C++ scanner that is
2250        more efficient than the old way of doing things because it avoids levels of
2251        recursion in the regex matching; (c) add a paragraph to the documentation
2252        for the FullMatch() function.
2253    
2254     6. The escape sequence \n was being treated as whatever was defined as
2255        "newline". Not only was this contrary to the documentation, which states
2256        that \n is character 10 (hex 0A), but it also went horribly wrong when
2257        "newline" was defined as CRLF. This has been fixed.
2258    
2259     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2260        was being set to -1 for the "end of line" case (supposedly a value that no
2261        character can have). Though this value is never used (the check for end of
2262        line is "zero bytes in current character"), it caused compiler complaints.
2263        I've changed it to 0xffffffff.
2264    
2265     8. In pcre_version.c, the version string was being built by a sequence of
2266        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2267        string (as it is for production releases) called a macro with an empty
2268        argument. The C standard says the result of this is undefined. The gcc
2269        compiler treats it as an empty string (which was what was wanted) but it is
2270        reported that Visual C gives an error. The source has been hacked around to
2271        avoid this problem.
2272    
2273     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2274        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2275        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2276        of them did).
2277    
2278    10. Originally, pcretest opened its input and output without "b"; then I was
2279        told that "b" was needed in some environments, so it was added for release
2280        5.0 to both the input and output. (It makes no difference on Unix-like
2281        systems.) Later I was told that it is wrong for the input on Windows. I've
2282        now abstracted the modes into two macros, to make it easier to fiddle with
2283        them, and removed "b" from the input mode under Windows.
2284    
2285    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2286    
2287    12. Added -help and --help to pcretest as an official way of being reminded
2288        of the options.
2289    
2290    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2291        and pcrecpp.cc because they annoy compilers at high warning levels.
2292    
2293    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2294    
2295    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2296        scripts are not C programs :-) and which was not noticed because it works
2297        on Linux.
2298    
2299    16. pcretest is supposed to handle any length of pattern and data line (as one
2300        line or as a continued sequence of lines) by extending its input buffer if
2301        necessary. This feature was broken for very long pattern lines, leading to
2302        a string of junk being passed to pcre_compile() if the pattern was longer
2303        than about 50K.
2304    
2305    17. I have done a major re-factoring of the way pcre_compile() computes the
2306        amount of memory needed for a compiled pattern. Previously, there was code
2307        that made a preliminary scan of the pattern in order to do this. That was
2308        OK when PCRE was new, but as the facilities have expanded, it has become
2309        harder and harder to keep it in step with the real compile phase, and there
2310        have been a number of bugs (see for example, 4 above). I have now found a
2311        cunning way of running the real compile function in a "fake" mode that
2312        enables it to compute how much memory it would need, while actually only
2313        ever using a few hundred bytes of working memory and without too many
2314        tests of the mode. This should make future maintenance and development
2315        easier. A side effect of this work is that the limit of 200 on the nesting
2316        depth of parentheses has been removed (though this was never a serious
2317        limitation, I suspect). However, there is a downside: pcre_compile() now
2318        runs more slowly than before (30% or more, depending on the pattern). I
2319        hope this isn't a big issue. There is no effect on runtime performance.
2320    
2321    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2322        newline (only possible for the last line of a file) and it was a
2323        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2324    
2325    19. Added additional timing features to pcretest. (1) The -tm option now times
2326        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2327        separate command line item, by a number that specifies the number of
2328        repeats to use when timing. The default is 50000; this gives better
2329        precision, but takes uncomfortably long for very large patterns.
2330    
2331    20. Extended pcre_study() to be more clever in cases where a branch of a
2332        subpattern has no definite first character. For example, (a*|b*)[cd] would
2333        previously give no result from pcre_study(). Now it recognizes that the
2334        first character must be a, b, c, or d.
2335    
2336    21. There was an incorrect error "recursive call could loop indefinitely" if
2337        a subpattern (or the entire pattern) that was being tested for matching an
2338        empty string contained only one non-empty item after a nested subpattern.
2339        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2340        incorrectly, because the \d was being skipped in the check.
2341    
2342    22. The pcretest program now has a new pattern option /B and a command line
2343        option -b, which is equivalent to adding /B to every pattern. This causes
2344        it to show the compiled bytecode, without the additional information that
2345        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2346        is the same as /B/I).
2347    
2348    23. A new optimization is now able automatically to treat some sequences such
2349        as a*b as a*+b. More specifically, if something simple (such as a character
2350        or a simple class like \d) has an unlimited quantifier, and is followed by
2351        something that cannot possibly match the quantified thing, the quantifier
2352        is automatically "possessified".
2353    
2354    24. A recursive reference to a subpattern whose number was greater than 39
2355        went wrong under certain circumstances in UTF-8 mode. This bug could also
2356        have affected the operation of pcre_study().
2357    
2358    25. Realized that a little bit of performance could be had by replacing
2359        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2360    
2361    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2362    
2363    27. Possessive quantifiers such as a++ were previously implemented by turning
2364        them into atomic groups such as ($>a+). Now they have their own opcodes,
2365        which improves performance. This includes the automatically created ones
2366        from 23 above.
2367    
2368    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2369        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2370        the first matched character to be a colon. This applied both to named and
2371        numbered groups.
2372    
2373    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2374    
2375    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2376        building PCRE on Windows easier, so I have included it in the distribution.
2377    
2378    31. There is now a check in pcretest against a ridiculously large number being
2379        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2380        loop, the loop is abandoned.
2381    
2382    32. Forward references to subpatterns in conditions such as (?(2)...) where
2383        subpattern 2 is defined later cause pcre_compile() to search forwards in
2384        the pattern for the relevant set of parentheses. This search went wrong
2385        when there were unescaped parentheses in a character class, parentheses
2386        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2387    
2388    33. "Subroutine" calls and backreferences were previously restricted to
2389        referencing subpatterns earlier in the regex. This restriction has now
2390        been removed.
2391    
2392    34. Added a number of extra features that are going to be in Perl 5.10. On the
2393        whole, these are just syntactic alternatives for features that PCRE had
2394        previously implemented using the Python syntax or my own invention. The
2395        other formats are all retained for compatibility.
2396    
2397        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2398            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2399            also .NET compatible.
2400    
2401        (b) A recursion or subroutine call to a named group can now be defined as
2402            (?&name) as well as (?P>name).
2403    
2404        (c) A backreference to a named group can now be defined as \k<name> or
2405            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2406            5.10, are also .NET compatible.
2407    
2408        (d) A conditional reference to a named group can now use the syntax
2409            (?(<name>) or (?('name') as well as (?(name).
2410    
2411        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2412            groups (named and numbered) that are never evaluated inline, but can be
2413            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2414            is always false. There may be only one alternative in such a group.
2415    
2416        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2417            as the simple (?(R). The condition is true only if the most recent
2418            recursion is that of the given number or name. It does not search out
2419            through the entire recursion stack.
2420    
2421        (g) The escape \gN or \g{N} has been added, where N is a positive or
2422            negative number, specifying an absolute or relative reference.
2423    
2424    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2425        some "unreachable code" warnings.
2426    
2427    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2428        things, this adds five new scripts.
2429    
2430    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2431        There were also incompatibilities regarding the handling of \Q..\E inside
2432        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2433        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2434    
2435    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2436        matches an empty string, and forcibly breaks the loop. There were bugs in
2437        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2438        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2439        separate and independent bugs (that affected different cases) have been
2440        fixed.
2441    
2442    39. Refactored the code to abolish the use of different opcodes for small
2443        capturing bracket numbers. This is a tidy that I avoided doing when I
2444        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2445        The new approach is not only tidier, it makes it possible to reduce the
2446        memory needed to fix the previous bug (38).
2447    
2448    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2449        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2450        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2451        mode.
2452    
2453    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2454        report.
2455    
2456    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2457        copy construction and assignment in the C++ wrapper.
2458    
2459    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2460        couple of bugs that could have given wrong results in the "--newline=crlf"
2461        case.
2462    
2463    44. Added a number of casts and did some reorganization of signed/unsigned int
2464        variables following suggestions from Dair Grant. Also renamed the variable
2465        "this" as "item" because it is a C++ keyword.
2466    
2467    45. Arranged for dftables to add
2468    
2469          #include "pcre_internal.h"
2470    
2471        to pcre_chartables.c because without it, gcc 4.x may remove the array
2472        definition from the final binary if PCRE is built into a static library and
2473        dead code stripping is activated.
2474    
2475    46. For an unanchored pattern, if a match attempt fails at the start of a
2476        newline sequence, and the newline setting is CRLF or ANY, and the next two
2477        characters are CRLF, advance by two characters instead of one.
2478    
2479    
2480    Version 6.7 04-Jul-06
2481    ---------------------
2482    
2483     1. In order to handle tests when input lines are enormously long, pcretest has
2484        been re-factored so that it automatically extends its buffers when
2485        necessary. The code is crude, but this _is_ just a test program. The
2486        default size has been increased from 32K to 50K.
2487    
2488     2. The code in pcre_study() was using the value of the re argument before
2489        testing it for NULL. (Of course, in any sensible call of the function, it
2490        won't be NULL.)
2491    
2492     3. The memmove() emulation function in pcre_internal.h, which is used on
2493        systems that lack both memmove() and bcopy() - that is, hardly ever -
2494        was missing a "static" storage class specifier.
2495    
2496     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2497        containing an extended class (one that cannot be represented by a bitmap
2498        because it contains high-valued characters or Unicode property items, e.g.
2499        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2500        pattern, but PCRE should not loop if you do not (it no longer does).
2501        [Detail: two cases were found: (a) a repeated subpattern containing an
2502        extended class; (b) a recursive reference to a subpattern that followed a
2503        previous extended class. It wasn't skipping over the extended class
2504        correctly when UTF-8 mode was not set.]
2505    
2506     5. A negated single-character class was not being recognized as fixed-length
2507        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2508        compile error "lookbehind assertion is not fixed length".
2509    
2510     6. The RunPerlTest auxiliary script was showing an unexpected difference
2511        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2512        write a Perl script that can interpret lines of an input file either as
2513        byte characters or as UTF-8, which is what "perltest" was being required to
2514        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2515        can't do is switch easily at run time between having the "use utf8;" pragma
2516        or not. In the end, I fudged it by using the RunPerlTest script to insert
2517        "use utf8;" explicitly for the UTF-8 tests.
2518    
2519     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2520        the end of the subject string, contrary to the documentation and to what
2521        Perl does. This was true of both matching functions. Now it matches only at
2522        the start of the subject and immediately after *internal* newlines.
2523    
2524     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2525        a pointer to an int instead of a pointer to an unsigned long int. This
2526        caused problems on 64-bit systems.
2527    
2528     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2529        instance of the 'standard' template library not being so standard".
2530    
2531    10. There was no check on the number of named subpatterns nor the maximum
2532        length of a subpattern name. The product of these values is used to compute
2533        the size of the memory block for a compiled pattern. By supplying a very
2534        long subpattern name and a large number of named subpatterns, the size
2535        computation could be caused to overflow. This is now prevented by limiting
2536        the length of names to 32 characters, and the number of named subpatterns
2537        to 10,000.
2538    
2539    11. Subpatterns that are repeated with specific counts have to be replicated in
2540        the compiled pattern. The size of memory for this was computed from the
2541        length of the subpattern and the repeat count. The latter is limited to
2542        65535, but there was no limit on the former, meaning that integer overflow
2543        could in principle occur. The compiled length of a repeated subpattern is
2544        now limited to 30,000 bytes in order to prevent this.
2545    
2546    12. Added the optional facility to have named substrings with the same name.
2547    
2548    13. Added the ability to use a named substring as a condition, using the
2549        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2550        are numbers (not recommended). Forward references are permitted.
2551    
2552    14. Added forward references in named backreferences (if you see what I mean).
2553    
2554    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2555        pattern could run off the end of the subject. For example, the pattern
2556        "(?s)(.{1,5})"8 did this with the subject "ab".
2557    
2558    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2559        PCRE_CASELESS was set when matching characters that were quantified with ?
2560        or *.
2561    
2562    17. A character class other than a single negated character that had a minimum
2563        but no maximum quantifier - for example [ab]{6,} - was not handled
2564        correctly by pce_dfa_exec(). It would match only one character.
2565    
2566    18. A valid (though odd) pattern that looked like a POSIX character
2567        class but used an invalid character after [ (for example [[,abc,]]) caused
2568        pcre_compile() to give the error "Failed: internal error: code overflow" or
2569        in some cases to crash with a glibc free() error. This could even happen if
2570        the pattern terminated after [[ but there just happened to be a sequence of
2571        letters, a binary zero, and a closing ] in the memory that followed.
2572    
2573    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2574        over the years. Originally (before any Unicode support), just the bottom 8
2575        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2576        output from "man perlunicode" includes this:
2577    
2578          The regular expression compiler produces polymorphic opcodes.  That
2579          is, the pattern adapts to the data and automatically switches to
2580          the Unicode character scheme when presented with Unicode data--or
2581          instead uses a traditional byte scheme when presented with byte
2582          data.
2583    
2584        Sadly, a wide octal escape does not cause a switch, and in a string with
2585        no other multibyte characters, these octal escapes are treated as before.
2586        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2587        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2588        Unicode string.
2589    
2590        I have not perpetrated such confusion in PCRE. Up till now, it took just
2591        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2592        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2593        translate to the appropriate multibyte character.
2594    
2595    29. Applied some refactoring to reduce the number of warnings from Microsoft
2596        and Borland compilers. This has included removing the fudge introduced
2597        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2598        a warning about an unused variable.
2599    
2600    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2601        characters since release 4.0, because Perl (from release 5.004) does not.
2602        [Or at least, is documented not to: some releases seem to be in conflict
2603        with the documentation.] However, when a pattern was studied with
2604        pcre_study() and all its branches started with \s, PCRE still included VT
2605        as a possible starting character. Of course, this did no harm; it just
2606        caused an unnecessary match attempt.
2607    
2608    22. Removed a now-redundant internal flag bit that recorded the fact that case
2609        dependency changed within the pattern. This was once needed for "required
2610        byte" processing, but is no longer used. This recovers a now-scarce options
2611        bit. Also moved the least significant internal flag bit to the most-
2612        significant bit of the word, which was not previously used (hangover from
2613        the days when it was an int rather than a uint) to free up another bit for
2614        the future.
2615    
2616    23. Added support for CRLF line endings as well as CR and LF. As well as the
2617        default being selectable at build time, it can now be changed at runtime
2618        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2619        specify that it is scanning data with non-default line endings.
2620    
2621    24. Changed the definition of CXXLINK to make it agree with the definition of
2622        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2623    
2624    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2625        recursions. This makes a big different to stack usage for some patterns.
2626    
2627    26. If a subpattern containing a named recursion or subroutine reference such
2628        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2629        the space required for the compiled pattern went wrong and gave too small a
2630        value. Depending on the environment, this could lead to "Failed: internal
2631        error: code overflow at offset 49" or "glibc detected double free or
2632        corruption" errors.
2633    
2634    27. Applied patches from Google (a) to support the new newline modes and (b) to
2635        advance over multibyte UTF-8 characters in GlobalReplace.
2636    
2637    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2638        difference for some implementation of PCRE in some Windows version.
2639    
2640    29. Added some extra testing facilities to pcretest:
2641    
2642        \q<number>   in a data line sets the "match limit" value
2643        \Q<number>   in a data line sets the "match recursion limt" value
2644        -S <number>  sets the stack size, where <number> is in megabytes
2645    
2646        The -S option isn't available for Windows.
2647    
2648    
2649    Version 6.6 06-Feb-06
2650    ---------------------
2651    
2652     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2653        in pcreposix.h. I have copied the definition from pcre.h.
2654    
2655     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2656        because pcre.h is no longer a built file.
2657    
2658     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2659        not normally included in the compiled code.
2660    
2661    
2662    Version 6.5 01-Feb-06
2663    ---------------------
2664    
2665     1. When using the partial match feature with pcre_dfa_exec(), it was not
2666        anchoring the second and subsequent partial matches at the new starting
2667        point. This could lead to incorrect results. For example, with the pattern
2668        /1234/, partially matching against "123" and then "a4" gave a match.
2669    
2670     2. Changes to pcregrep:
2671    
2672        (a) All non-match returns from pcre_exec() were being treated as failures
2673            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2674            error message is output. Some extra information is given for the
2675            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2676            probably the only errors that are likely to be caused by users (by
2677            specifying a regex that has nested indefinite repeats, for instance).
2678            If there are more than 20 of these errors, pcregrep is abandoned.
2679    
2680        (b) A binary zero was treated as data while matching, but terminated the
2681            output line if it was written out. This has been fixed: binary zeroes
2682            are now no different to any other data bytes.
2683    
2684        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2685            used to set a locale for matching. The --locale=xxxx long option has
2686            been added (no short equivalent) to specify a locale explicitly on the
2687            pcregrep command, overriding the environment variables.
2688    
2689        (d) When -B was used with -n, some line numbers in the output were one less
2690            than they should have been.
2691    
2692        (e) Added the -o (--only-matching) option.
2693    
2694        (f) If -A or -C was used with -c (count only), some lines of context were
2695            accidentally printed for the final match.
2696    
2697        (g) Added the -H (--with-filename) option.
2698    
2699        (h) The combination of options -rh failed to suppress file names for files
2700            that were found from directory arguments.
2701    
2702        (i) Added the -D (--devices) and -d (--directories) options.
2703    
2704        (j) Added the -F (--fixed-strings) option.
2705    
2706        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2707    
2708        (l) Added the --colo(u)r option.
2709    
2710        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2711            is not present by default.
2712    
2713     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2714        items such as (?R) or (?1), when the recursion could match a number of
2715        alternatives. If it matched one of the alternatives, but subsequently,
2716        outside the recursion, there was a failure, the code tried to back up into
2717        the recursion. However, because of the way PCRE is implemented, this is not
2718        possible, and the result was an incorrect result from the match.
2719    
2720        In order to prevent this happening, the specification of recursion has
2721        been changed so that all such subpatterns are automatically treated as
2722        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2723    
2724     4. I had overlooked the fact that, in some locales, there are characters for
2725        which isalpha() is true but neither isupper() nor islower() are true. In
2726        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2727        and ordfeminine) are like this. This affected the treatment of \w and \W
2728        when they appeared in character classes, but not when they appeared outside
2729        a character class. The bit map for "word" characters is now created
2730        separately from the results of isalnum() instead of just taking it from the
2731        upper, lower, and digit maps. (Plus the underscore character, of course.)
2732    
2733     5. The above bug also affected the handling of POSIX character classes such as
2734        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2735        permanent tables. Instead, the bit maps for such a class were previously
2736        created as the appropriate unions of the upper, lower, and digit bitmaps.
2737        Now they are created by subtraction from the [[:word:]] class, which has
2738        its own bitmap.
2739    
2740     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2741        It is created by subtracting the vertical space characters (\x09, \x0a,
2742        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2743        subtraction was done in the overall bitmap for a character class, meaning
2744        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2745        be recognized. This bug has been fixed.
2746    
2747     7. Patches from the folks at Google:
2748    
2749          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2750          real life, but is still worth protecting against".
2751    
2752          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2753          regular expressions".
2754    
2755          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2756          have it.
2757    
2758          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2759          "configure" and the latter not, in order to fix a problem somebody had
2760          with compiling the Arg class on HP-UX.
2761    
2762          (e) Improve the error-handling of the C++ wrapper a little bit.
2763    
2764          (f) New tests for checking recursion limiting.
2765    
2766     8. The pcre_memmove() function, which is used only if the environment does not
2767        have a standard memmove() function (and is therefore rarely compiled),
2768        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2769        returning a result (though PCRE never actually uses the result).
2770    
2771     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2772        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2773        returned instead of calling malloc() with an overflowing number that would
2774        most likely cause subsequent chaos.
2775    
2776    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2777    
2778    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2779        with this option is matched, the nmatch and pmatch options of regexec() are
2780        ignored.
2781    
2782    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2783        provided in case anyone wants to the the POSIX interface with UTF-8
2784        strings.
2785    
2786    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2787        C++ linking (needed for some HP-UX environments).
2788    
2789    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2790        (unused parameter) and in the pcre_printint() function (omitted "default"
2791        switch label when the default is to do nothing).
2792    
2793    15. Added some code to make it possible, when PCRE is compiled as a C++
2794        library, to replace subject pointers for pcre_exec() with a smart pointer
2795        class, thus making it possible to process discontinuous strings.
2796    
2797    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2798        much the same function. They were added by different people who were trying
2799        to make PCRE easy to compile on non-Unix systems. It has been suggested
2800        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2801        for compiling on Windows systems. I have therefore replaced it with
2802        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2803        defaults to "extern" for C or "extern C" for C++, which works fine on
2804        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2805        SCOPE with something explicit in config.h. In addition:
2806    
2807        (a) pcreposix.h still had just "extern" instead of either of these macros;
2808            I have replaced it with PCRE_DATA_SCOPE.
2809    
2810        (b) Functions such as _pcre_xclass(), which are internal to the library,
2811            but external in the C sense, all had PCRE_EXPORT in their definitions.
2812            This is apparently wrong for the Windows case, so I have removed it.
2813            (It makes no difference on Unix-like systems.)
2814    
2815    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2816        of recursive calls to match(). This is different to MATCH_LIMIT because
2817        that limits the total number of calls to match(), not all of which increase
2818        the depth of recursion. Limiting the recursion depth limits the amount of
2819        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2820        when PCRE is compiled, and changed at run time. A patch from Google adds
2821        this functionality to the C++ interface.
2822    
2823    18. Changes to the handling of Unicode character properties:
2824    
2825        (a) Updated the table to Unicode 4.1.0.
2826    
2827        (b) Recognize characters that are not in the table as "Cn" (undefined).
2828    
2829        (c) I revised the way the table is implemented to a much improved format
2830            which includes recognition of ranges. It now supports the ranges that
2831            are defined in UnicodeData.txt, and it also amalgamates other
2832            characters into ranges. This has reduced the number of entries in the
2833            table from around 16,000 to around 3,000, thus reducing its size
2834            considerably. I realized I did not need to use a tree structure after
2835            all - a binary chop search is just as efficient. Having reduced the
2836            number of entries, I extended their size from 6 bytes to 8 bytes to
2837            allow for more data.
2838    
2839        (d) Added support for Unicode script names via properties such as \p{Han}.
2840    
2841    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2842        matching that character.
2843    
2844    20. When matching a repeated Unicode property with a minimum greater than zero,
2845        (for example \pL{2,}), PCRE could look past the end of the subject if it
2846        reached it while seeking the minimum number of characters. This could
2847        happen only if some of the characters were more than one byte long, because
2848        there is a check for at least the minimum number of bytes.
2849    
2850    21. Refactored the implementation of \p and \P so as to be more general, to
2851        allow for more different types of property in future. This has changed the
2852        compiled form incompatibly. Anybody with saved compiled patterns that use
2853        \p or \P will have to recompile them.
2854    
2855    22. Added "Any" and "L&" to the supported property types.
2856    
2857    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2858        but give a compile time error if the value is greater than 0xff.
2859    
2860    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2861        accidentally not being installed or uninstalled.
2862    
2863    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2864        made were to insert the current release number. This seemed silly, because
2865        it made things harder for people building PCRE on systems that don't run
2866        "configure". I have turned pcre.h into a distributed file, no longer built
2867        by "configure", with the version identification directly included. There is
2868        no longer a pcre.h.in file.
2869    
2870        However, this change necessitated a change to the pcre-config script as
2871        well. It is built from pcre-config.in, and one of the substitutions was the
2872        release number. I have updated configure.ac so that ./configure now finds
2873        the release number by grepping pcre.h.
2874    
2875    26. Added the ability to run the tests under valgrind.
2876    
2877    
2878    Version 6.4 05-Sep-05
2879    ---------------------
2880    
2881     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2882        "--" to be printed when multiple files were scanned, even when none of the
2883        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2884        consider it to be a bug, and have restored the previous behaviour.
2885    
2886     2. A couple of code tidies to get rid of compiler warnings.
2887    
2888     3. The pcretest program used to cheat by referring to symbols in the library
2889        whose names begin with _pcre_. These are internal symbols that are not
2890        really supposed to be visible externally, and in some environments it is
2891        possible to suppress them. The cheating is now confined to including
2892        certain files from the library's source, which is a bit cleaner.
2893    
2894     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2895        file's purpose clearer.
2896    
2897     5. Reorganized pcre_ucp_findchar().
2898    
2899    
2900    Version 6.3 15-Aug-05
2901    ---------------------
2902    
2903     1. The file libpcre.pc.in did not have general read permission in the tarball.
2904    
2905     2. There were some problems when building without C++ support:
2906    
2907        (a) If C++ support was not built, "make install" and "make test" still
2908            tried to test it.
2909    
2910        (b) There were problems when the value of CXX was explicitly set. Some
2911            changes have been made to try to fix these, and ...
2912    
2913        (c) --disable-cpp can now be used to explicitly disable C++ support.
2914    
2915        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2916            backslash in a target when C++ was disabled. This confuses some
2917            versions of "make", apparently. Using an intermediate variable solves
2918            this. (Same for CPP_LOBJ.)
2919    
2920     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2921        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2922        necessary on certain architectures.
2923    
2924     4. Added a setting of -export-symbols-regex to the link command to remove
2925        those symbols that are exported in the C sense, but actually are local
2926        within the library, and not documented. Their names all begin with
2927        "_pcre_". This is not a perfect job, because (a) we have to except some
2928        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2929        available (and never for static libraries). I have made a note to try to
2930        find a way round (a) in the future.
2931    
2932    
2933    Version 6.2 01-Aug-05
2934    ---------------------
2935    
2936     1. There was no test for integer overflow of quantifier values. A construction
2937        such as {1111111111111111} would give undefined results. What is worse, if
2938        a minimum quantifier for a parenthesized subpattern overflowed and became
2939        negative, the calculation of the memory size went wrong. This could have
2940        led to memory overwriting.
2941    
2942     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2943    
2944     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2945        operating environments where this matters.
2946    
2947     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2948        PCRE options from within the C++ wrapper.
2949    
2950     5. Named capturing subpatterns were not being correctly counted when a pattern
2951        was compiled. This caused two problems: (a) If there were more than 100
2952        such subpatterns, the calculation of the memory needed for the whole
2953        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2954        back references of the form \12, where the number was greater than 9, were
2955        not recognized as back references, even though there were sufficient
2956        previous subpatterns.
2957    
2958     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2959        versions of gcc, e.g. 2.95.4.
2960    
2961    
2962    Version 6.1 21-Jun-05
2963    ---------------------
2964    
2965     1. There was one reference to the variable "posix" in pcretest.c that was not
2966        surrounded by "#if !defined NOPOSIX".
2967    
2968     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2969        the cross-check on the old pcre_info() function, for the benefit of the
2970        cut-down version of PCRE that is currently imported into Exim.
2971    
2972     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2973        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2974        patterns that start (?i) but I don't think that matters. The use of (?i) is
2975        just an example; this all applies to the other options as well.
2976    
2977     4. Since libtool seems to echo the compile commands it is issuing, the output
2978        from "make" can be reduced a bit by putting "@" in front of each libtool
2979        compile command.
2980    
2981     5. Patch from the folks at Google for configure.in to be a bit more thorough
2982        in checking for a suitable C++ installation before trying to compile the
2983        C++ stuff. This should fix a reported problem when a compiler was present,
2984        but no suitable headers.
2985    
2986     6. The man pages all had just "PCRE" as their title. I have changed them to
2987        be the relevant file name. I have also arranged that these names are
2988        retained in the file doc/pcre.txt, which is a concatenation in text format
2989        of all the man pages except the little individual ones for each function.
2990    
2991     7. The NON-UNIX-USE file had not been updated for the different set of source
2992        files that come with release 6. I also added a few comments about the C++
2993        wrapper.
2994    
2995    
2996    Version 6.0 07-Jun-05
2997    ---------------------
2998    
2999     1. Some minor internal re-organization to help with my DFA experiments.
3000    
3001     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3002        didn't matter for the library itself when fully configured, but did matter
3003        when compiling without UCP support, or within Exim, where the ucp files are
3004        not imported.
3005    
3006     3. Refactoring of the library code to split up the various functions into
3007        different source modules. The addition of the new DFA matching code (see
3008        below) to a single monolithic source would have made it really too
3009        unwieldy, quite apart from causing all the code to be include in a
3010        statically linked application, when only some functions are used. This is
3011        relevant even without the DFA addition now that patterns can be compiled in
3012        one application and matched in another.
3013    
3014        The downside of splitting up is that there have to be some external
3015        functions and data tables that are used internally in different modules of
3016        the library but which are not part of the API. These have all had their
3017        names changed to start with "_pcre_" so that they are unlikely to clash
3018        with other external names.
3019    
3020     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3021        a different (DFA) algorithm. Although it is slower than the original
3022        function, it does have some advantages for certain types of matching
3023        problem.
3024    
3025     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3026        including restarting after a partial match.
3027    
3028     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3029        defined when compiling for Windows was sent to me. I have put it into the
3030        code, though I have no means of testing or verifying it.
3031    
3032     7. Added the pcre_refcount() auxiliary function.
3033    
3034     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3035        match before or at the first newline in the subject string. In pcretest,
3036        the /f option on a pattern can be used to set this.
3037    
3038     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3039        would behave wrongly. This has been present in PCRE since release 4.0.
3040    
3041    10. A number of changes to the pcregrep command:
3042    
3043        (a) Refactored how -x works; insert ^(...)$ instead of setting
3044            PCRE_ANCHORED and checking the length, in preparation for adding
3045            something similar for -w.
3046    
3047        (b) Added the -w (match as a word) option.
3048    
3049        (c) Refactored the way lines are read and buffered so as to have more
3050            than one at a time available.
3051    
3052        (d) Implemented a pcregrep test script.
3053    
3054        (e) Added the -M (multiline match) option. This allows patterns to match
3055            over several lines of the subject. The buffering ensures that at least
3056            8K, or the rest of the document (whichever is the shorter) is available
3057            for matching (and similarly the previous 8K for lookbehind assertions).
3058    
3059        (f) Changed the --help output so that it now says
3060    
3061              -w, --word-regex(p)
3062    
3063            instead of two lines, one with "regex" and the other with "regexp"
3064            because that confused at least one person since the short forms are the
3065            same. (This required a bit of code, as the output is generated
3066            automatically from a table. It wasn't just a text change.)
3067    
3068        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3069            option but starts with a hyphen. Could be a pattern or a path name
3070            starting with a hyphen, for instance.
3071    
3072        (h) "-" can be given as a file name to represent stdin.
3073    
3074        (i) When file names are being printed, "(standard input)" is used for
3075            the standard input, for compatibility with GNU grep. Previously
3076            "<stdin>" was used.
3077    
3078        (j) The option --label=xxx can be used to supply a name to be used for
3079            stdin when file names are being printed. There is no short form.
3080    
3081        (k) Re-factored the options decoding logic because we are going to add
3082            two more options that take data. Such options can now be given in four
3083            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3084    
3085        (l) Added the -A, -B, and -C options for requesting that lines of context
3086            around matches be printed.
3087    
3088        (m) Added the -L option to print the names of files that do not contain
3089            any matching lines, that is, the complement of -l.
3090    
3091        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3092            continue to scan other files.
3093    
3094        (o) The -s option was incorrectly implemented. For compatibility with other
3095            greps, it now suppresses the error message for a non-existent or non-
3096            accessible file (but not the return code). There is a new option called
3097            -q that suppresses the output of matching lines, which was what -s was
3098            previously doing.
3099    
3100        (p) Added --include and --exclude options to specify files for inclusion
3101            and exclusion when recursing.
3102    
3103    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3104        Hopefully, it now does.
3105    
3106    12. Missing cast in pcre_study().
3107    
3108    13. Added an "uninstall" target to the makefile.
3109    
3110    14. Replaced "extern" in the function prototypes in Makefile.in with
3111        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3112        world, but is set differently for Windows.
3113    
3114    15. Added a second compiling function called pcre_compile2(). The only
3115        difference is that it has an extra argument, which is a pointer to an
3116        integer error code. When there is a compile-time failure, this is set
3117        non-zero, in addition to the error test pointer being set to point to an
3118        error message. The new argument may be NULL if no error number is required
3119        (but then you may as well call pcre_compile(), which is now just a
3120        wrapper). This facility is provided because some applications need a
3121        numeric error indication, but it has also enabled me to tidy up the way
3122        compile-time errors are handled in the POSIX wrapper.
3123    
3124    16. Added VPATH=.libs to the makefile; this should help when building with one
3125        prefix path and installing with another. (Or so I'm told by someone who
3126        knows more about this stuff than I do.)
3127    
3128    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3129        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3130        match everything, including newlines. This is not POSIX-compatible, but
3131        somebody wanted the feature. From pcretest it can be activated by using
3132        both the P and the s flags.
3133    
3134    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3135    
3136    19. libpcre.pc was being incorrectly installed as executable.
3137    
3138    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3139        it now also looks for '\r' so that it will work unmodified on Windows.
3140    
3141    21. Added Google's contributed C++ wrapper to the distribution.
3142    
3143    22. Added some untidy missing memory free() calls in pcretest, to keep
3144        Electric Fence happy when testing.
3145    
3146    
3147    
3148    Version 5.0 13-Sep-04
3149    ---------------------
3150    
3151     1. Internal change: literal characters are no longer packed up into items
3152        containing multiple characters in a single byte-string. Each character
3153        is now matched using a separate opcode. However, there may be more than one
3154        byte in the character in UTF-8 mode.
3155    
3156     2. The pcre_callout_block structure has two new fields: pattern_position and
3157        next_item_length. These contain the offset in the pattern to the next match
3158        item, and its length, respectively.
3159    
3160     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3161        insertion of callouts before each pattern item. Added the /C option to
3162        pcretest to make use of this.
3163    
3164     4. On the advice of a Windows user, the lines
3165    
3166          #if defined(_WIN32) || defined(WIN32)
3167          _setmode( _fileno( stdout ), 0x8000 );
3168          #endif  /* defined(_WIN32) || defined(WIN32) */
3169    
3170        have been added to the source of pcretest. This apparently does useful
3171        magic in relation to line terminators.
3172    
3173     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3174        for the benefit of those environments where the "b" makes a difference.
3175    
3176     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3177        to know about it. I have put a hack into configure.in that adds in code
3178        to set GCC=yes if CC=icc. This seems to end up at a point in the
3179        generated configure script that is early enough to affect the setting of
3180        compiler options, which is what is needed, but I have no means of testing
3181        whether it really works. (The user who reported this had patched the
3182        generated configure script, which of course I cannot do.)
3183    
3184        LATER: After change 22 below (new libtool files), the configure script
3185        seems to know about icc (and also ecc). Therefore, I have commented out
3186        this hack in configure.in.
3187    
3188     7. Added support for pkg-config (2 patches were sent in).
3189    
3190     8. Negated POSIX character classes that used a combination of internal tables
3191        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3192        [[:^ascii]]. Typically, they would match almost any characters. The other
3193        POSIX classes were not broken in this way.
3194    
3195     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3196        to find the match, as PCRE was deluded into thinking that the match had to
3197        start at the start point or following a newline. The same bug applied to
3198        patterns with negative forward assertions or any backward assertions
3199        preceding ".*" at the start, unless the pattern required a fixed first
3200        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3201    
3202    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3203        starting at the last subject character, bytes beyond the end of the subject
3204        string were read.
3205    
3206    11. Renamed the variable "class" as "classbits" to make life easier for C++
3207        users. (Previously there was a macro definition, but it apparently wasn't
3208        enough.)
3209    
3210    12. Added the new field "tables" to the extra data so that tables can be passed
3211        in at exec time, or the internal tables can be re-selected. This allows
3212        a compiled regex to be saved and re-used at a later time by a different
3213        program that might have everything at different addresses.
3214    
3215    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3216        -R library as well as a -L library.
3217    
3218    14. The debugging options of pcretest (-d on the command line or D on a
3219        pattern) showed incorrect output for anything following an extended class
3220        that contained multibyte characters and which was followed by a quantifier.
3221    
3222    15. Added optional support for general category Unicode character properties
3223        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3224        support. It adds about 90K to the size of the library. The meanings of the
3225        inbuilt class escapes such as \d and \s have NOT been changed.
3226    
3227    16. Updated pcredemo.c to include calls to free() to release the memory for the
3228        compiled pattern.
3229    
3230    17. The generated file chartables.c was being created in the source directory
3231        instead of in the building directory. This caused the build to fail if the
3232        source directory was different from the building directory, and was
3233        read-only.
3234    
3235    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3236        file. No doubt somebody will tell me if they don't make sense... Also added
3237        Dan Mooney's comments about building on OpenVMS.
3238    
3239    19. Added support for partial matching via the PCRE_PARTIAL option for
3240        pcre_exec() and the \P data escape in pcretest.
3241    
3242    20. Extended pcretest with 3 new pattern features:
3243    
3244        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3245              write the compiled pattern to the file whose name is "rest-of-line".
3246              This is a straight binary dump of the data, with the saved pointer to
3247              the character tables forced to be NULL. The study data, if any, is
3248              written too. After writing, pcretest reads a new pattern.
3249    
3250        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3251              compiled pattern from the given file. There must not be any
3252              occurrences of "<" in the file name (pretty unlikely); if there are,
3253              pcretest will instead treat the initial "<" as a pattern delimiter.
3254              After reading in the pattern, pcretest goes on to read data lines as
3255              usual.
3256    
3257        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3258              and 16-bit fields in a compiled pattern, to simulate a pattern that
3259              was compiled on a host of opposite endianness.
3260    
3261    21. The pcre-exec() function can now cope with patterns that were compiled on
3262        hosts of opposite endianness, with this restriction:
3263    
3264          As for any compiled expression that is saved and used later, the tables
3265          pointer field cannot be preserved; the extra_data field in the arguments
3266          to pcre_exec() should be used to pass in a tables address if a value
3267          other than the default internal tables were used at compile time.
3268    
3269    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3270        now diagnosed as an error. Previously, most of the time, a negative number
3271        would have been treated as zero, but if in addition "ovector" was passed as
3272        NULL, a crash could occur.
3273    
3274    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3275        new versions from the libtool 1.5 distribution (the last one is a copy of
3276        a file called libtool.m4). This seems to have fixed the need to patch
3277        "configure" to support Darwin 1.3 (which I used to do). However, I still
3278        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3279        workstation).
3280    
3281    24. Changed the PCRE licence to be the more standard "BSD" licence.
3282    
3283    
3284    Version 4.5 01-Dec-03
3285    ---------------------
3286    
3287     1. There has been some re-arrangement of the code for the match() function so
3288        that it can be compiled in a version that does not call itself recursively.
3289        Instead, it keeps those local variables that need separate instances for
3290        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3291        needs to "recurse". Keeping track of where control must go is done by means
3292        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3293        hide most of the details from the main code, and operates only if
3294        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3295        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3296        operating.
3297    
3298        To make it easier for callers to provide specially tailored get/free
3299        functions for this usage, two new functions, pcre_stack_malloc, and
3300        pcre_stack_free, are used. They are always called in strict stacking order,
3301        and the size of block requested is always the same.
3302    
3303        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3304        PCRE has been compiled to use the stack or the heap for recursion. The
3305        -C option of pcretest uses this to show which version is compiled.
3306    
3307        A new data escape \S, is added to pcretest; it causes the amounts of store
3308        obtained and freed by both kinds of malloc/free at match time to be added
3309        to the output.
3310    
3311     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3312        what's available on my current Linux desktop machine.
3313    
3314     3. When matching a UTF-8 string, the test for a valid string at the start has
3315        been extended. If start_offset is not zero, PCRE now checks that it points
3316        to a byte that is the start of a UTF-8 character. If not, it returns
3317        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3318        this is necessary because there may be backward assertions in the pattern.
3319        When matching the same subject several times, it may save resources to use
3320        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3321    
3322     4. The code for checking the validity of UTF-8 strings has been tightened so
3323        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3324        containing "overlong sequences".
3325    
3326     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3327        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3328        should have been "&", but it just so happened that all the cases this let
3329        through by mistake were picked up later in the function.
3330    
3331     6. I had used a variable called "isblank" - this is a C99 function, causing
3332        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3333    
3334     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3335        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3336        so the version is shown; (c) stop "make test" echoing "./RunTest".
3337    
3338     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3339    
3340     9. The prototype for memmove() for systems that don't have it was using
3341        size_t, but the inclusion of the header that defines size_t was later. I've
3342        moved the #includes for the C headers earlier to avoid this.
3343    
3344    10. Added some adjustments to the code to make it easier to compiler on certain
3345        special systems:
3346    
3347          (a) Some "const" qualifiers were missing.
3348          (b) Added the macro EXPORT before all exported functions; by default this
3349              is defined to be empty.
3350          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3351              that it reads its output file name as an argument instead of writing
3352              to the standard output and assuming this can be redirected.
3353    
3354    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3355        class containing characters with values greater than 255, PCRE compilation
3356        went into a loop.
3357    
3358    12. A recursive reference to a subpattern that was within another subpattern
3359        that had a minimum quantifier of zero caused PCRE to crash. For example,
3360        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3361        recursion. If the recursively-called subpattern itself had a zero repeat,
3362        that was OK.
3363    
3364    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3365        buffer into which it was copied (for escape processing) was still set at
3366        1024, so long lines caused crashes.
3367    
3368    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3369        "internal error: code overflow...". This applied to any character class
3370        that was followed by a possessive quantifier.
3371    
3372    15. Modified the Makefile to add libpcre.la as a prerequisite for
3373        libpcreposix.la because I was told this is needed for a parallel build to
3374        work.
3375    
3376    16. If a pattern that contained .* following optional items at the start was
3377        studied, the wrong optimizing data was generated, leading to matching
3378        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3379        matching string must start with a or b or c. The correct conclusion for
3380        this pattern is that a match can start with any character.
3381    
3382    
3383    Version 4.4 13-Aug-03
3384    ---------------------
3385    
3386     1. In UTF-8 mode, a character class containing characters with values between
3387        127 and 255 was not handled correctly if the compiled pattern was studied.
3388        In fixing this, I have also improved the studying algorithm for such
3389        classes (slightly).
3390    
3391     2. Three internal functions had redundant arguments passed to them. Removal
3392        might give a very teeny performance improvement.
3393    
3394     3. Documentation bug: the value of the capture_top field in a callout is *one
3395        more than* the number of the hightest numbered captured substring.
3396    
3397     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3398        in incorrectly linking with a previously installed version. They now link
3399        explicitly with libpcre.la.
3400    
3401     5. configure.in no longer needs to recognize Cygwin specially.
3402    
3403     6. A problem in pcre.in for Windows platforms is fixed.
3404    
3405     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3406        pcretest, it used to include the size of the study block as part of its
3407        output. Unfortunately, the structure contains a field that has a different
3408        size on different hardware architectures. This meant that the tests that
3409        showed this size failed. As the block is currently always of a fixed size,
3410        this information isn't actually particularly useful in pcretest output, so
3411        I have just removed it.
3412    
3413     8. Three pre-processor statements accidentally did not start in column 1.
3414        Sadly, there are *still* compilers around that complain, even though
3415        standard C has not required this for well over a decade. Sigh.
3416    
3417     9. In pcretest, the code for checking callouts passed small integers in the
3418        callout_data field, which is a void * field. However, some picky compilers
3419        complained about the casts involved for this on 64-bit systems. Now
3420        pcretest passes the address of the small integer instead, which should get
3421        rid of the warnings.
3422    
3423    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3424        both compile and run time, and gives an error if an invalid UTF-8 sequence
3425        is found. There is a option for disabling this check in cases where the
3426        string is known to be correct and/or the maximum performance is wanted.
3427    
3428    11. In response to a bug report, I changed one line in Makefile.in from
3429    
3430            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3431        to
3432            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3433    
3434        to look similar to other lines, but I have no way of telling whether this
3435        is the right thing to do, as I do not use Windows. No doubt I'll get told
3436        if it's wrong...
3437    
3438    
3439    Version 4.3 21-May-03
3440    ---------------------
3441    
3442    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3443       Makefile.
3444    
3445    2. Some refactoring to improve the quality of the code:
3446    
3447       (i)   The utf8_table... variables are now declared "const".
3448    
3449       (ii)  The code for \cx, which used the "case flipping" table to upper case
3450             lower case letters, now just substracts 32. This is ASCII-specific,
3451             but the whole concept of \cx is ASCII-specific, so it seems
3452             reasonable.
3453    
3454       (iii) PCRE was using its character types table to recognize decimal and
3455             hexadecimal digits in the pattern. This is silly, because it handles
3456             only 0-9, a-f, and A-F, but the character types table is locale-
3457             specific, which means strange things might happen. A private
3458             table is now used for this - though it costs 256 bytes, a table is
3459             much faster than multiple explicit tests. Of course, the standard
3460             character types table is still used for matching digits in subject
3461             strings against \d.
3462    
3463       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3464             ending in _t are). So I've renamed it as ESC_tee.
3465    
3466    3. The first argument for regexec() in the POSIX wrapper should have been
3467       defined as "const".
3468    
3469    4. Changed pcretest to use malloc() for its buffers so that they can be
3470       Electric Fenced for debugging.
3471    
3472    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3473       to read one or more bytes before the start of the subject string. Often this
3474       had no effect on PCRE's behaviour, but in some circumstances it could
3475       provoke a segmentation fault.
3476    
3477    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3478       to try to read one or more bytes before the start of the subject string.
3479    
3480    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3481       UTF-8 support could misbehave in various ways if the subject string
3482       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3483       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3484       back over UTF-8 characters.)
3485    
3486    
3487    Version 4.2 14-Apr-03
3488    ---------------------
3489    
3490    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3491    
3492    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3493         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3494         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3495         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3496         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3497           and BUILD_EXEEXT
3498         Note: automatic setting of the BUILD variables is not yet working
3499         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3500           compile-time but not at link-time
3501         [LINK]: use for linking executables only
3502         make different versions for Windows and non-Windows
3503         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3504           libraries
3505         [LINK_FOR_BUILD]: new variable
3506         [OBJEXT]: use throughout
3507         [EXEEXT]: use throughout
3508         <winshared>: new target
3509         <wininstall>: new target
3510         <dftables.o>: use native compiler
3511         <dftables>: use native linker
3512         <install>: handle Windows platform correctly
3513         <clean>: ditto
3514         <check>: ditto
3515         copy DLL to top builddir before testing
3516    
3517       As part of these changes, -no-undefined was removed again. This was reported
3518       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3519       in any case.
3520    
3521    3. Some tidies to get rid of compiler warnings:
3522    
3523       . In the match_data structure, match_limit was an unsigned long int, whereas
3524         match_call_count was an int. I've made them both unsigned long ints.
3525    
3526       . In pcretest the fact that a const uschar * doesn't automatically cast to
3527         a void * provoked a warning.
3528    
3529       . Turning on some more compiler warnings threw up some "shadow" variables
3530         and a few more missing casts.
3531    
3532    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3533       option, a class that contained a single character with a value between 128
3534       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3535    
3536    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3537       option, a class that contained several characters, but with at least one
3538       whose value was between 128 and 255 caused PCRE to crash.
3539    
3540    
3541    Version 4.1 12-Mar-03
3542    ---------------------
3543    
3544    1. Compiling with gcc -pedantic found a couple of places where casts were
3545    needed, and a string in dftables.c that was longer than standard compilers are
3546    required to support.
3547    
3548    2. Compiling with Sun's compiler found a few more places where the code could
3549    be tidied up in order to avoid warnings.
3550    
3551    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3552    first of these names is deprecated in the latest Autoconf in favour of the name
3553    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3554    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3555    analogy I have changed it to CFLAGS_FOR_BUILD.
3556    
3557    4. Added -no-undefined to the linking command in the Makefile, because this is
3558    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3559    linking step for the pcreposix library.
3560    
3561    5. PCRE was failing to diagnose the case of two named groups with the same
3562    name.
3563    
3564    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3565    literal character that is needed in the subject for a match, and scans along to
3566    ensure that it is present before embarking on the full matching process. This
3567    saves time in cases of nested unlimited repeats that are never going to match.
3568    Problem: the scan can take a lot of time if the subject is very long (e.g.
3569    megabytes), thus penalizing straightforward matches. It is now done only if the
3570    amount of subject to be scanned is less than 1000 bytes.
3571    
3572    7. A lesser problem with the same optimization is that it was recording the
3573    first character of an anchored pattern as "needed", thus provoking a search
3574    right along the subject, even when the first match of the pattern was going to
3575    fail. The "needed" character is now not set for anchored patterns, unless it
3576    follows something in the pattern that is of non-fixed length. Thus, it still
3577    fulfils its original purpose of finding quick non-matches in cases of nested
3578    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3579    
3580    
3581    Version 4.0 17-Feb-03
3582    ---------------------
3583    
3584    1. If a comment in an extended regex that started immediately after a meta-item
3585    extended to the end of string, PCRE compiled incorrect data. This could lead to
3586    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3587    
3588    2. Moved to autoconf 2.53 and libtool 1.4.2.
3589    
3590    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3591    the special perltest8 script is no longer needed - all the tests can be run
3592    from a single perltest script.
3593    
3594    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3595    by \s. It has now been removed in PCRE. This means it isn't recognized as
3596    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3597    class [:space:] *does* include VT, thereby creating a mess.
3598    
3599    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3600    space and tab.
3601    
3602    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3603    its new features into the main test script, reducing the number of scripts.
3604    
3605    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3606    were backward compatible, and made the (?i) apply to the whole pattern, as if
3607    /i were given. Now it behaves more logically, and applies the option setting
3608    only to what follows. PCRE has been changed to follow suit. However, if it
3609    finds options settings right at the start of the pattern, it extracts them into
3610    the global options, as before. Thus, they show up in the info data.
3611    
3612    8. Added support for the \Q...\E escape sequence. Characters in between are
3613    treated as literals. This is slightly different from Perl in that $ and @ are
3614    also handled as literals inside the quotes. In Perl, they will cause variable
3615    interpolation. Note the following examples:
3616    
3617        Pattern            PCRE matches      Perl matches
3618    
3619        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3620        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3621        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3622    
3623    For compatibility with Perl, \Q...\E sequences are recognized inside character
3624    classes as well as outside them.
3625    
3626    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3627    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3628    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3629    signed/unsigned warnings.
3630    
3631    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3632    option for pcretest, so I've replaced it by a simple function that does just
3633    that job.
3634    
3635    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3636    "pcregrep -".
3637    
3638    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3639    Java package. This provides some syntactic sugar for simple cases of what my
3640    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3641    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3642    item, you can use this simplified notation. Note that only makes sense with
3643    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3644    greediness, whatever the setting of the PCRE_UNGREEDY option.
3645    
3646    13. A change of greediness default within a pattern was not taking effect at
3647    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3648    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3649    was abstracted outside.
3650    
3651    14. PCRE now supports the \G assertion. It is true when the current matching
3652    position is at the start point of the match. This differs from \A when the
3653    starting offset is non-zero. Used with the /g option of pcretest (or similar
3654    code), it works in the same way as it does for Perl's /g option. If all
3655    alternatives of a regex begin with \G, the expression is anchored to the start
3656    match position, and the "anchored" flag is set in the compiled expression.
3657    
3658    15. Some bugs concerning the handling of certain option changes within patterns
3659    have been fixed. These applied to options other than (?ims). For example,
3660    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3661    been the other way round. Some of this was related to change 7 above.
3662    
3663    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3664    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3665    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3666    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3667    
3668    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3669    mode. Unlike ".", it always matches newline, whatever the setting of
3670    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3671    assertions. Perl allows it, but it doesn't (in general) work because it can't
3672    calculate the length of the lookbehind. At least, that's the case for Perl
3673    5.8.0 - I've been told they are going to document that it doesn't work in
3674    future.
3675    
3676    18. Added an error diagnosis for escapes that PCRE does not support: these are
3677    \L, \l, \N, \P, \p, \U, \u, and \X.
3678    
3679    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3680    reading past the end of the pattern in cases such as /[abcd/.
3681    
3682    20. PCRE was getting more memory than necessary for patterns with classes that
3683    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3684    
3685    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3686    compiling PCRE for use with Virtual Pascal.
3687    
3688    22. Small fix to the Makefile to make it work properly if the build is done
3689    outside the source tree.
3690    
3691    23. Added a new extension: a condition to go with recursion. If a conditional
3692    subpattern starts with (?(R) the "true" branch is used if recursion has
3693    happened, whereas the "false" branch is used only at the top level.
3694    
3695    24. When there was a very long string of literal characters (over 255 bytes
3696    without UTF support, over 250 bytes with UTF support), the computation of how
3697    much memory was required could be incorrect, leading to segfaults or other
3698    strange effects.
3699    
3700    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3701    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3702    there was a subsequent back reference to those brackets. This meant that, for
3703    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3704    possible to check for precisely this case. All we can do is abandon the
3705    optimization if .* occurs inside capturing brackets when there are any back
3706    references whatsoever. (See below for a better fix that came later.)
3707    
3708    26. The handling of the optimization for finding the first character of a
3709    non-anchored pattern, and for finding a character that is required later in the
3710    match were failing in some cases. This didn't break the matching; it just
3711    failed to optimize when it could. The way this is done has been re-implemented.
3712    
3713    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3714    
3715    28. Added a new feature that provides some of the functionality that Perl
3716    provides with (?{...}). The facility is termed a "callout". The way it is done
3717    in PCRE is for the caller to provide an optional function, by setting
3718    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3719    global variable. By default it is unset, which disables all calling out. To get
3720    the function called, the regex must include (?C) at appropriate points. This
3721    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3722    This provides a means of identifying different callout points. When PCRE
3723    reaches such a point in the regex, if pcre_callout has been set, the external
3724    function is called. It is provided with data in a structure called
3725    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3726    matching continues; if it returns a non-zero value, the match at the current
3727    point fails. However, backtracking will occur if possible. [This was changed
3728    later and other features added - see item 49 below.]
3729    
3730    29. pcretest is upgraded to test the callout functionality. It provides a
3731    callout function that displays information. By default, it shows the start of
3732    the match and the current position in the text. There are some new data escapes
3733    to vary what happens:
3734    
3735        \C+         in addition, show current contents of captured substrings
3736        \C-         do not supply a callout function
3737        \C!n        return 1 when callout number n is reached
3738        \C!n!m      return 1 when callout number n is reached for the mth time
3739    
3740    30. If pcregrep was called with the -l option and just a single file name, it
3741    output "<stdin>" if a match was found, instead of the file name.
3742    
3743    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3744    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3745    pcre_exec(). This saves a malloc/free per call. The default value of
3746    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3747    when configuring.
3748    
3749    32. The default maximum size of a compiled pattern is 64K. There have been a
3750    few cases of people hitting this limit. The code now uses macros to handle the
3751    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3752    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3753    configuring. Tests 2 and 5 work only with 2-byte links because they output
3754    debugging information about compiled patterns.
3755    
3756    33. Internal code re-arrangements:
3757    
3758    (a) Moved the debugging function for printing out a compiled regex into
3759        its own source file (printint.c) and used #include to pull it into
3760        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3761        separate copies.
3762    
3763    (b) Defined the list of op-code names for debugging as a macro in
3764        internal.h so that it is next to the definition of the opcodes.
3765    
3766    (c) Defined a table of op-code lengths for simpler skipping along compiled
3767        code. This is again a macro in internal.h so that it is next to the
3768        definition of the opcodes.
3769    
3770    34. Added support for recursive calls to individual subpatterns, along the
3771    lines of Robin Houston's patch (but implemented somewhat differently).
3772    
3773    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3774    allow it to read and process whole directories in Win32. This code was
3775    contributed by Lionel Fourquaux; it has not been tested by me.
3776    
3777    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3778    used to name a group. Names consist of alphanumerics and underscores, and must
3779    be unique. Back references use the syntax (?P=name) and recursive calls use
3780    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3781    numbers. The function pcre_fullinfo() can be used after compilation to extract
3782    a name/number map. There are three relevant calls:
3783    
3784      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3785      PCRE_INFO_NAMECOUNT            yields the number of entries
3786      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3787    
3788    The map is a vector of fixed-size entries. The size of each entry depends on
3789    the length of the longest name used. The first two bytes of each entry are the
3790    group number, most significant byte first. There follows the corresponding
3791    name, zero terminated. The names are in alphabetical order.
3792    
3793    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3794    case instead of 255. Making it the same both with and without UTF-8 support
3795    means that the same test output works with both.
3796    
3797    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3798    calling malloc() with a zero argument.
3799    
3800    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3801    optimization. I've improved things by keeping a bitmap of backreferences with
3802    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3803    fact referenced, the optimization can be applied. It is unlikely that a
3804    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3805    the match to follow \n) will appear inside brackets with a number greater than
3806    31, but if it does, any back reference > 31 suppresses the optimization.
3807    
3808    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3809    of disabling numbered capturing parentheses. Any opening parenthesis that is
3810    not followed by ? behaves as if it were followed by ?: but named parentheses
3811    can still be used for capturing (and they will acquire numbers in the usual
3812    way).
3813    
3814    41. Redesigned the return codes from the match() function into yes/no/error so
3815    that errors can be passed back from deep inside the nested calls. A malloc
3816    failure while inside a recursive subpattern call now causes the
3817    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3818    
3819    42. It is now possible to set a limit on the number of times the match()
3820    function is called in a call to pcre_exec(). This facility makes it possible to
3821    limit the amount of recursion and backtracking, though not in a directly
3822    obvious way, because the match() function is used in a number of different
3823    circumstances. The count starts from zero for each position in the subject
3824    string (for non-anchored patterns). The default limit is, for compatibility, a
3825    large number, namely 10 000 000. You can change this in two ways:
3826    
3827    (a) When configuring PCRE before making, you can use --with-match-limit=n
3828        to set a default value for the compiled library.
3829    
3830    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3831        a different value is set. See 45 below.
3832    
3833    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3834    
3835    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3836    of things that can be changed at compile time. The first argument specifies
3837    what is wanted and the second points to where the information is to be placed.
3838    The current list of available information is:
3839    
3840      PCRE_CONFIG_UTF8
3841    
3842    The output is an integer that is set to one if UTF-8 support is available;
3843    otherwise it is set to zero.
3844    
3845      PCRE_CONFIG_NEWLINE
3846    
3847    The output is an integer that it set to the value of the code that is used for
3848    newline. It is either LF (10) or CR (13).
3849    
3850      PCRE_CONFIG_LINK_SIZE
3851    
3852    The output is an integer that contains the number of bytes used for internal
3853    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3854    
3855      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3856    
3857    The output is an integer that contains the threshold above which the POSIX
3858    interface uses malloc() for output vectors. See item 31 above.
3859    
3860      PCRE_CONFIG_MATCH_LIMIT
3861    
3862    The output is an unsigned integer that contains the default limit of the number
3863    of match() calls in a pcre_exec() execution. See 42 above.
3864    
3865    44. pcretest has been upgraded by the addition of the -C option. This causes it
3866    to extract all the available output from the new pcre_config() function, and to
3867    output it. The program then exits immediately.
3868    
3869    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3870    order to support additional features. One way would have been to define
3871    pcre_exec2() (for example) with extra arguments, but this would not have been
3872    extensible, and would also have required all calls to the original function to
3873    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3874    is used for passing in "extra" data from pcre_study().
3875    
3876    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3877    contains the following fields:
3878    
3879      flags         a bitmap indicating which of the following fields are set
3880      study_data    opaque data from pcre_study()
3881      match_limit   a way of specifying a limit on match() calls for a specific
3882                      call to pcre_exec()
3883      callout_data  data for callouts (see 49 below)
3884    
3885    The flag bits are also defined in pcre.h, and are
3886    
3887      PCRE_EXTRA_STUDY_DATA
3888      PCRE_EXTRA_MATCH_LIMIT
3889      PCRE_EXTRA_CALLOUT_DATA
3890    
3891    The pcre_study() function now returns one of these new pcre_extra blocks, with
3892    the actual study data pointed to by the study_data field, and the
3893    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3894    before. That is, this change is entirely upwards-compatible and requires no
3895    change to existing code.
3896    
3897    If you want to pass in additional data to pcre_exec(), you can either place it
3898    in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3899    block.
3900    
3901    46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3902    data string contains the escape sequence \M, pcretest calls pcre_exec() several
3903    times with different match limits, until it finds the minimum value needed for
3904    pcre_exec() to complete. The value is then output. This can be instructive; for
3905    most simple matches the number is quite small, but for pathological cases it
3906    gets very large very quickly.
3907    
3908    47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3909    returns the size of the data block pointed to by the study_data field in a
3910    pcre_extra block, that is, the value that was passed as the argument to
3911    pcre_malloc() when PCRE was getting memory in which to place the information
3912    created by pcre_study(). The fourth argument should point to a size_t variable.
3913    pcretest has been extended so that this information is shown after a successful
3914    pcre_study() call when information about the compiled regex is being displayed.
3915    
3916    48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3917    because what follows is always an absolute path. (Later: it turns out that this
3918    is more than cosmetic for MinGW, because it doesn't like empty path
3919    components.)
3920    
3921    49. Some changes have been made to the callout feature (see 28 above):
3922    
3923    (i)  A callout function now has three choices for what it returns:
3924    
3925           0  =>  success, carry on matching
3926         > 0  =>  failure at this point, but backtrack if possible
3927         < 0  =>  serious error, return this value from pcre_exec()
3928    
3929         Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3930         values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3931         "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3932         use by callout functions. It will never be used by PCRE itself.
3933    
3934    (ii) The pcre_extra structure (see 45 above) has a void * field called
3935         callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3936         pcre_callout_block structure has a field of the same name. The contents of
3937         the field passed in the pcre_extra structure are passed to the callout
3938         function in the corresponding field in the callout block. This makes it
3939         easier to use the same callout-containing regex from multiple threads. For
3940