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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 966 by zherczeg, Mon May 14 06:27:21 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  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    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
95        \w+ when the character tables indicated that \x{c4} was a word character.
96        There were several related cases, all because the tests for doing a table
97        lookup were testing for characters less than 127 instead of 255.
98    
99    27. If a pattern contains capturing parentheses that are not used in a match,
100        their slots in the ovector are set to -1. For those that are higher than
101        any matched groups, this happens at the end of processing. In the case when
102        there were back references that the ovector was too small to contain
103        (causing temporary malloc'd memory to be used during matching), and the
104        highest capturing number was not used, memory off the end of the ovector
105        was incorrectly being set to -1. (It was using the size of the temporary
106        memory instead of the true size.)
107    
108    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
109        ovector size, it uses memory at the end of the block that it has got.
110    
111    29. Check for an overlong MARK name and give an error at compile time. The
112        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
113    
114    30. JIT compiler update.
115    
116    
117    Version 8.30 04-February-2012
118    -----------------------------
119    
120    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
121        name is defined in ctype.h.
122    
123    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
124        only in quite long subpatterns.
125    
126    3.  Removed the function pcre_info(), which has been obsolete and deprecated
127        since it was replaced by pcre_fullinfo() in February 2000.
128    
129    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
130        match a (*MARK), and the match failed at the start of the subject, a
131        reference to memory before the start of the subject could occur. This bug
132        was introduced by fix 17 of release 8.21.
133    
134    5.  A reference to an unset group with zero minimum repetition was giving
135        totally wrong answers (in non-JavaScript-compatibility mode). For example,
136        /(another)?(\1?)test/ matched against "hello world test". This bug was
137        introduced in release 8.13.
138    
139    6.  Add support for 16-bit character strings (a large amount of work involving
140        many changes and refactorings).
141    
142    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
143        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
144        from a file.
145    
146    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
147        rounding is not applied in this particular case).
148    
149    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
150        if they appear, or are escaped, in patterns.
151    
152    10. Get rid of a number of -Wunused-but-set-variable warnings.
153    
154    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
155        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
156        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
157        also returns the mark "x". This bug applied to capturing parentheses,
158        non-capturing parentheses, and atomic parentheses. It also applied to some
159        assertions.
160    
161    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
162        information out of configure.ac instead of relying on pcre.h.generic, which
163        is not stored in the repository.
164    
165    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
166        -lreadline.
167    
168    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
169    
170    15. Applied Graycode's patch to put the top-level frame on the stack rather
171        than the heap when not using the stack for recursion. This gives a
172        performance improvement in many cases when recursion is not deep.
173    
174    16. Experimental code added to "pcretest -C" to output the stack frame size.
175    
176    
177    Version 8.21 12-Dec-2011
178    ------------------------
179    
180    1.  Updating the JIT compiler.
181    
182    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
183        are added as well.
184    
185    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
186        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
187        calling _pcre_jit_exec. Some extra comments are added.
188    
189    4.  (*MARK) settings inside atomic groups that do not contain any capturing
190        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
191        was introduced by change 18 for 8.20.
192    
193    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
194        ECMA-262 standard.
195    
196    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
197        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
198        This bug was probably introduced by change 9 of 8.13.
199    
200    7.  While fixing 6 above, I noticed that a number of other items were being
201        incorrectly rejected as "not fixed length". This arose partly because newer
202        opcodes had not been added to the fixed-length checking code. I have (a)
203        corrected the bug and added tests for these items, and (b) arranged for an
204        error to occur if an unknown opcode is encountered while checking for fixed
205        length instead of just assuming "not fixed length". The items that were
206        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
207        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
208        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
209    
210    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
211        being incorrectly compiled and would have given unpredicatble results.
212    
213    9.  A possessively repeated subpattern with minimum repeat count greater than
214        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
215        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
216        the first (A) could occur when it should not.
217    
218    10. Add a cast and remove a redundant test from the code.
219    
220    11. JIT should use pcre_malloc/pcre_free for allocation.
221    
222    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
223        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
224        is anything other than /usr, -L is still shown).
225    
226    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
227    
228    14. Perl does not support \N without a following name in a [] class; PCRE now
229        also gives an error.
230    
231    15. If a forward reference was repeated with an upper limit of around 2000,
232        it caused the error "internal error: overran compiling workspace". The
233        maximum number of forward references (including repeats) was limited by the
234        internal workspace, and dependent on the LINK_SIZE. The code has been
235        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
236        the default depends on LINK_SIZE. There is a new upper limit (for safety)
237        of around 200,000 forward references. While doing this, I also speeded up
238        the filling in of repeated forward references.
239    
240    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
241        incorrectly expecting the subject to contain another "a" after the start.
242    
243    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
244        in the match, the SKIP should be ignored. This was not happening; instead
245        the SKIP was being treated as NOMATCH. For patterns such as
246        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
247        tested.
248    
249    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
250        now much more compatible with Perl, in particular in cases where the result
251        is a non-match for a non-anchored pattern. For example, if
252        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
253        "m", where previously it did not return a name. A side effect of this
254        change is that for partial matches, the last encountered mark name is
255        returned, as for non matches. A number of tests that were previously not
256        Perl-compatible have been moved into the Perl-compatible test files. The
257        refactoring has had the pleasing side effect of removing one argument from
258        the match() function, thus reducing its stack requirements.
259    
260    19. If the /S+ option was used in pcretest to study a pattern using JIT,
261        subsequent uses of /S (without +) incorrectly behaved like /S+.
262    
263    21. Retrieve executable code size support for the JIT compiler and fixing
264        some warnings.
265    
266    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
267        not work when the shorter character appeared right at the end of the
268        subject string.
269    
270    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
271        systems.
272    
273    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
274        output it when the /M option is used in pcretest.
275    
276    25. The CheckMan script was not being included in the distribution. Also, added
277        an explicit "perl" to run Perl scripts from the PrepareRelease script
278        because this is reportedly needed in Windows.
279    
280    26. If study data was being save in a file and studying had not found a set of
281        "starts with" bytes for the pattern, the data written to the file (though
282        never used) was taken from uninitialized memory and so caused valgrind to
283        complain.
284    
285    27. Updated RunTest.bat as provided by Sheri Pierce.
286    
287    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
288    
289    29. Computation of memory usage for the table of capturing group names was
290        giving an unnecessarily large value.
291    
292    
293    Version 8.20 21-Oct-2011
294    ------------------------
295    
296    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
297        a POSIX class. After further experiments with Perl, which convinced me that
298        Perl has bugs and confusions, a closing square bracket is no longer allowed
299        in a POSIX name. This bug also affected patterns with classes that started
300        with full stops.
301    
302    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
303        captured substring, but while checking the failing first alternative,
304        substring 1 is temporarily captured. If the output vector supplied to
305        pcre_exec() was not big enough for this capture, the yield of the function
306        was still zero ("insufficient space for captured substrings"). This cannot
307        be totally fixed without adding another stack variable, which seems a lot
308        of expense for a edge case. However, I have improved the situation in cases
309        such as /(a)(b)x|abc/ matched against "abc", where the return code
310        indicates that fewer than the maximum number of slots in the ovector have
311        been set.
312    
313    3.  Related to (2) above: when there are more back references in a pattern than
314        slots in the output vector, pcre_exec() uses temporary memory during
315        matching, and copies in the captures as far as possible afterwards. It was
316        using the entire output vector, but this conflicts with the specification
317        that only 2/3 is used for passing back captured substrings. Now it uses
318        only the first 2/3, for compatibility. This is, of course, another edge
319        case.
320    
321    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
322        main code base, and can be used by building with --enable-jit. When this is
323        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
324        runtime --no-jit option is given.
325    
326    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
327        ovector, the return from the function was zero, implying that there were
328        other matches that did not fit. The correct "exactly full" value is now
329        returned.
330    
331    6.  If a subpattern that was called recursively or as a subroutine contained
332        (*PRUNE) or any other control that caused it to give a non-standard return,
333        invalid errors such as "Error -26 (nested recursion at the same subject
334        position)" or even infinite loops could occur.
335    
336    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
337        computing the minimum length on reaching *ACCEPT, and so ended up with the
338        wrong value of 1 rather than 0. Further investigation indicates that
339        computing a minimum subject length in the presence of *ACCEPT is difficult
340        (think back references, subroutine calls), and so I have changed the code
341        so that no minimum is registered for a pattern that contains *ACCEPT.
342    
343    8.  If (*THEN) was present in the first (true) branch of a conditional group,
344        it was not handled as intended. [But see 16 below.]
345    
346    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
347        Sheri Pierce.
348    
349    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
350        the first byte in a match must be "a".
351    
352    11. Change 17 for 8.13 increased the recursion depth for patterns like
353        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
354        pattern contains any instances of (*THEN). If it does not, the old
355        optimizations are restored. It would be nice to do this on a per-group
356        basis, but at the moment that is not feasible.
357    
358    12. In some environments, the output of pcretest -C is CRLF terminated. This
359        broke RunTest's code that checks for the link size. A single white space
360        character after the value is now allowed for.
361    
362    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
363        For "fr", it uses the Windows-specific input and output files.
364    
365    14. If (*THEN) appeared in a group that was called recursively or as a
366        subroutine, it did not work as intended. [But see next item.]
367    
368    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
369        pattern fragments (but not containing any | characters). If A and B are
370        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
371        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
372        D. In other words, Perl considers parentheses that do not contain any |
373        characters to be part of a surrounding alternative, whereas PCRE was
374        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
375        differently. PCRE now behaves in the same way as Perl, except in the case
376        of subroutine/recursion calls such as (?1) which have in any case always
377        been different (but PCRE had them first :-).
378    
379    16. Related to 15 above: Perl does not treat the | in a conditional group as
380        creating alternatives. Such a group is treated in the same way as an
381        ordinary group without any | characters when processing (*THEN). PCRE has
382        been changed to match Perl's behaviour.
383    
384    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
385        RunGrepTest script failed.
386    
387    18. Change 22 for version 13 caused atomic groups to use more stack. This is
388        inevitable for groups that contain captures, but it can lead to a lot of
389        stack use in large patterns. The old behaviour has been restored for atomic
390        groups that do not contain any capturing parentheses.
391    
392    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
393        suppress the check for a minimum subject length at run time. (If it was
394        given to pcre_exec() or pcre_dfa_exec() it did work.)
395    
396    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
397        fail to work when decoding hex characters in data strings in EBCDIC
398        environments.
399    
400    21. It appears that in at least one Mac OS environment, the isxdigit() function
401        is implemented as a macro that evaluates to its argument more than once,
402        contravening the C 90 Standard (I haven't checked a later standard). There
403        was an instance in pcretest which caused it to go wrong when processing
404        \x{...} escapes in subject strings. The has been rewritten to avoid using
405        things like p++ in the argument of isxdigit().
406    
407    
408    Version 8.13 16-Aug-2011
409    ------------------------
410    
411    1.  The Unicode data tables have been updated to Unicode 6.0.0.
412    
413    2.  Two minor typos in pcre_internal.h have been fixed.
414    
415    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
416        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
417        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
418    
419    4.  There were a number of related bugs in the code for matching backrefences
420        caselessly in UTF-8 mode when codes for the characters concerned were
421        different numbers of bytes. For example, U+023A and U+2C65 are an upper
422        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
423        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
424        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
425        2-byte code at the end of the subject (it thought there wasn't enough data
426        left).
427    
428    5.  Comprehensive information about what went wrong is now returned by
429        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
430        as the output vector has at least 2 elements. The offset of the start of
431        the failing character and a reason code are placed in the vector.
432    
433    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
434        now returned is for the first byte of the failing character, instead of the
435        last byte inspected. This is an incompatible change, but I hope it is small
436        enough not to be a problem. It makes the returned offset consistent with
437        pcre_exec() and pcre_dfa_exec().
438    
439    7.  pcretest now gives a text phrase as well as the error number when
440        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
441        failure, the offset and reason code are output.
442    
443    8.  When \R was used with a maximizing quantifier it failed to skip backwards
444        over a \r\n pair if the subsequent match failed. Instead, it just skipped
445        back over a single character (\n). This seems wrong (because it treated the
446        two characters as a single entity when going forwards), conflicts with the
447        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
448        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
449        has been changed.
450    
451    9.  Some internal refactoring has changed the processing so that the handling
452        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
453        time (the PCRE_DOTALL option was changed this way some time ago: version
454        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
455        which was always a bit of a fudge. It also means that there is one less
456        argument for the match() function, which reduces its stack requirements
457        slightly. This change also fixes an incompatibility with Perl: the pattern
458        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
459    
460    10. More internal refactoring has drastically reduced the number of recursive
461        calls to match() for possessively repeated groups such as (abc)++ when
462        using pcre_exec().
463    
464    11. While implementing 10, a number of bugs in the handling of groups were
465        discovered and fixed:
466    
467        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
468        (a|)*(?1) gave a compile-time internal error.
469        ((a|)+)+  did not notice that the outer group could match an empty string.
470        (^a|^)+   was not marked as anchored.
471        (.*a|.*)+ was not marked as matching at start or after a newline.
472    
473    12. Yet more internal refactoring has removed another argument from the match()
474        function. Special calls to this function are now indicated by setting a
475        value in a variable in the "match data" data block.
476    
477    13. Be more explicit in pcre_study() instead of relying on "default" for
478        opcodes that mean there is no starting character; this means that when new
479        ones are added and accidentally left out of pcre_study(), testing should
480        pick them up.
481    
482    14. The -s option of pcretest has been documented for ages as being an old
483        synonym of -m (show memory usage). I have changed it to mean "force study
484        for every regex", that is, assume /S for every regex. This is similar to -i
485        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
486        using it. It makes it easier to run collections of tests with and without
487        study enabled, and thereby test pcre_study() more easily. All the standard
488        tests are now run with and without -s (but some patterns can be marked as
489        "never study" - see 20 below).
490    
491    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
492        restoration of the capturing data to the outer values was not happening
493        correctly.
494    
495    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
496        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
497        pattern had matched an empty string, and so incorrectly returned a no
498        match.
499    
500    17. There was optimizing code for the last branch of non-capturing parentheses,
501        and also for the obeyed branch of a conditional subexpression, which used
502        tail recursion to cut down on stack usage. Unfortunately, now that there is
503        the possibility of (*THEN) occurring in these branches, tail recursion is
504        no longer possible because the return has to be checked for (*THEN). These
505        two optimizations have therefore been removed. [But see 8.20/11 above.]
506    
507    18. If a pattern containing \R was studied, it was assumed that \R always
508        matched two bytes, thus causing the minimum subject length to be
509        incorrectly computed because \R can also match just one byte.
510    
511    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
512        was incorrectly computed.
513    
514    20. If /S is present twice on a test pattern in pcretest input, it now
515        *disables* studying, thereby overriding the use of -s on the command line
516        (see 14 above). This is necessary for one or two tests to keep the output
517        identical in both cases.
518    
519    21. When (*ACCEPT) was used in an assertion that matched an empty string and
520        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
521    
522    22. When an atomic group that contained a capturing parenthesis was
523        successfully matched, but the branch in which it appeared failed, the
524        capturing was not being forgotten if a higher numbered group was later
525        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
526        group 1 to "a", when in fact it should be unset. This applied to multi-
527        branched capturing and non-capturing groups, repeated or not, and also to
528        positive assertions (capturing in negative assertions does not happen
529        in PCRE) and also to nested atomic groups.
530    
531    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
532        subject after a captured substring, to make it easier to tell which of a
533        number of identical substrings has been captured.
534    
535    24. The way atomic groups are processed by pcre_exec() has been changed so that
536        if they are repeated, backtracking one repetition now resets captured
537        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
538        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
539        "aaa". Previously, it would have been "a". As part of this code
540        refactoring, the way recursive calls are handled has also been changed.
541    
542    25. If an assertion condition captured any substrings, they were not passed
543        back unless some other capturing happened later. For example, if
544        (?(?=(a))a) was matched against "a", no capturing was returned.
545    
546    26. When studying a pattern that contained subroutine calls or assertions,
547        the code for finding the minimum length of a possible match was handling
548        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
549        group 1 called group 2 while simultaneously a separate group 2 called group
550        1). A stack overflow occurred in this case. I have fixed this by limiting
551        the recursion depth to 10.
552    
553    27. Updated RunTest.bat in the distribution to the version supplied by Tom
554        Fortmann. This supports explicit test numbers on the command line, and has
555        argument validation and error reporting.
556    
557    28. An instance of \X with an unlimited repeat could fail if at any point the
558        first character it looked at was a mark character.
559    
560    29. Some minor code refactoring concerning Unicode properties and scripts
561        should reduce the stack requirement of match() slightly.
562    
563    30. Added the '=' option to pcretest to check the setting of unused capturing
564        slots at the end of the pattern, which are documented as being -1, but are
565        not included in the return count.
566    
567    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
568        compiled something random. Now it gives a compile-time error (as does
569        Perl).
570    
571    32. A *MARK encountered during the processing of a positive assertion is now
572        recorded and passed back (compatible with Perl).
573    
574    33. If --only-matching or --colour was set on a pcregrep call whose pattern
575        had alternative anchored branches, the search for a second match in a line
576        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
577        matched the line "0102" twice. The same bug affected patterns that started
578        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
579        twice.
580    
581    34. Previously, PCRE did not allow quantification of assertions. However, Perl
582        does, and because of capturing effects, quantifying parenthesized
583        assertions may at times be useful. Quantifiers are now allowed for
584        parenthesized assertions.
585    
586    35. A minor code tidy in pcre_compile() when checking options for \R usage.
587    
588    36. \g was being checked for fancy things in a character class, when it should
589        just be a literal "g".
590    
591    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
592        appearance of a nested POSIX class supersedes an apparent external class.
593        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
594        unescaped square brackets may also appear as part of class names. For
595        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
596        more like Perl. (But see 8.20/1 above.)
597    
598    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
599        was because it thought it was \N{name}, which is not supported).
600    
601    39. Add minix to OS list not supporting the -S option in pcretest.
602    
603    40. PCRE tries to detect cases of infinite recursion at compile time, but it
604        cannot analyze patterns in sufficient detail to catch mutual recursions
605        such as ((?1))((?2)). There is now a runtime test that gives an error if a
606        subgroup is called recursively as a subpattern for a second time at the
607        same position in the subject string. In previous releases this might have
608        been caught by the recursion limit, or it might have run out of stack.
609    
610    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
611        happen only once. PCRE was, however incorrectly giving a compile time error
612        "recursive call could loop indefinitely" because it cannot analyze the
613        pattern in sufficient detail. The compile time test no longer happens when
614        PCRE is compiling a conditional subpattern, but actual runaway loops are
615        now caught at runtime (see 40 above).
616    
617    42. It seems that Perl allows any characters other than a closing parenthesis
618        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
619        has been changed to be the same.
620    
621    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
622        as not to get warnings when autogen.sh is called. Also changed
623        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
624    
625    44. To help people who use pcregrep to scan files containing exceedingly long
626        lines, the following changes have been made:
627    
628        (a) The default value of the buffer size parameter has been increased from
629            8K to 20K. (The actual buffer used is three times this size.)
630    
631        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
632            PCRE is built.
633    
634        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
635            to be set at run time.
636    
637        (d) Numerical values in pcregrep options can be followed by K or M, for
638            example --buffer-size=50K.
639    
640        (e) If a line being scanned overflows pcregrep's buffer, an error is now
641            given and the return code is set to 2.
642    
643    45. Add a pointer to the latest mark to the callout data block.
644    
645    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
646        partial match of an empty string instead of no match. This was specific to
647        the use of ".".
648    
649    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
650        complete match instead of a partial match. This bug was dependent on both
651        the PCRE_UTF8 and PCRE_DOTALL options being set.
652    
653    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
654        starting byte set, because \b was not being ignored.
655    
656    
657    Version 8.12 15-Jan-2011
658    ------------------------
659    
660    1.  Fixed some typos in the markup of the man pages, and wrote a script that
661        checks for such things as part of the documentation building process.
662    
663    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
664        --match-limit and --recursion-limit options (added for 8.11). In
665        particular, this made one of the standard tests fail. (The integer value
666        went into the wrong half of a long int.)
667    
668    3.  If the --colour option was given to pcregrep with -v (invert match), it
669        did strange things, either producing crazy output, or crashing. It should,
670        of course, ignore a request for colour when reporting lines that do not
671        match.
672    
673    4.  Another pcregrep bug caused similar problems if --colour was specified with
674        -M (multiline) and the pattern match finished with a line ending.
675    
676    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
677        matched in multiline mode, the following line was shown as part of the
678        match. This seems wrong, so I have changed it.
679    
680    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
681        the check for further matches in the same line (so they could be coloured)
682        to overrun the end of the current line. If another match was found, it was
683        incorrectly shown (and then shown again when found in the next line).
684    
685    7.  If pcregrep was compiled under Windows, there was a reference to the
686        function pcregrep_exit() before it was defined. I am assuming this was
687        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
688        reported by a user. I've moved the definition above the reference.
689    
690    
691    Version 8.11 10-Dec-2010
692    ------------------------
693    
694    1.  (*THEN) was not working properly if there were untried alternatives prior
695        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
696        backtracked to try for "b" instead of moving to the next alternative branch
697        at the same level (in this case, to look for "c"). The Perl documentation
698        is clear that when (*THEN) is backtracked onto, it goes to the "next
699        alternative in the innermost enclosing group".
700    
701    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
702        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
703        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
704        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
705        (*THEN).
706    
707    3.  If \s appeared in a character class, it removed the VT character from
708        the class, even if it had been included by some previous item, for example
709        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
710        of \s, but is part of the POSIX "space" class.)
711    
712    4.  A partial match never returns an empty string (because you can always
713        match an empty string at the end of the subject); however the checking for
714        an empty string was starting at the "start of match" point. This has been
715        changed to the "earliest inspected character" point, because the returned
716        data for a partial match starts at this character. This means that, for
717        example, /(?<=abc)def/ gives a partial match for the subject "abc"
718        (previously it gave "no match").
719    
720    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
721        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
722        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
723        has an implication that the given string is incomplete (because a partial
724        match is preferred over a full match). For this reason, these items now
725        give a partial match in this situation. [Aside: previously, the one case
726        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
727        match rather than a full match, which was wrong by the old rules, but is
728        now correct.]
729    
730    6.  There was a bug in the handling of #-introduced comments, recognized when
731        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
732        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
733        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
734        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
735        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
736        places in pcre_compile().
737    
738    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
739        comments when looking ahead for named forward references to subpatterns,
740        the only newline sequence it recognized was NL. It now handles newlines
741        according to the set newline convention.
742    
743    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
744        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
745        cater for a lack of strerror(). These oversights have been fixed.
746    
747    9.  Added --match-limit and --recursion-limit to pcregrep.
748    
749    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
750    
751    11. When the -o option was used, pcregrep was setting a return code of 1, even
752        when matches were found, and --line-buffered was not being honoured.
753    
754    12. Added an optional parentheses number to the -o and --only-matching options
755        of pcregrep.
756    
757    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
758        can match an empty string. The code to do it in pcretest and pcredemo
759        needed fixing:
760    
761        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
762            only one byte after an empty string match just before CRLF (this case
763            just got forgotten; "any" and "anycrlf" were OK).
764    
765        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
766            mode when an empty string match preceded an ASCII character followed by
767            a non-ASCII character. (The code for advancing by one character rather
768            than one byte was nonsense.)
769    
770        (c) The pcredemo.c sample program did not have any code at all to handle
771            the cases when CRLF is a valid newline sequence.
772    
773    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
774        as a starting offset was within the subject string. There is now a new
775        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
776        negative or greater than the length of the string. In order to test this,
777        pcretest is extended to allow the setting of negative starting offsets.
778    
779    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
780        starting offset points to the beginning of a UTF-8 character was
781        unnecessarily clumsy. I tidied it up.
782    
783    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
784        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
785    
786    17. Nobody had reported that the --include_dir option, which was added in
787        release 7.7 should have been called --include-dir (hyphen, not underscore)
788        for compatibility with GNU grep. I have changed it to --include-dir, but
789        left --include_dir as an undocumented synonym, and the same for
790        --exclude-dir, though that is not available in GNU grep, at least as of
791        release 2.5.4.
792    
793    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
794        characters from a string of bytes) have been redefined so as not to use
795        loops, in order to improve performance in some environments. At the same
796        time, I abstracted some of the common code into auxiliary macros to save
797        repetition (this should not affect the compiled code).
798    
799    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
800        compile-time error is now given if \c is not followed by an ASCII
801        character, that is, a byte less than 128. (In EBCDIC mode, the code is
802        different, and any byte value is allowed.)
803    
804    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
805        START_OPTIMIZE option, which is now allowed at compile time - but just
806        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
807        to pcregrep and other applications that have no direct access to PCRE
808        options. The new /Y option in pcretest sets this option when calling
809        pcre_compile().
810    
811    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
812        back references. Groups containing recursive back references were forced to
813        be atomic by that change, but in the case of named groups, the amount of
814        memory required was incorrectly computed, leading to "Failed: internal
815        error: code overflow". This has been fixed.
816    
817    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
818        pcretest.c, to avoid build problems in some Borland environments.
819    
820    
821    Version 8.10 25-Jun-2010
822    ------------------------
823    
824    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
825        THEN.
826    
827    2.  (*ACCEPT) was not working when inside an atomic group.
828    
829    3.  Inside a character class, \B is treated as a literal by default, but
830        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
831        causes the error). The code is unchanged, but I tidied the documentation.
832    
833    4.  Inside a character class, PCRE always treated \R and \X as literals,
834        whereas Perl faults them if its -w option is set. I have changed PCRE so
835        that it faults them when PCRE_EXTRA is set.
836    
837    5.  Added support for \N, which always matches any character other than
838        newline. (It is the same as "." when PCRE_DOTALL is not set.)
839    
840    6.  When compiling pcregrep with newer versions of gcc which may have
841        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
842        declared with attribute warn_unused_result" were given. Just casting the
843        result to (void) does not stop the warnings; a more elaborate fudge is
844        needed. I've used a macro to implement this.
845    
846    7.  Minor change to pcretest.c to avoid a compiler warning.
847    
848    8.  Added four artifical Unicode properties to help with an option to make
849        \s etc use properties (see next item). The new properties are: Xan
850        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
851    
852    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
853        use Unicode properties. (*UCP) at the start of a pattern can be used to set
854        this option. Modified pcretest to add /W to test this facility. Added
855        REG_UCP to make it available via the POSIX interface.
856    
857    10. Added --line-buffered to pcregrep.
858    
859    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
860        studied, and the match started with a letter with a code point greater than
861        127 whose first byte was different to the first byte of the other case of
862        the letter, the other case of this starting letter was not recognized
863        (#976).
864    
865    12. If a pattern that was studied started with a repeated Unicode property
866        test, for example, \p{Nd}+, there was the theoretical possibility of
867        setting up an incorrect bitmap of starting bytes, but fortunately it could
868        not have actually happened in practice until change 8 above was made (it
869        added property types that matched character-matching opcodes).
870    
871    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
872        possible starting bytes for non-anchored patterns.
873    
874    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
875        \R, and also a number of cases that involve Unicode properties, both
876        explicit and implicit when PCRE_UCP is set.
877    
878    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
879        input, it could crash or give wrong results if characters with values
880        greater than 0xc0 were present in the subject string. (Detail: it assumed
881        UTF-8 input when processing these items.)
882    
883    16. Added a lot of (int) casts to avoid compiler warnings in systems where
884        size_t is 64-bit (#991).
885    
886    17. Added a check for running out of memory when PCRE is compiled with
887        --disable-stack-for-recursion (#990).
888    
889    18. If the last data line in a file for pcretest does not have a newline on
890        the end, a newline was missing in the output.
891    
892    19. The default pcre_chartables.c file recognizes only ASCII characters (values
893        less than 128) in its various bitmaps. However, there is a facility for
894        generating tables according to the current locale when PCRE is compiled. It
895        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
896        characters, are recognized by isspace() and therefore were getting set in
897        these tables, and indeed these tables seem to approximate to ISO 8859. This
898        caused a problem in UTF-8 mode when pcre_study() was used to create a list
899        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
900        which of course cannot start UTF-8 characters. I have changed the code so
901        that only real ASCII characters (less than 128) and the correct starting
902        bytes for UTF-8 encodings are set for characters greater than 127 when in
903        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
904        altogether.)
905    
906    20. Added the /T option to pcretest so as to be able to run tests with non-
907        standard character tables, thus making it possible to include the tests
908        used for 19 above in the standard set of tests.
909    
910    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
911        reference to a subpattern the other side of a comment that contains an
912        opening parenthesis caused either an internal compiling error, or a
913        reference to the wrong subpattern.
914    
915    
916    Version 8.02 19-Mar-2010
917    ------------------------
918    
919    1.  The Unicode data tables have been updated to Unicode 5.2.0.
920    
921    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
922        configured.
923    
924    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
925        original author of that file, following a query about its status.
926    
927    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
928        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
929    
930    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
931        quantifier applied to a forward-referencing subroutine call, could compile
932        incorrect code or give the error "internal error: previously-checked
933        referenced subpattern not found".
934    
935    6.  Both MS Visual Studio and Symbian OS have problems with initializing
936        variables to point to external functions. For these systems, therefore,
937        pcre_malloc etc. are now initialized to local functions that call the
938        relevant global functions.
939    
940    7.  There were two entries missing in the vectors called coptable and poptable
941        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
942        I've fixed the data, and added a kludgy way of testing at compile time that
943        the lengths are correct (equal to the number of opcodes).
944    
945    8.  Following on from 7, I added a similar kludge to check the length of the
946        eint vector in pcreposix.c.
947    
948    9.  Error texts for pcre_compile() are held as one long string to avoid too
949        much relocation at load time. To find a text, the string is searched,
950        counting zeros. There was no check for running off the end of the string,
951        which could happen if a new error number was added without updating the
952        string.
953    
954    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
955    
956    11. \K was not working if it appeared in an atomic group or in a group that
957        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
958        \K is "not well defined" if used in an assertion. PCRE now accepts it if
959        the assertion is positive, but not if it is negative.
960    
961    12. Change 11 fortuitously reduced the size of the stack frame used in the
962        "match()" function of pcre_exec.c by one pointer. Forthcoming
963        implementation of support for (*MARK) will need an extra pointer on the
964        stack; I have reserved it now, so that the stack frame size does not
965        decrease.
966    
967    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
968        item in branch that calls a recursion is a subroutine call - as in the
969        second branch in the above example - was incorrectly given the compile-
970        time error "recursive call could loop indefinitely" because pcre_compile()
971        was not correctly checking the subroutine for matching a non-empty string.
972    
973    14. The checks for overrunning compiling workspace could trigger after an
974        overrun had occurred. This is a "should never occur" error, but it can be
975        triggered by pathological patterns such as hundreds of nested parentheses.
976        The checks now trigger 100 bytes before the end of the workspace.
977    
978    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
979    
980    
981    Version 8.01 19-Jan-2010
982    ------------------------
983    
984    1.  If a pattern contained a conditional subpattern with only one branch (in
985        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
986        computed the wrong minimum data length (which is of course zero for such
987        subpatterns). This could cause incorrect "no match" results.
988    
989    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
990        the pattern is reset in the first branch, pcre_compile() failed with
991        "internal error: code overflow at offset...". This happened only when
992        the reset was to the original external option setting. (An optimization
993        abstracts leading options settings into an external setting, which was the
994        cause of this.)
995    
996    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
997        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
998        assertion pattern did not match (meaning that the assertion was true), it
999        was incorrectly treated as false if the SKIP had been reached during the
1000        matching. This also applied to assertions used as conditions.
1001    
1002    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1003        assertion subpattern, including such a pattern used as a condition,
1004        unpredictable results occurred, instead of the error return
1005        PCRE_ERROR_DFA_UITEM.
1006    
1007    5.  The C++ GlobalReplace function was not working like Perl for the special
1008        situation when an empty string is matched. It now does the fancy magic
1009        stuff that is necessary.
1010    
1011    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1012        removed. (These were left over from very, very early versions of PCRE.)
1013    
1014    7.  Some cosmetic changes to the code to make life easier when compiling it
1015        as part of something else:
1016    
1017        (a) Change DEBUG to PCRE_DEBUG.
1018    
1019        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1020            called "current" as "current_branch", to prevent a collision with the
1021            Linux macro when compiled as a kernel module.
1022    
1023        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1024            prevent a collision with the Linux macro when compiled as a kernel
1025            module.
1026    
1027    8.  In pcre_compile() there are some checks for integer overflows that used to
1028        cast potentially large values to (double). This has been changed to that
1029        when building, a check for int64_t is made, and if it is found, it is used
1030        instead, thus avoiding the use of floating point arithmetic. (There is no
1031        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1032        double.
1033    
1034    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1035        2005 (difference between two addresses compared to an unsigned value).
1036    
1037    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1038        custom one, because of the following reported problem in Windows:
1039    
1040          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1041              under Win32.
1042          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1043              therefore missing the function definition.
1044          - The compiler thus generates a "C" signature for the test function.
1045          - The linker fails to find the "C" function.
1046          - PCRE fails to configure if asked to do so against libbz2.
1047    
1048    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1049        messages were output:
1050    
1051          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1052          rerunning libtoolize, to keep the correct libtool macros in-tree.
1053          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1054    
1055        I have done both of these things.
1056    
1057    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1058        most of the time, it *can* run out if it is given a pattern that contains a
1059        runaway infinite recursion. I updated the discussion in the pcrestack man
1060        page.
1061    
1062    13. Now that we have gone to the x.xx style of version numbers, the minor
1063        version may start with zero. Using 08 or 09 is a bad idea because users
1064        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1065        interpreted as invalid octal numbers. I've updated the previous comment in
1066        configure.ac, and also added a check that gives an error if 08 or 09 are
1067        used.
1068    
1069    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1070        causing partial matching to fail when the end of the subject matched \W
1071        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1072    
1073    15. There were some discrepancies between the declarations in pcre_internal.h
1074        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1075        their definitions. The declarations used "const uschar *" and the
1076        definitions used USPTR. Even though USPTR is normally defined as "const
1077        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1078        reported that: "This difference in casting confuses some C++ compilers, for
1079        example, SunCC recognizes above declarations as different functions and
1080        generates broken code for hbpcre." I have changed the declarations to use
1081        USPTR.
1082    
1083    16. GNU libtool is named differently on some systems. The autogen.sh script now
1084        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1085        (FreeBSD).
1086    
1087    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1088        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1089        comment: "Figure out how to create a longlong from a string: strtoll and
1090        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1091        instance, but it only takes 2 args instead of 3!"
1092    
1093    18. A subtle bug concerned with back references has been fixed by a change of
1094        specification, with a corresponding code fix. A pattern such as
1095        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1096        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1097        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1098        same bug. Such groups have to be quantified to be useful, or contained
1099        inside another quantified group. (If there's no repetition, the reference
1100        can never match.) The problem arises because, having left the group and
1101        moved on to the rest of the pattern, a later failure that backtracks into
1102        the group uses the captured value from the final iteration of the group
1103        rather than the correct earlier one. I have fixed this in PCRE by forcing
1104        any group that contains a reference to itself to be an atomic group; that
1105        is, there cannot be any backtracking into it once it has completed. This is
1106        similar to recursive and subroutine calls.
1107    
1108    
1109    Version 8.00 19-Oct-09
1110    ----------------------
1111    
1112    1.  The table for translating pcre_compile() error codes into POSIX error codes
1113        was out-of-date, and there was no check on the pcre_compile() error code
1114        being within the table. This could lead to an OK return being given in
1115        error.
1116    
1117    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1118        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1119        in a Windows environment.
1120    
1121    3.  The pcregrep --count option prints the count for each file even when it is
1122        zero, as does GNU grep. However, pcregrep was also printing all files when
1123        --files-with-matches was added. Now, when both options are given, it prints
1124        counts only for those files that have at least one match. (GNU grep just
1125        prints the file name in this circumstance, but including the count seems
1126        more useful - otherwise, why use --count?) Also ensured that the
1127        combination -clh just lists non-zero counts, with no names.
1128    
1129    4.  The long form of the pcregrep -F option was incorrectly implemented as
1130        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1131        but it seems right to fix it, and I didn't think it was worth preserving
1132        the old behaviour.
1133    
1134    5.  The command line items --regex=pattern and --regexp=pattern were not
1135        recognized by pcregrep, which required --regex pattern or --regexp pattern
1136        (with a space rather than an '='). The man page documented the '=' forms,
1137        which are compatible with GNU grep; these now work.
1138    
1139    6.  No libpcreposix.pc file was created for pkg-config; there was just
1140        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1141    
1142    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1143        when UCP support is not needed, by modifying the Python script that
1144        generates it from Unicode data files. This should not matter if the module
1145        is correctly used as a library, but I received one complaint about 50K of
1146        unwanted data. My guess is that the person linked everything into his
1147        program rather than using a library. Anyway, it does no harm.
1148    
1149    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1150        was a minimum greater than 1 for a wide character in a possessive
1151        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1152        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1153        character. Chaos in the form of incorrect output or a compiling loop could
1154        result.
1155    
1156    9.  The restrictions on what a pattern can contain when partial matching is
1157        requested for pcre_exec() have been removed. All patterns can now be
1158        partially matched by this function. In addition, if there are at least two
1159        slots in the offset vector, the offset of the earliest inspected character
1160        for the match and the offset of the end of the subject are set in them when
1161        PCRE_ERROR_PARTIAL is returned.
1162    
1163    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1164        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1165        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1166        and may be more useful for multi-segment matching.
1167    
1168    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1169        used to be given if ever the end of the subject was reached; now it is
1170        given only if matching could not proceed because another character was
1171        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1172        string "Z", which now yields "no match" instead of "partial match". In the
1173        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1174        final character ended with (*FAIL).
1175    
1176    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1177        if the pattern had a "must contain" character that was already found in the
1178        earlier partial match, unless partial matching was again requested. For
1179        example, with the pattern /dog.(body)?/, the "must contain" character is
1180        "g". If the first part-match was for the string "dog", restarting with
1181        "sbody" failed. This bug has been fixed.
1182    
1183    13. The string returned by pcre_dfa_exec() after a partial match has been
1184        changed so that it starts at the first inspected character rather than the
1185        first character of the match. This makes a difference only if the pattern
1186        starts with a lookbehind assertion or \b or \B (\K is not supported by
1187        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1188        matching functions compatible, and I think it's the right thing to do.
1189    
1190    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1191        so that the demonstration program is easily available in environments where
1192        PCRE has not been installed from source.
1193    
1194    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1195        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1196        library.
1197    
1198    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1199        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1200        is not the first non-POSIX option to be added. Clearly some people find
1201        these options useful.
1202    
1203    17. If a caller to the POSIX matching function regexec() passes a non-zero
1204        value for nmatch with a NULL value for pmatch, the value of
1205        nmatch is forced to zero.
1206    
1207    18. RunGrepTest did not have a test for the availability of the -u option of
1208        the diff command, as RunTest does. It now checks in the same way as
1209        RunTest, and also checks for the -b option.
1210    
1211    19. If an odd number of negated classes containing just a single character
1212        interposed, within parentheses, between a forward reference to a named
1213        subpattern and the definition of the subpattern, compilation crashed with
1214        an internal error, complaining that it could not find the referenced
1215        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1216        [The bug was that it was starting one character too far in when skipping
1217        over the character class, thus treating the ] as data rather than
1218        terminating the class. This meant it could skip too much.]
1219    
1220    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1221        /g option in pcretest when the pattern contains \K, which makes it possible
1222        to have an empty string match not at the start, even when the pattern is
1223        anchored. Updated pcretest and pcredemo to use this option.
1224    
1225    21. If the maximum number of capturing subpatterns in a recursion was greater
1226        than the maximum at the outer level, the higher number was returned, but
1227        with unset values at the outer level. The correct (outer level) value is
1228        now given.
1229    
1230    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1231        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1232        make it do so. The string so far is captured, making this feature
1233        compatible with Perl.
1234    
1235    23. The tests have been re-organized, adding tests 11 and 12, to make it
1236        possible to check the Perl 5.10 features against Perl 5.10.
1237    
1238    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1239        pattern matches a fixed length string. PCRE did not allow this; now it
1240        does. Neither allows recursion.
1241    
1242    25. I finally figured out how to implement a request to provide the minimum
1243        length of subject string that was needed in order to match a given pattern.
1244        (It was back references and recursion that I had previously got hung up
1245        on.) This code has now been added to pcre_study(); it finds a lower bound
1246        to the length of subject needed. It is not necessarily the greatest lower
1247        bound, but using it to avoid searching strings that are too short does give
1248        some useful speed-ups. The value is available to calling programs via
1249        pcre_fullinfo().
1250    
1251    26. While implementing 25, I discovered to my embarrassment that pcretest had
1252        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1253        study optimizations had never been tested with that matching function.
1254        Oops. What is worse, even when it was passed study data, there was a bug in
1255        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1256        were also very few tests of studied patterns with pcre_dfa_exec().
1257    
1258    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1259        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1260        on the other side of the coin, they are no longer allowed to have different
1261        names, because these cannot be distinguished in PCRE, and this has caused
1262        confusion. (This is a difference from Perl.)
1263    
1264    28. When duplicate subpattern names are present (necessarily with different
1265        numbers, as required by 27 above), and a test is made by name in a
1266        conditional pattern, either for a subpattern having been matched, or for
1267        recursion in such a pattern, all the associated numbered subpatterns are
1268        tested, and the overall condition is true if the condition is true for any
1269        one of them. This is the way Perl works, and is also more like the way
1270        testing by number works.
1271    
1272    
1273    Version 7.9 11-Apr-09
1274    ---------------------
1275    
1276    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1277        (pcretest), all targets were linked against these libraries. This included
1278        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1279        libraries. This caused unwanted dependencies to be created. This problem
1280        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1281        pcretest is linked with readline.
1282    
1283    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1284        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1285        moved outside it again, because FALSE and TRUE are already defined in AIX,
1286        but BOOL is not.
1287    
1288    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1289        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1290    
1291    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1292        hyphens) following file names and line numbers when outputting matching
1293        lines. This is not true; no spaces are inserted. I have also clarified the
1294        wording for the --colour (or --color) option.
1295    
1296    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1297        was not coloured; this is different to GNU grep, so I have changed it to be
1298        the same.
1299    
1300    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1301        each matching line was coloured. Now it goes on to look for further matches
1302        of any of the test patterns, which is the same behaviour as GNU grep.
1303    
1304    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1305        doesn't make sense to accept an empty string match in pcregrep, so I have
1306        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1307        seems to be how GNU grep behaves.
1308    
1309    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1310        start or after a newline", because the conditional assertion was not being
1311        correctly handled. The rule now is that both the assertion and what follows
1312        in the first alternative must satisfy the test.
1313    
1314    9.  If auto-callout was enabled in a pattern with a conditional group whose
1315        condition was an assertion, PCRE could crash during matching, both with
1316        pcre_exec() and pcre_dfa_exec().
1317    
1318    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1319        used for matching.
1320    
1321    11. Unicode property support in character classes was not working for
1322        characters (bytes) greater than 127 when not in UTF-8 mode.
1323    
1324    12. Added the -M command line option to pcretest.
1325    
1326    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1327    
1328    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1329    
1330    16. Added comments and documentation about mis-use of no_arg in the C++
1331        wrapper.
1332    
1333    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1334        from Martin Jerabek that uses macro names for all relevant character and
1335        string constants.
1336    
1337    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1338        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1339        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1340        these, but not everybody uses configure.
1341    
1342    19. A conditional group that had only one branch was not being correctly
1343        recognized as an item that could match an empty string. This meant that an
1344        enclosing group might also not be so recognized, causing infinite looping
1345        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1346        with the subject "ab", where knowledge that the repeated group can match
1347        nothing is needed in order to break the loop.
1348    
1349    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1350        exec(), but without supplying a callout function, matching went wrong.
1351    
1352    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1353        leak if the size of the offset vector was greater than 30. When the vector
1354        is smaller, the saved offsets during recursion go onto a local stack
1355        vector, but for larger vectors malloc() is used. It was failing to free
1356        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1357        error, in fact).
1358    
1359    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1360        heapframe that is used only when UTF-8 support is enabled. This caused no
1361        problem, but was untidy.
1362    
1363    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1364        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1365        included within another project.
1366    
1367    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1368        slightly modified by me:
1369    
1370          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1371              not building pcregrep.
1372    
1373          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1374              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1375    
1376    25. Forward references, both numeric and by name, in patterns that made use of
1377        duplicate group numbers, could behave incorrectly or give incorrect errors,
1378        because when scanning forward to find the reference group, PCRE was not
1379        taking into account the duplicate group numbers. A pattern such as
1380        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1381    
1382    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1383        the feature of a custom pointer more persuasive (as requested by a user).
1384    
1385    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1386        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1387        pre-defined.
1388    
1389    28. Added support for (*UTF8) at the start of a pattern.
1390    
1391    29. Arrange for flags added by the "release type" setting in CMake to be shown
1392        in the configuration summary.
1393    
1394    
1395    Version 7.8 05-Sep-08
1396    ---------------------
1397    
1398    1.  Replaced UCP searching code with optimized version as implemented for Ad
1399        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1400        stage table and inline lookup instead of a function, giving speed ups of 2
1401        to 5 times on some simple patterns that I tested. Permission was given to
1402        distribute the MultiStage2.py script that generates the tables (it's not in
1403        the tarball, but is in the Subversion repository).
1404    
1405    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1406        scripts.
1407    
1408    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1409        a group with a zero qualifier. The result of the study could be incorrect,
1410        or the function might crash, depending on the pattern.
1411    
1412    4.  Caseless matching was not working for non-ASCII characters in back
1413        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1414        It now works when Unicode Property Support is available.
1415    
1416    5.  In pcretest, an escape such as \x{de} in the data was always generating
1417        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1418        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1419        truncation.
1420    
1421    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1422    
1423    7.  Added two (int) casts to pcregrep when printing the difference of two
1424        pointers, in case they are 64-bit values.
1425    
1426    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1427        test 2 if it fails.
1428    
1429    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1430        and a #define of that name to empty if it is not externally set. This is to
1431        allow users of MSVC to set it if necessary.
1432    
1433    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1434        the convenience functions in the pcre_get.c source file.
1435    
1436    11. An option change at the start of a pattern that had top-level alternatives
1437        could cause overwriting and/or a crash. This command provoked a crash in
1438        some environments:
1439    
1440          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1441    
1442        This potential security problem was recorded as CVE-2008-2371.
1443    
1444    12. For a pattern where the match had to start at the beginning or immediately
1445        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1446        pcre_dfa_exec() could read past the end of the passed subject if there was
1447        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1448        pcretest so that it places the subject at the end of its malloc-ed buffer.
1449    
1450    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1451        exec() might read past the end of the data buffer in UTF-8 mode.
1452    
1453    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1454        the data contained the byte 0x85 as part of a UTF-8 character within its
1455        first line. This applied both to normal and DFA matching.
1456    
1457    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1458        /^[^d]*?$/8 failed to match "abc".
1459    
1460    16. Added a missing copyright notice to pcrecpp_internal.h.
1461    
1462    17. Make it more clear in the documentation that values returned from
1463        pcre_exec() in ovector are byte offsets, not character counts.
1464    
1465    18. Tidied a few places to stop certain compilers from issuing warnings.
1466    
1467    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1468        supplied by Stefan Weber. I made a further small update for 7.8 because
1469        there is a change of source arrangements: the pcre_searchfuncs.c module is
1470        replaced by pcre_ucd.c.
1471    
1472    
1473    Version 7.7 07-May-08
1474    ---------------------
1475    
1476    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1477        a string to a long long, pretend we don't even have a long long." This is
1478        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1479    
1480    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1481        pre-7.6 versions, which defined a global no_arg variable instead of putting
1482        it in the RE class. (See also #8 below.)
1483    
1484    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1485        Lopes.
1486    
1487    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1488    
1489        (1) The include/exclude patterns were being applied to the whole pathnames
1490            of files, instead of just to the final components.
1491    
1492        (2) If there was more than one level of directory, the subdirectories were
1493            skipped unless they satisfied the include/exclude conditions. This is
1494            inconsistent with GNU grep (and could even be seen as contrary to the
1495            pcregrep specification - which I improved to make it absolutely clear).
1496            The action now is always to scan all levels of directory, and just
1497            apply the include/exclude patterns to regular files.
1498    
1499    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1500        --exclude_dir in the tests to avoid scanning .svn directories.
1501    
1502    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1503        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1504        doesn't support NULs in patterns.
1505    
1506    7.  Added some missing "const"s to declarations of static tables in
1507        pcre_compile.c and pcre_dfa_exec.c.
1508    
1509    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1510        caused by fix #2  above. (Subsequently also a second patch to fix the
1511        first patch. And a third patch - this was a messy problem.)
1512    
1513    9.  Applied Craig's patch to remove the use of push_back().
1514    
1515    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1516        matching function regexec().
1517    
1518    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1519        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1520        references. PCRE supports relative numbers with this syntax (I don't think
1521        Oniguruma does).
1522    
1523    12. Previously, a group with a zero repeat such as (...){0} was completely
1524        omitted from the compiled regex. However, this means that if the group
1525        was called as a subroutine from elsewhere in the pattern, things went wrong
1526        (an internal error was given). Such groups are now left in the compiled
1527        pattern, with a new opcode that causes them to be skipped at execution
1528        time.
1529    
1530    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1531        to the way PCRE behaves:
1532    
1533        (a) A lone ] character is dis-allowed (Perl treats it as data).
1534    
1535        (b) A back reference to an unmatched subpattern matches an empty string
1536            (Perl fails the current match path).
1537    
1538        (c) A data ] in a character class must be notated as \] because if the
1539            first data character in a class is ], it defines an empty class. (In
1540            Perl it is not possible to have an empty class.) The empty class []
1541            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1542            The negative empty class [^] matches any one character, independently
1543            of the DOTALL setting.
1544    
1545    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1546        non-existent subpattern following a character class starting with ']' and
1547        containing () gave an internal compiling error instead of "reference to
1548        non-existent subpattern". Fortunately, when the pattern did exist, the
1549        compiled code was correct. (When scanning forwards to check for the
1550        existencd of the subpattern, it was treating the data ']' as terminating
1551        the class, so got the count wrong. When actually compiling, the reference
1552        was subsequently set up correctly.)
1553    
1554    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1555        it was being rejected as not supported by pcre_dfa_exec(), even though
1556        other assertions are supported. I have made pcre_dfa_exec() support
1557        (*FAIL).
1558    
1559    16. The implementation of 13c above involved the invention of a new opcode,
1560        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1561        cannot be changed at match time, I realized I could make a small
1562        improvement to matching performance by compiling OP_ALLANY instead of
1563        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1564        on the OP_ANY path.
1565    
1566    17. Compiling pcretest on Windows with readline support failed without the
1567        following two fixes: (1) Make the unistd.h include conditional on
1568        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1569    
1570    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1571        ncurses library to be included for pcretest when ReadLine support is
1572        requested, but also to allow for it to be overridden. This patch came from
1573        Daniel Bergström.
1574    
1575    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1576        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1577        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1578        spotting this.
1579    
1580    
1581    Version 7.6 28-Jan-08
1582    ---------------------
1583    
1584    1.  A character class containing a very large number of characters with
1585        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1586        overflow.
1587    
1588    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1589        HAVE_LONG_LONG is not defined.
1590    
1591    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1592        bring them up to date and include new features. This patch includes:
1593    
1594        - Fixed PH's badly added libz and libbz2 support.
1595        - Fixed a problem with static linking.
1596        - Added pcredemo. [But later removed - see 7 below.]
1597        - Fixed dftables problem and added an option.
1598        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1599            HAVE_LONG_LONG.
1600        - Added readline support for pcretest.
1601        - Added an listing of the option settings after cmake has run.
1602    
1603    4.  A user submitted a patch to Makefile that makes it easy to create
1604        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1605        Makefile.am that cause it to include this special target, without
1606        affecting anything else. Note that the same mingw target plus all
1607        the other distribution libraries and programs are now supported
1608        when configuring with CMake (see 6 below) instead of with
1609        Configure/Make.
1610    
1611    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1612        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1613        exported in the Windows port". It has not yet been confirmed that the patch
1614        solves the problem, but it does no harm.
1615    
1616    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1617        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1618        with CMake, and also correct the comment about stack recursion.
1619    
1620    7.  Remove the automatic building of pcredemo from the ./configure system and
1621        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1622        of a program that users should build themselves after PCRE is installed, so
1623        building it automatically is not really right. What is more, it gave
1624        trouble in some build environments.
1625    
1626    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1627    
1628    
1629    Version 7.5 10-Jan-08
1630    ---------------------
1631    
1632    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1633        values in parens when parsing an RE using the C++ wrapper."
1634    
1635    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1636        Characters greater than 255 were excluded from the class instead of being
1637        included.
1638    
1639    3.  The same bug as (2) above applied to negated POSIX classes such as
1640        [:^space:].
1641    
1642    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1643        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1644        I have changed it.
1645    
1646    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1647        first named subpattern) and a construct such as (?&a) would reference the
1648        first named subpattern whose name started with "a" (in other words, the
1649        length check was missing). Both these problems are fixed. "Subpattern name
1650        expected" is now given for (?&) (a zero-length name), and this patch also
1651        makes it give the same error for \k'' (previously it complained that that
1652        was a reference to a non-existent subpattern).
1653    
1654    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1655        this is right because (?- can be followed by option settings as well as by
1656        digits. I have, however, made the messages clearer.
1657    
1658    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1659        than the number used in the conditional) now cause a compile-time error.
1660        This is actually not compatible with Perl, which accepts such patterns, but
1661        treats the conditional as always being FALSE (as PCRE used to), but it
1662        seems to me that giving a diagnostic is better.
1663    
1664    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1665        and messages.
1666    
1667    9.  Fix two occurrences of "backslash" in comments that should have been
1668        "backspace".
1669    
1670    10. Remove two redundant lines of code that can never be obeyed (their function
1671        was moved elsewhere).
1672    
1673    11. The program that makes PCRE's Unicode character property table had a bug
1674        which caused it to generate incorrect table entries for sequences of
1675        characters that have the same character type, but are in different scripts.
1676        It amalgamated them into a single range, with the script of the first of
1677        them. In other words, some characters were in the wrong script. There were
1678        thirteen such cases, affecting characters in the following ranges:
1679    
1680          U+002b0 - U+002c1
1681          U+0060c - U+0060d
1682          U+0061e - U+00612
1683          U+0064b - U+0065e
1684          U+0074d - U+0076d
1685          U+01800 - U+01805
1686          U+01d00 - U+01d77
1687          U+01d9b - U+01dbf
1688          U+0200b - U+0200f
1689          U+030fc - U+030fe
1690          U+03260 - U+0327f
1691          U+0fb46 - U+0fbb1
1692          U+10450 - U+1049d
1693    
1694    12. The -o option (show only the matching part of a line) for pcregrep was not
1695        compatible with GNU grep in that, if there was more than one match in a
1696        line, it showed only the first of them. It now behaves in the same way as
1697        GNU grep.
1698    
1699    13. If the -o and -v options were combined for pcregrep, it printed a blank
1700        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1701        does the same. The return code can be used to tell if there were any
1702        non-matching lines.
1703    
1704    14. Added --file-offsets and --line-offsets to pcregrep.
1705    
1706    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1707        infinitely looping recursion. The bug was that positive lookaheads were not
1708        being skipped when checking for a possible empty match (negative lookaheads
1709        and both kinds of lookbehind were skipped).
1710    
1711    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1712        inclusion of <windows.h> to before rather than after the definition of
1713        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1714    
1715    17. Specifying a possessive quantifier with a specific limit for a Unicode
1716        character property caused pcre_compile() to compile bad code, which led at
1717        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1718        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1719        caused the error; without that there was no problem.
1720    
1721    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1722    
1723    19. Added --enable-pcretest-libreadline.
1724    
1725    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1726        RE::GlobalReplace(). As a result, the number of replacements returned was
1727        double what it should be. I removed one of the increments, but Craig sent a
1728        later patch that removed the other one (the right fix) and added unit tests
1729        that check the return values (which was not done before).
1730    
1731    21. Several CMake things:
1732    
1733        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1734            the names libpcre and libpcreposix, not just pcre and pcreposix.
1735    
1736        (2) The above change means that pcretest and pcregrep are now correctly
1737            linked with the newly-built libraries, not previously installed ones.
1738    
1739        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1740    
1741    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1742        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1743        UTF-8 newline character). The key issue is that the pattern starts .*;
1744        this means that the match must be either at the beginning, or after a
1745        newline. The bug was in the code for advancing after a failed match and
1746        checking that the new position followed a newline. It was not taking
1747        account of UTF-8 characters correctly.
1748    
1749    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1750        character classes. PCRE was not treating the sequence [:...:] as a
1751        character class unless the ... were all letters. Perl, however, seems to
1752        allow any characters between [: and :], though of course it rejects as
1753        unknown any "names" that contain non-letters, because all the known class
1754        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1755        for example, whereas PCRE did not - it did not recognize a POSIX character
1756        class. This seemed a bit dangerous, so the code has been changed to be
1757        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1758        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1759        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1760        Perl does, and where it didn't before.
1761    
1762    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1763        Windows environments %n is disabled by default.
1764    
1765    
1766    Version 7.4 21-Sep-07
1767    ---------------------
1768    
1769    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1770        means that a class such as [\s] counted as "explicit reference to CR or
1771        LF". That isn't really right - the whole point of the change was to try to
1772        help when there was an actual mention of one of the two characters. So now
1773        the change happens only if \r or \n (or a literal CR or LF) character is
1774        encountered.
1775    
1776    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1777        of both had grown to the point where there were only 3 bits left.
1778        Fortunately, there was spare space in the data structure, and so I have
1779        moved the internal flags into a new 16-bit field to free up more option
1780        bits.
1781    
1782    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1783        but did not set the internal JCHANGED flag - either of these is enough to
1784        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1785        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1786        start sets both bits.
1787    
1788    4.  Added options (at build time, compile time, exec time) to change \R from
1789        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1790    
1791    5.  doc/pcresyntax.html was missing from the distribution.
1792    
1793    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1794        compatibility, even though it is no longer used.
1795    
1796    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1797        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1798        windows.h file is present (where different names are used). [This was
1799        reversed later after testing - see 16 below.]
1800    
1801    8.  Changed all #include <config.h> to #include "config.h". There were also
1802        some further <pcre.h> cases that I changed to "pcre.h".
1803    
1804    9.  When pcregrep was used with the --colour option, it missed the line ending
1805        sequence off the lines that it output.
1806    
1807    10. It was pointed out to me that arrays of string pointers cause lots of
1808        relocations when a shared library is dynamically loaded. A technique of
1809        using a single long string with a table of offsets can drastically reduce
1810        these. I have refactored PCRE in four places to do this. The result is
1811        dramatic:
1812    
1813          Originally:                          290
1814          After changing UCP table:            187
1815          After changing error message table:   43
1816          After changing table of "verbs"       36
1817          After changing table of Posix names   22
1818    
1819        Thanks to the folks working on Gregex for glib for this insight.
1820    
1821    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1822        unicode-properties was also set.
1823    
1824    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1825    
1826    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1827        checked only for CRLF.
1828    
1829    14. Added casts to pcretest.c to avoid compiler warnings.
1830    
1831    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1832    
1833    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1834        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1835        entirely. This removes changes made in 7 above.
1836    
1837    17. The CMake files have been updated, and there is now more information about
1838        building with CMake in the NON-UNIX-USE document.
1839    
1840    
1841    Version 7.3 28-Aug-07
1842    ---------------------
1843    
1844     1. In the rejigging of the build system that eventually resulted in 7.1, the
1845        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1846        brackets there is not right, since it causes compilers to look for an
1847        installed pcre.h, not the version that is in the source that is being
1848        compiled (which of course may be different). I have changed it back to:
1849    
1850          #include "pcre.h"
1851    
1852        I have a vague recollection that the change was concerned with compiling in
1853        different directories, but in the new build system, that is taken care of
1854        by the VPATH setting the Makefile.
1855    
1856     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1857        when the subject happened to end in the byte 0x85 (e.g. if the last
1858        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1859        characters but of course it shouldn't be taken as a newline when it is part
1860        of another character. The bug was that, for an unlimited repeat of . in
1861        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1862        characters when looking for a newline.
1863    
1864     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1865    
1866     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1867        in debug output.
1868    
1869     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1870        long printing in the pcrecpp unittest when running under MinGW.
1871    
1872     6. ESC_K was left out of the EBCDIC table.
1873    
1874     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1875        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1876        limit also applies to "virtual nesting" when a pattern is recursive, and in
1877        this case 1000 isn't so big. I have been able to remove this limit at the
1878        expense of backing off one optimization in certain circumstances. Normally,
1879        when pcre_exec() would call its internal match() function recursively and
1880        immediately return the result unconditionally, it uses a "tail recursion"
1881        feature to save stack. However, when a subpattern that can match an empty
1882        string has an unlimited repetition quantifier, it no longer makes this
1883        optimization. That gives it a stack frame in which to save the data for
1884        checking that an empty string has been matched. Previously this was taken
1885        from the 1000-entry workspace that had been reserved. So now there is no
1886        explicit limit, but more stack is used.
1887    
1888     8. Applied Daniel's patches to solve problems with the import/export magic
1889        syntax that is required for Windows, and which was going wrong for the
1890        pcreposix and pcrecpp parts of the library. These were overlooked when this
1891        problem was solved for the main library.
1892    
1893     9. There were some crude static tests to avoid integer overflow when computing
1894        the size of patterns that contain repeated groups with explicit upper
1895        limits. As the maximum quantifier is 65535, the maximum group length was
1896        set at 30,000 so that the product of these two numbers did not overflow a
1897        32-bit integer. However, it turns out that people want to use groups that
1898        are longer than 30,000 bytes (though not repeat them that many times).
1899        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1900        made it possible to implement the integer overflow checks in a much more
1901        dynamic way, which I have now done. The artificial limitation on group
1902        length has been removed - we now have only the limit on the total length of
1903        the compiled pattern, which depends on the LINK_SIZE setting.
1904    
1905    10. Fixed a bug in the documentation for get/copy named substring when
1906        duplicate names are permitted. If none of the named substrings are set, the
1907        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1908        empty string.
1909    
1910    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1911        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1912        because the ] is interpreted as the first data character and the
1913        terminating ] is not found. PCRE has been made compatible with Perl in this
1914        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1915        cause memory overwriting.
1916    
1917    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1918        string has been matched (to stop an infinite loop). It was not recognizing
1919        a conditional subpattern that could match an empty string if that
1920        subpattern was within another subpattern. For example, it looped when
1921        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1922        condition was not nested. This bug has been fixed.
1923    
1924    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1925        past the start of the subject in the presence of bytes with the top bit
1926        set, for example "\x8aBCD".
1927    
1928    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1929        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1930    
1931    14. Optimized (?!) to (*FAIL).
1932    
1933    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1934        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1935        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1936        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1937        does: it's just the validity check that is more restrictive.
1938    
1939    16. Inserted checks for integer overflows during escape sequence (backslash)
1940        processing, and also fixed erroneous offset values for syntax errors during
1941        backslash processing.
1942    
1943    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1944        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1945    
1946    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1947        caused an overrun.
1948    
1949    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1950        something other than just ASCII characters) inside a group that had an
1951        unlimited repeat caused a loop at compile time (while checking to see
1952        whether the group could match an empty string).
1953    
1954    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1955        [\P{Any}] did so. (Error in the code for printing property names.)
1956    
1957    21. An orphan \E inside a character class could cause a crash.
1958    
1959    22. A repeated capturing bracket such as (A)? could cause a wild memory
1960        reference during compilation.
1961    
1962    23. There are several functions in pcre_compile() that scan along a compiled
1963        expression for various reasons (e.g. to see if it's fixed length for look
1964        behind). There were bugs in these functions when a repeated \p or \P was
1965        present in the pattern. These operators have additional parameters compared
1966        with \d, etc, and these were not being taken into account when moving along
1967        the compiled data. Specifically:
1968    
1969        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1970            length.
1971    
1972        (b) An item such as \pL+ within a repeated group could cause crashes or
1973            loops.
1974    
1975        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1976            "reference to non-existent subpattern" error.
1977    
1978        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1979    
1980    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1981        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1982    
1983    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1984    
1985    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1986        character were causing crashes (broken optimization).
1987    
1988    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1989        \p or \P) caused a compile-time loop.
1990    
1991    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1992        break. For example, the unstudied pattern [\r\n]A does not match the string
1993        "\r\nA" because change 7.0/46 below moves the current point on by two
1994        characters after failing to match at the start. However, the pattern \nA
1995        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1996        the same is true. There doesn't seem any very clean way out of this, but
1997        what I have chosen to do makes the common cases work: PCRE now takes note
1998        of whether there can be an explicit match for \r or \n anywhere in the
1999        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2000        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2001        pattern has explicit CR or LF references.
2002    
2003    29. Added (*CR) etc for changing newline setting at start of pattern.
2004    
2005    
2006    Version 7.2 19-Jun-07
2007    ---------------------
2008    
2009     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2010        which is apparently normally available under Windows.
2011    
2012     2. Re-jig the pcregrep tests with different newline settings in an attempt
2013        to make them independent of the local environment's newline setting.
2014    
2015     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2016    
2017     4. Some of the "internals" tests were previously cut out when the link size
2018        was not 2, because the output contained actual offsets. The recent new
2019        "Z" feature of pcretest means that these can be cut out, making the tests
2020        usable with all link sizes.
2021    
2022     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2023        stack recursion. This gives a massive performance boost under BSD, but just
2024        a small improvement under Linux. However, it saves one field in the frame
2025        in all cases.
2026    
2027     6. Added more features from the forthcoming Perl 5.10:
2028    
2029        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2030            recursion call. It refers to the nth most recently opened parentheses.
2031    
2032        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2033            to be opened parentheses.
2034    
2035        (c) Conditions that refer to capturing parentheses can be specified
2036            relatively, for example, (?(-2)... or (?(+3)...
2037    
2038        (d) \K resets the start of the current match so that everything before
2039            is not part of it.
2040    
2041        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2042    
2043        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2044            reference syntax.
2045    
2046        (g) (?| introduces a group in which the numbering of parentheses in each
2047            alternative starts with the same number.
2048    
2049        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2050    
2051     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2052        PCRE_INFO_JCHANGED.
2053    
2054     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2055        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2056        for detecting groups that can match an empty string.
2057    
2058     9. A pattern with a very large number of alternatives (more than several
2059        hundred) was running out of internal workspace during the pre-compile
2060        phase, where pcre_compile() figures out how much memory will be needed. A
2061        bit of new cunning has reduced the workspace needed for groups with
2062        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2063        workspace instead of running out of the 4096 that are available.
2064    
2065    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2066    
2067    11. Applied patch from Google to remove an optimization that didn't quite work.
2068        The report of the bug said:
2069    
2070          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2071          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2072          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2073    
2074    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2075        it matched the wrong number of bytes.
2076    
2077    
2078    Version 7.1 24-Apr-07
2079    ---------------------
2080    
2081     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2082        that is more "standard", making use of automake and other Autotools. There
2083        is some re-arrangement of the files and adjustment of comments consequent
2084        on this.
2085    
2086     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2087        for recursive directory scanning broke on some systems because the files
2088        are not scanned in any specific order and on different systems the order
2089        was different. A call to "sort" has been inserted into RunGrepTest for the
2090        approprate test as a short-term fix. In the longer term there may be an
2091        alternative.
2092    
2093     3. I had an email from Eric Raymond about problems translating some of PCRE's
2094        man pages to HTML (despite the fact that I distribute HTML pages, some
2095        people do their own conversions for various reasons). The problems
2096        concerned the use of low-level troff macros .br and .in. I have therefore
2097        removed all such uses from the man pages (some were redundant, some could
2098        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2099        HTML has been updated to handle .nf/.fi and to complain if it encounters
2100        .br or .in.
2101    
2102     4. Updated comments in configure.ac that get placed in config.h.in and also
2103        arranged for config.h to be included in the distribution, with the name
2104        config.h.generic, for the benefit of those who have to compile without
2105        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2106    
2107     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2108        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2109        makevp.bat for the current PCRE, using the additional files
2110        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2111    
2112     6. A Windows user reported a minor discrepancy with test 2, which turned out
2113        to be caused by a trailing space on an input line that had got lost in his
2114        copy. The trailing space was an accident, so I've just removed it.
2115    
2116     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2117        that is needed.
2118    
2119     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2120        as "const" (a) because they are and (b) because it helps the PHP
2121        maintainers who have recently made a script to detect big data structures
2122        in the php code that should be moved to the .rodata section. I remembered
2123        to update Builducptable as well, so it won't revert if ucptable.h is ever
2124        re-created.
2125    
2126     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2127        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2128        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2129        support is not required. This saves 1.5-2K of code, which is important in
2130        some applications.
2131    
2132        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2133        so as not to refer to the tables, even though these functions will never be
2134        called when UTF-8 support is disabled. Otherwise there are problems with a
2135        shared library.
2136    
2137    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2138    
2139        (a) It was defining its arguments as char * instead of void *.
2140    
2141        (b) It was assuming that all moves were upwards in memory; this was true
2142            a long time ago when I wrote it, but is no longer the case.
2143    
2144        The emulated memove() is provided for those environments that have neither
2145        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2146        is clearly not the case, as these two bugs were recently reported.
2147    
2148    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2149        and Detrail to create the HTML documentation, the .txt form of the man
2150        pages, and it removes trailing spaces from listed files. It also creates
2151        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2152        case, it wraps all the #defines with #ifndefs. This script should be run
2153        before "make dist".
2154    
2155    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2156        with Unicode property support.
2157    
2158        (a) For a maximizing quantifier, if the two different cases of the
2159            character were of different lengths in their UTF-8 codings (there are
2160            some cases like this - I found 11), and the matching function had to
2161            back up over a mixture of the two cases, it incorrectly assumed they
2162            were both the same length.
2163    
2164        (b) When PCRE was configured to use the heap rather than the stack for
2165            recursion during matching, it was not correctly preserving the data for
2166            the other case of a UTF-8 character when checking ahead for a match
2167            while processing a minimizing repeat. If the check also involved
2168            matching a wide character, but failed, corruption could cause an
2169            erroneous result when trying to check for a repeat of the original
2170            character.
2171    
2172    13. Some tidying changes to the testing mechanism:
2173    
2174        (a) The RunTest script now detects the internal link size and whether there
2175            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2176            values substituted by "configure". (The RunGrepTest script already did
2177            this for UTF-8.) The configure.ac script no longer substitutes the
2178            relevant variables.
2179    
2180        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2181            with length and offset values. This means that the output is different
2182            for different internal link sizes. Test 2 is skipped for link sizes
2183            other than 2 because of this, bypassing the problem. Unfortunately,
2184            there was also a test in test 3 (the locale tests) that used /B and
2185            failed for link sizes other than 2. Rather than cut the whole test out,
2186            I have added a new /Z option to pcretest that replaces the length and
2187            offset values with spaces. This is now used to make test 3 independent
2188            of link size. (Test 2 will be tidied up later.)
2189    
2190    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2191        segmentation fault instead of returning the appropriate error message.
2192    
2193    15. In multiline mode when the newline sequence was set to "any", the pattern
2194        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2195        This doesn't seem right; it now treats the CRLF combination as the line
2196        ending, and so does not match in that case. It's only a pattern such as ^$
2197        that would hit this one: something like ^ABC$ would have failed after \r
2198        and then tried again after \r\n.
2199    
2200    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2201        in an attempt to make files that differ only in their line terminators
2202        compare equal. This works on Linux.
2203    
2204    17. Under certain error circumstances pcregrep might try to free random memory
2205        as it exited. This is now fixed, thanks to valgrind.
2206    
2207    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2208        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2209        was because its rules for how to advance for /g after matching an empty
2210        string at the end of a line did not allow for this case. They now check for
2211        it specially.
2212    
2213    20. pcretest is supposed to handle patterns and data of any length, by
2214        extending its buffers when necessary. It was getting this wrong when the
2215        buffer for a data line had to be extended.
2216    
2217    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2218        CRLF as a newline sequence.
2219    
2220    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2221        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2222        I have nevertheless tidied it up.
2223    
2224    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2225    
2226    24. Added a man page for pcre-config.
2227    
2228    
2229    Version 7.0 19-Dec-06
2230    ---------------------
2231    
2232     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2233        moving to gcc 4.1.1.
2234    
2235     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2236        sys/time.h, which is documented as needed for this function. It doesn't
2237        seem to matter on Linux, but it showed up on some releases of OS X.
2238    
2239     3. It seems that there are systems where bytes whose values are greater than
2240        127 match isprint() in the "C" locale. The "C" locale should be the
2241        default when a C program starts up. In most systems, only ASCII printing
2242        characters match isprint(). This difference caused the output from pcretest
2243        to vary, making some of the tests fail. I have changed pcretest so that:
2244    
2245        (a) When it is outputting text in the compiled version of a pattern, bytes
2246            other than 32-126 are always shown as hex escapes.
2247    
2248        (b) When it is outputting text that is a matched part of a subject string,
2249            it does the same, unless a different locale has been set for the match
2250            (using the /L modifier). In this case, it uses isprint() to decide.
2251    
2252     4. Fixed a major bug that caused incorrect computation of the amount of memory
2253        required for a compiled pattern when options that changed within the
2254        pattern affected the logic of the preliminary scan that determines the
2255        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2256        that the computed length was too small. The symptoms of this bug were
2257        either the PCRE error "internal error: code overflow" from pcre_compile(),
2258        or a glibc crash with a message such as "pcretest: free(): invalid next
2259        size (fast)". Examples of patterns that provoked this bug (shown in
2260        pcretest format) are:
2261    
2262          /(?-x: )/x
2263          /(?x)(?-x: \s*#\s*)/
2264          /((?i)[\x{c0}])/8
2265          /(?i:[\x{c0}])/8
2266    
2267        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2268        is now done differently.
2269    
2270     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2271        wrapper classes; (b) implement a new function in the C++ scanner that is
2272        more efficient than the old way of doing things because it avoids levels of
2273        recursion in the regex matching; (c) add a paragraph to the documentation
2274        for the FullMatch() function.
2275    
2276     6. The escape sequence \n was being treated as whatever was defined as
2277        "newline". Not only was this contrary to the documentation, which states
2278        that \n is character 10 (hex 0A), but it also went horribly wrong when
2279        "newline" was defined as CRLF. This has been fixed.
2280    
2281     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2282        was being set to -1 for the "end of line" case (supposedly a value that no
2283        character can have). Though this value is never used (the check for end of
2284        line is "zero bytes in current character"), it caused compiler complaints.
2285        I've changed it to 0xffffffff.
2286    
2287     8. In pcre_version.c, the version string was being built by a sequence of
2288        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2289        string (as it is for production releases) called a macro with an empty
2290        argument. The C standard says the result of this is undefined. The gcc
2291        compiler treats it as an empty string (which was what was wanted) but it is
2292        reported that Visual C gives an error. The source has been hacked around to
2293        avoid this problem.
2294    
2295     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2296        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2297        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2298        of them did).
2299    
2300    10. Originally, pcretest opened its input and output without "b"; then I was
2301        told that "b" was needed in some environments, so it was added for release
2302        5.0 to both the input and output. (It makes no difference on Unix-like
2303        systems.) Later I was told that it is wrong for the input on Windows. I've
2304        now abstracted the modes into two macros, to make it easier to fiddle with
2305        them, and removed "b" from the input mode under Windows.
2306    
2307    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2308    
2309    12. Added -help and --help to pcretest as an official way of being reminded
2310        of the options.
2311    
2312    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2313        and pcrecpp.cc because they annoy compilers at high warning levels.
2314    
2315    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2316    
2317    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2318        scripts are not C programs :-) and which was not noticed because it works
2319        on Linux.
2320    
2321    16. pcretest is supposed to handle any length of pattern and data line (as one
2322        line or as a continued sequence of lines) by extending its input buffer if
2323        necessary. This feature was broken for very long pattern lines, leading to
2324        a string of junk being passed to pcre_compile() if the pattern was longer
2325        than about 50K.
2326    
2327    17. I have done a major re-factoring of the way pcre_compile() computes the
2328        amount of memory needed for a compiled pattern. Previously, there was code
2329        that made a preliminary scan of the pattern in order to do this. That was
2330        OK when PCRE was new, but as the facilities have expanded, it has become
2331        harder and harder to keep it in step with the real compile phase, and there
2332        have been a number of bugs (see for example, 4 above). I have now found a
2333        cunning way of running the real compile function in a "fake" mode that
2334        enables it to compute how much memory it would need, while actually only
2335        ever using a few hundred bytes of working memory and without too many
2336        tests of the mode. This should make future maintenance and development
2337        easier. A side effect of this work is that the limit of 200 on the nesting
2338        depth of parentheses has been removed (though this was never a serious
2339        limitation, I suspect). However, there is a downside: pcre_compile() now
2340        runs more slowly than before (30% or more, depending on the pattern). I
2341        hope this isn't a big issue. There is no effect on runtime performance.
2342    
2343    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2344        newline (only possible for the last line of a file) and it was a
2345        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2346    
2347    19. Added additional timing features to pcretest. (1) The -tm option now times
2348        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2349        separate command line item, by a number that specifies the number of
2350        repeats to use when timing. The default is 50000; this gives better
2351        precision, but takes uncomfortably long for very large patterns.
2352    
2353    20. Extended pcre_study() to be more clever in cases where a branch of a
2354        subpattern has no definite first character. For example, (a*|b*)[cd] would
2355        previously give no result from pcre_study(). Now it recognizes that the
2356        first character must be a, b, c, or d.
2357    
2358    21. There was an incorrect error "recursive call could loop indefinitely" if
2359        a subpattern (or the entire pattern) that was being tested for matching an
2360        empty string contained only one non-empty item after a nested subpattern.
2361        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2362        incorrectly, because the \d was being skipped in the check.
2363    
2364    22. The pcretest program now has a new pattern option /B and a command line
2365        option -b, which is equivalent to adding /B to every pattern. This causes
2366        it to show the compiled bytecode, without the additional information that
2367        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2368        is the same as /B/I).
2369    
2370    23. A new optimization is now able automatically to treat some sequences such
2371        as a*b as a*+b. More specifically, if something simple (such as a character
2372        or a simple class like \d) has an unlimited quantifier, and is followed by
2373        something that cannot possibly match the quantified thing, the quantifier
2374        is automatically "possessified".
2375    
2376    24. A recursive reference to a subpattern whose number was greater than 39
2377        went wrong under certain circumstances in UTF-8 mode. This bug could also
2378        have affected the operation of pcre_study().
2379    
2380    25. Realized that a little bit of performance could be had by replacing
2381        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2382    
2383    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2384    
2385    27. Possessive quantifiers such as a++ were previously implemented by turning
2386        them into atomic groups such as ($>a+). Now they have their own opcodes,
2387        which improves performance. This includes the automatically created ones
2388        from 23 above.
2389    
2390    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2391        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2392        the first matched character to be a colon. This applied both to named and
2393        numbered groups.
2394    
2395    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2396    
2397    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2398        building PCRE on Windows easier, so I have included it in the distribution.
2399    
2400    31. There is now a check in pcretest against a ridiculously large number being
2401        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2402        loop, the loop is abandoned.
2403    
2404    32. Forward references to subpatterns in conditions such as (?(2)...) where
2405        subpattern 2 is defined later cause pcre_compile() to search forwards in
2406        the pattern for the relevant set of parentheses. This search went wrong
2407        when there were unescaped parentheses in a character class, parentheses
2408        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2409    
2410    33. "Subroutine" calls and backreferences were previously restricted to
2411        referencing subpatterns earlier in the regex. This restriction has now
2412        been removed.
2413    
2414    34. Added a number of extra features that are going to be in Perl 5.10. On the
2415        whole, these are just syntactic alternatives for features that PCRE had
2416        previously implemented using the Python syntax or my own invention. The
2417        other formats are all retained for compatibility.
2418    
2419        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2420            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2421            also .NET compatible.
2422    
2423        (b) A recursion or subroutine call to a named group can now be defined as
2424            (?&name) as well as (?P>name).
2425    
2426        (c) A backreference to a named group can now be defined as \k<name> or
2427            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2428            5.10, are also .NET compatible.
2429    
2430        (d) A conditional reference to a named group can now use the syntax
2431            (?(<name>) or (?('name') as well as (?(name).
2432    
2433        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2434            groups (named and numbered) that are never evaluated inline, but can be
2435            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2436            is always false. There may be only one alternative in such a group.
2437    
2438        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2439            as the simple (?(R). The condition is true only if the most recent
2440            recursion is that of the given number or name. It does not search out
2441            through the entire recursion stack.
2442    
2443        (g) The escape \gN or \g{N} has been added, where N is a positive or
2444            negative number, specifying an absolute or relative reference.
2445    
2446    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2447        some "unreachable code" warnings.
2448    
2449    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2450        things, this adds five new scripts.
2451    
2452    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2453        There were also incompatibilities regarding the handling of \Q..\E inside
2454        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2455        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2456    
2457    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2458        matches an empty string, and forcibly breaks the loop. There were bugs in
2459        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2460        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2461        separate and independent bugs (that affected different cases) have been
2462        fixed.
2463    
2464    39. Refactored the code to abolish the use of different opcodes for small
2465        capturing bracket numbers. This is a tidy that I avoided doing when I
2466        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2467        The new approach is not only tidier, it makes it possible to reduce the
2468        memory needed to fix the previous bug (38).
2469    
2470    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2471        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2472        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2473        mode.
2474    
2475    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2476        report.
2477    
2478    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2479        copy construction and assignment in the C++ wrapper.
2480    
2481    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2482        couple of bugs that could have given wrong results in the "--newline=crlf"
2483        case.
2484    
2485    44. Added a number of casts and did some reorganization of signed/unsigned int
2486        variables following suggestions from Dair Grant. Also renamed the variable
2487        "this" as "item" because it is a C++ keyword.
2488    
2489    45. Arranged for dftables to add
2490    
2491          #include "pcre_internal.h"
2492    
2493        to pcre_chartables.c because without it, gcc 4.x may remove the array
2494        definition from the final binary if PCRE is built into a static library and
2495        dead code stripping is activated.
2496    
2497    46. For an unanchored pattern, if a match attempt fails at the start of a
2498        newline sequence, and the newline setting is CRLF or ANY, and the next two
2499        characters are CRLF, advance by two characters instead of one.
2500    
2501    
2502    Version 6.7 04-Jul-06
2503    ---------------------
2504    
2505     1. In order to handle tests when input lines are enormously long, pcretest has
2506        been re-factored so that it automatically extends its buffers when
2507        necessary. The code is crude, but this _is_ just a test program. The
2508        default size has been increased from 32K to 50K.
2509    
2510     2. The code in pcre_study() was using the value of the re argument before
2511        testing it for NULL. (Of course, in any sensible call of the function, it
2512        won't be NULL.)
2513    
2514     3. The memmove() emulation function in pcre_internal.h, which is used on
2515        systems that lack both memmove() and bcopy() - that is, hardly ever -
2516        was missing a "static" storage class specifier.
2517    
2518     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2519        containing an extended class (one that cannot be represented by a bitmap
2520        because it contains high-valued characters or Unicode property items, e.g.
2521        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2522        pattern, but PCRE should not loop if you do not (it no longer does).
2523        [Detail: two cases were found: (a) a repeated subpattern containing an
2524        extended class; (b) a recursive reference to a subpattern that followed a
2525        previous extended class. It wasn't skipping over the extended class
2526        correctly when UTF-8 mode was not set.]
2527    
2528     5. A negated single-character class was not being recognized as fixed-length
2529        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2530        compile error "lookbehind assertion is not fixed length".
2531    
2532     6. The RunPerlTest auxiliary script was showing an unexpected difference
2533        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2534        write a Perl script that can interpret lines of an input file either as
2535        byte characters or as UTF-8, which is what "perltest" was being required to
2536        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2537        can't do is switch easily at run time between having the "use utf8;" pragma
2538        or not. In the end, I fudged it by using the RunPerlTest script to insert
2539        "use utf8;" explicitly for the UTF-8 tests.
2540    
2541     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2542        the end of the subject string, contrary to the documentation and to what
2543        Perl does. This was true of both matching functions. Now it matches only at
2544        the start of the subject and immediately after *internal* newlines.
2545    
2546     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2547        a pointer to an int instead of a pointer to an unsigned long int. This
2548        caused problems on 64-bit systems.
2549    
2550     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2551        instance of the 'standard' template library not being so standard".
2552    
2553    10. There was no check on the number of named subpatterns nor the maximum
2554        length of a subpattern name. The product of these values is used to compute
2555        the size of the memory block for a compiled pattern. By supplying a very
2556        long subpattern name and a large number of named subpatterns, the size
2557        computation could be caused to overflow. This is now prevented by limiting
2558        the length of names to 32 characters, and the number of named subpatterns
2559        to 10,000.
2560    
2561    11. Subpatterns that are repeated with specific counts have to be replicated in
2562        the compiled pattern. The size of memory for this was computed from the
2563        length of the subpattern and the repeat count. The latter is limited to
2564        65535, but there was no limit on the former, meaning that integer overflow
2565        could in principle occur. The compiled length of a repeated subpattern is
2566        now limited to 30,000 bytes in order to prevent this.
2567    
2568    12. Added the optional facility to have named substrings with the same name.
2569    
2570    13. Added the ability to use a named substring as a condition, using the
2571        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2572        are numbers (not recommended). Forward references are permitted.
2573    
2574    14. Added forward references in named backreferences (if you see what I mean).
2575    
2576    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2577        pattern could run off the end of the subject. For example, the pattern
2578        "(?s)(.{1,5})"8 did this with the subject "ab".
2579    
2580    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2581        PCRE_CASELESS was set when matching characters that were quantified with ?
2582        or *.
2583    
2584    17. A character class other than a single negated character that had a minimum
2585        but no maximum quantifier - for example [ab]{6,} - was not handled
2586        correctly by pce_dfa_exec(). It would match only one character.
2587    
2588    18. A valid (though odd) pattern that looked like a POSIX character
2589        class but used an invalid character after [ (for example [[,abc,]]) caused
2590        pcre_compile() to give the error "Failed: internal error: code overflow" or
2591        in some cases to crash with a glibc free() error. This could even happen if
2592        the pattern terminated after [[ but there just happened to be a sequence of
2593        letters, a binary zero, and a closing ] in the memory that followed.
2594    
2595    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2596        over the years. Originally (before any Unicode support), just the bottom 8
2597        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2598        output from "man perlunicode" includes this:
2599    
2600          The regular expression compiler produces polymorphic opcodes.  That
2601          is, the pattern adapts to the data and automatically switches to
2602          the Unicode character scheme when presented with Unicode data--or
2603          instead uses a traditional byte scheme when presented with byte
2604          data.
2605    
2606        Sadly, a wide octal escape does not cause a switch, and in a string with
2607        no other multibyte characters, these octal escapes are treated as before.
2608        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2609        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2610        Unicode string.
2611    
2612        I have not perpetrated such confusion in PCRE. Up till now, it took just
2613        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2614        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2615        translate to the appropriate multibyte character.
2616    
2617    29. Applied some refactoring to reduce the number of warnings from Microsoft
2618        and Borland compilers. This has included removing the fudge introduced
2619        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2620        a warning about an unused variable.
2621    
2622    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2623        characters since release 4.0, because Perl (from release 5.004) does not.
2624        [Or at least, is documented not to: some releases seem to be in conflict
2625        with the documentation.] However, when a pattern was studied with
2626        pcre_study() and all its branches started with \s, PCRE still included VT
2627        as a possible starting character. Of course, this did no harm; it just
2628        caused an unnecessary match attempt.
2629    
2630    22. Removed a now-redundant internal flag bit that recorded the fact that case
2631        dependency changed within the pattern. This was once needed for "required
2632        byte" processing, but is no longer used. This recovers a now-scarce options
2633        bit. Also moved the least significant internal flag bit to the most-
2634        significant bit of the word, which was not previously used (hangover from
2635        the days when it was an int rather than a uint) to free up another bit for
2636        the future.
2637    
2638    23. Added support for CRLF line endings as well as CR and LF. As well as the
2639        default being selectable at build time, it can now be changed at runtime
2640        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2641        specify that it is scanning data with non-default line endings.
2642    
2643    24. Changed the definition of CXXLINK to make it agree with the definition of
2644        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2645    
2646    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2647        recursions. This makes a big different to stack usage for some patterns.
2648    
2649    26. If a subpattern containing a named recursion or subroutine reference such
2650        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2651        the space required for the compiled pattern went wrong and gave too small a
2652        value. Depending on the environment, this could lead to "Failed: internal
2653        error: code overflow at offset 49" or "glibc detected double free or
2654        corruption" errors.
2655    
2656    27. Applied patches from Google (a) to support the new newline modes and (b) to
2657        advance over multibyte UTF-8 characters in GlobalReplace.
2658    
2659    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2660        difference for some implementation of PCRE in some Windows version.
2661    
2662    29. Added some extra testing facilities to pcretest:
2663    
2664        \q<number>   in a data line sets the "match limit" value
2665        \Q<number>   in a data line sets the "match recursion limt" value
2666        -S <number>  sets the stack size, where <number> is in megabytes
2667    
2668        The -S option isn't available for Windows.
2669    
2670    
2671    Version 6.6 06-Feb-06
2672    ---------------------
2673    
2674     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2675        in pcreposix.h. I have copied the definition from pcre.h.
2676    
2677     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2678        because pcre.h is no longer a built file.
2679    
2680     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2681        not normally included in the compiled code.
2682    
2683    
2684    Version 6.5 01-Feb-06
2685    ---------------------
2686    
2687     1. When using the partial match feature with pcre_dfa_exec(), it was not
2688        anchoring the second and subsequent partial matches at the new starting
2689        point. This could lead to incorrect results. For example, with the pattern
2690        /1234/, partially matching against "123" and then "a4" gave a match.
2691    
2692     2. Changes to pcregrep:
2693    
2694        (a) All non-match returns from pcre_exec() were being treated as failures
2695            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2696            error message is output. Some extra information is given for the
2697            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2698            probably the only errors that are likely to be caused by users (by
2699            specifying a regex that has nested indefinite repeats, for instance).
2700            If there are more than 20 of these errors, pcregrep is abandoned.
2701    
2702        (b) A binary zero was treated as data while matching, but terminated the
2703            output line if it was written out. This has been fixed: binary zeroes
2704            are now no different to any other data bytes.
2705    
2706        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2707            used to set a locale for matching. The --locale=xxxx long option has
2708            been added (no short equivalent) to specify a locale explicitly on the
2709            pcregrep command, overriding the environment variables.
2710    
2711        (d) When -B was used with -n, some line numbers in the output were one less
2712            than they should have been.
2713    
2714        (e) Added the -o (--only-matching) option.
2715    
2716        (f) If -A or -C was used with -c (count only), some lines of context were
2717            accidentally printed for the final match.
2718    
2719        (g) Added the -H (--with-filename) option.
2720    
2721        (h) The combination of options -rh failed to suppress file names for files
2722            that were found from directory arguments.
2723    
2724        (i) Added the -D (--devices) and -d (--directories) options.
2725    
2726        (j) Added the -F (--fixed-strings) option.
2727    
2728        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2729    
2730        (l) Added the --colo(u)r option.
2731    
2732        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2733            is not present by default.
2734    
2735     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2736        items such as (?R) or (?1), when the recursion could match a number of
2737        alternatives. If it matched one of the alternatives, but subsequently,
2738        outside the recursion, there was a failure, the code tried to back up into
2739        the recursion. However, because of the way PCRE is implemented, this is not
2740        possible, and the result was an incorrect result from the match.
2741    
2742        In order to prevent this happening, the specification of recursion has
2743        been changed so that all such subpatterns are automatically treated as
2744        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2745    
2746     4. I had overlooked the fact that, in some locales, there are characters for
2747        which isalpha() is true but neither isupper() nor islower() are true. In
2748        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2749        and ordfeminine) are like this. This affected the treatment of \w and \W
2750        when they appeared in character classes, but not when they appeared outside
2751        a character class. The bit map for "word" characters is now created
2752        separately from the results of isalnum() instead of just taking it from the
2753        upper, lower, and digit maps. (Plus the underscore character, of course.)
2754    
2755     5. The above bug also affected the handling of POSIX character classes such as
2756        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2757        permanent tables. Instead, the bit maps for such a class were previously
2758        created as the appropriate unions of the upper, lower, and digit bitmaps.
2759        Now they are created by subtraction from the [[:word:]] class, which has
2760        its own bitmap.
2761    
2762     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2763        It is created by subtracting the vertical space characters (\x09, \x0a,
2764        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2765        subtraction was done in the overall bitmap for a character class, meaning
2766        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2767        be recognized. This bug has been fixed.
2768    
2769     7. Patches from the folks at Google:
2770    
2771          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2772          real life, but is still worth protecting against".
2773    
2774          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2775          regular expressions".
2776    
2777          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2778          have it.
2779    
2780          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2781          "configure" and the latter not, in order to fix a problem somebody had
2782          with compiling the Arg class on HP-UX.
2783    
2784          (e) Improve the error-handling of the C++ wrapper a little bit.
2785    
2786          (f) New tests for checking recursion limiting.
2787    
2788     8. The pcre_memmove() function, which is used only if the environment does not
2789        have a standard memmove() function (and is therefore rarely compiled),
2790        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2791        returning a result (though PCRE never actually uses the result).
2792    
2793     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2794        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2795        returned instead of calling malloc() with an overflowing number that would
2796        most likely cause subsequent chaos.
2797    
2798    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2799    
2800    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2801        with this option is matched, the nmatch and pmatch options of regexec() are
2802        ignored.
2803    
2804    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2805        provided in case anyone wants to the the POSIX interface with UTF-8
2806        strings.
2807    
2808    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2809        C++ linking (needed for some HP-UX environments).
2810    
2811    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2812        (unused parameter) and in the pcre_printint() function (omitted "default"
2813        switch label when the default is to do nothing).
2814    
2815    15. Added some code to make it possible, when PCRE is compiled as a C++
2816        library, to replace subject pointers for pcre_exec() with a smart pointer
2817        class, thus making it possible to process discontinuous strings.
2818    
2819    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2820        much the same function. They were added by different people who were trying
2821        to make PCRE easy to compile on non-Unix systems. It has been suggested
2822        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2823        for compiling on Windows systems. I have therefore replaced it with
2824        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2825        defaults to "extern" for C or "extern C" for C++, which works fine on
2826        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2827        SCOPE with something explicit in config.h. In addition:
2828    
2829        (a) pcreposix.h still had just "extern" instead of either of these macros;
2830            I have replaced it with PCRE_DATA_SCOPE.
2831    
2832        (b) Functions such as _pcre_xclass(), which are internal to the library,
2833            but external in the C sense, all had PCRE_EXPORT in their definitions.
2834            This is apparently wrong for the Windows case, so I have removed it.
2835            (It makes no difference on Unix-like systems.)
2836    
2837    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2838        of recursive calls to match(). This is different to MATCH_LIMIT because
2839        that limits the total number of calls to match(), not all of which increase
2840        the depth of recursion. Limiting the recursion depth limits the amount of
2841        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2842        when PCRE is compiled, and changed at run time. A patch from Google adds
2843        this functionality to the C++ interface.
2844    
2845    18. Changes to the handling of Unicode character properties:
2846    
2847        (a) Updated the table to Unicode 4.1.0.
2848    
2849        (b) Recognize characters that are not in the table as "Cn" (undefined).
2850    
2851        (c) I revised the way the table is implemented to a much improved format
2852            which includes recognition of ranges. It now supports the ranges that
2853            are defined in UnicodeData.txt, and it also amalgamates other
2854            characters into ranges. This has reduced the number of entries in the
2855            table from around 16,000 to around 3,000, thus reducing its size
2856            considerably. I realized I did not need to use a tree structure after
2857            all - a binary chop search is just as efficient. Having reduced the
2858            number of entries, I extended their size from 6 bytes to 8 bytes to
2859            allow for more data.
2860    
2861        (d) Added support for Unicode script names via properties such as \p{Han}.
2862    
2863    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2864        matching that character.
2865    
2866    20. When matching a repeated Unicode property with a minimum greater than zero,
2867        (for example \pL{2,}), PCRE could look past the end of the subject if it
2868        reached it while seeking the minimum number of characters. This could
2869        happen only if some of the characters were more than one byte long, because
2870        there is a check for at least the minimum number of bytes.
2871    
2872    21. Refactored the implementation of \p and \P so as to be more general, to
2873        allow for more different types of property in future. This has changed the
2874        compiled form incompatibly. Anybody with saved compiled patterns that use
2875        \p or \P will have to recompile them.
2876    
2877    22. Added "Any" and "L&" to the supported property types.
2878    
2879    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2880        but give a compile time error if the value is greater than 0xff.
2881    
2882    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2883        accidentally not being installed or uninstalled.
2884    
2885    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2886        made were to insert the current release number. This seemed silly, because
2887        it made things harder for people building PCRE on systems that don't run
2888        "configure". I have turned pcre.h into a distributed file, no longer built
2889        by "configure", with the version identification directly included. There is
2890        no longer a pcre.h.in file.
2891    
2892        However, this change necessitated a change to the pcre-config script as
2893        well. It is built from pcre-config.in, and one of the substitutions was the
2894        release number. I have updated configure.ac so that ./configure now finds
2895        the release number by grepping pcre.h.
2896    
2897    26. Added the ability to run the tests under valgrind.
2898    
2899    
2900    Version 6.4 05-Sep-05
2901    ---------------------
2902    
2903     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2904        "--" to be printed when multiple files were scanned, even when none of the
2905        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2906        consider it to be a bug, and have restored the previous behaviour.
2907    
2908     2. A couple of code tidies to get rid of compiler warnings.
2909    
2910     3. The pcretest program used to cheat by referring to symbols in the library
2911        whose names begin with _pcre_. These are internal symbols that are not
2912        really supposed to be visible externally, and in some environments it is
2913        possible to suppress them. The cheating is now confined to including
2914        certain files from the library's source, which is a bit cleaner.
2915    
2916     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2917        file's purpose clearer.
2918    
2919     5. Reorganized pcre_ucp_findchar().
2920    
2921    
2922    Version 6.3 15-Aug-05
2923    ---------------------
2924    
2925     1. The file libpcre.pc.in did not have general read permission in the tarball.
2926    
2927     2. There were some problems when building without C++ support:
2928    
2929        (a) If C++ support was not built, "make install" and "make test" still
2930            tried to test it.
2931    
2932        (b) There were problems when the value of CXX was explicitly set. Some
2933            changes have been made to try to fix these, and ...
2934    
2935        (c) --disable-cpp can now be used to explicitly disable C++ support.
2936    
2937        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2938            backslash in a target when C++ was disabled. This confuses some
2939            versions of "make", apparently. Using an intermediate variable solves
2940            this. (Same for CPP_LOBJ.)
2941    
2942     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2943        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2944        necessary on certain architectures.
2945    
2946     4. Added a setting of -export-symbols-regex to the link command to remove
2947        those symbols that are exported in the C sense, but actually are local
2948        within the library, and not documented. Their names all begin with
2949        "_pcre_". This is not a perfect job, because (a) we have to except some
2950        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2951        available (and never for static libraries). I have made a note to try to
2952        find a way round (a) in the future.
2953    
2954    
2955    Version 6.2 01-Aug-05
2956    ---------------------
2957    
2958     1. There was no test for integer overflow of quantifier values. A construction
2959        such as {1111111111111111} would give undefined results. What is worse, if
2960        a minimum quantifier for a parenthesized subpattern overflowed and became
2961        negative, the calculation of the memory size went wrong. This could have
2962        led to memory overwriting.
2963    
2964     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2965    
2966     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2967        operating environments where this matters.
2968    
2969     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2970        PCRE options from within the C++ wrapper.
2971    
2972     5. Named capturing subpatterns were not being correctly counted when a pattern
2973        was compiled. This caused two problems: (a) If there were more than 100
2974        such subpatterns, the calculation of the memory needed for the whole
2975        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2976        back references of the form \12, where the number was greater than 9, were
2977        not recognized as back references, even though there were sufficient
2978        previous subpatterns.
2979    
2980     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2981        versions of gcc, e.g. 2.95.4.
2982    
2983    
2984    Version 6.1 21-Jun-05
2985    ---------------------
2986    
2987     1. There was one reference to the variable "posix" in pcretest.c that was not
2988        surrounded by "#if !defined NOPOSIX".
2989    
2990     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2991        the cross-check on the old pcre_info() function, for the benefit of the
2992        cut-down version of PCRE that is currently imported into Exim.
2993    
2994     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2995        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2996        patterns that start (?i) but I don't think that matters. The use of (?i) is
2997        just an example; this all applies to the other options as well.
2998    
2999     4. Since libtool seems to echo the compile commands it is issuing, the output
3000        from "make" can be reduced a bit by putting "@" in front of each libtool
3001        compile command.
3002    
3003     5. Patch from the folks at Google for configure.in to be a bit more thorough
3004        in checking for a suitable C++ installation before trying to compile the
3005        C++ stuff. This should fix a reported problem when a compiler was present,
3006        but no suitable headers.
3007    
3008     6. The man pages all had just "PCRE" as their title. I have changed them to
3009        be the relevant file name. I have also arranged that these names are
3010        retained in the file doc/pcre.txt, which is a concatenation in text format
3011        of all the man pages except the little individual ones for each function.
3012    
3013     7. The NON-UNIX-USE file had not been updated for the different set of source
3014        files that come with release 6. I also added a few comments about the C++
3015        wrapper.
3016    
3017    
3018    Version 6.0 07-Jun-05
3019    ---------------------
3020    
3021     1. Some minor internal re-organization to help with my DFA experiments.
3022    
3023     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3024        didn't matter for the library itself when fully configured, but did matter
3025        when compiling without UCP support, or within Exim, where the ucp files are
3026        not imported.
3027    
3028     3. Refactoring of the library code to split up the various functions into
3029        different source modules. The addition of the new DFA matching code (see
3030        below) to a single monolithic source would have made it really too
3031        unwieldy, quite apart from causing all the code to be include in a
3032        statically linked application, when only some functions are used. This is
3033        relevant even without the DFA addition now that patterns can be compiled in
3034        one application and matched in another.
3035    
3036        The downside of splitting up is that there have to be some external
3037        functions and data tables that are used internally in different modules of
3038        the library but which are not part of the API. These have all had their
3039        names changed to start with "_pcre_" so that they are unlikely to clash
3040        with other external names.
3041    
3042     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3043        a different (DFA) algorithm. Although it is slower than the original
3044        function, it does have some advantages for certain types of matching
3045        problem.
3046    
3047     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3048        including restarting after a partial match.
3049    
3050     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3051        defined when compiling for Windows was sent to me. I have put it into the
3052        code, though I have no means of testing or verifying it.
3053    
3054     7. Added the pcre_refcount() auxiliary function.
3055    
3056     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3057        match before or at the first newline in the subject string. In pcretest,
3058        the /f option on a pattern can be used to set this.
3059    
3060     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3061        would behave wrongly. This has been present in PCRE since release 4.0.
3062    
3063    10. A number of changes to the pcregrep command:
3064    
3065        (a) Refactored how -x works; insert ^(...)$ instead of setting
3066            PCRE_ANCHORED and checking the length, in preparation for adding
3067            something similar for -w.
3068    
3069        (b) Added the -w (match as a word) option.
3070    
3071        (c) Refactored the way lines are read and buffered so as to have more
3072            than one at a time available.
3073    
3074        (d) Implemented a pcregrep test script.
3075    
3076        (e) Added the -M (multiline match) option. This allows patterns to match
3077            over several lines of the subject. The buffering ensures that at least
3078            8K, or the rest of the document (whichever is the shorter) is available
3079            for matching (and similarly the previous 8K for lookbehind assertions).
3080    
3081        (f) Changed the --help output so that it now says
3082    
3083              -w, --word-regex(p)
3084    
3085            instead of two lines, one with "regex" and the other with "regexp"
3086            because that confused at least one person since the short forms are the
3087            same. (This required a bit of code, as the output is generated
3088            automatically from a table. It wasn't just a text change.)
3089    
3090        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3091            option but starts with a hyphen. Could be a pattern or a path name
3092            starting with a hyphen, for instance.
3093    
3094        (h) "-" can be given as a file name to represent stdin.
3095    
3096        (i) When file names are being printed, "(standard input)" is used for
3097            the standard input, for compatibility with GNU grep. Previously
3098            "<stdin>" was used.
3099    
3100        (j) The option --label=xxx can be used to supply a name to be used for
3101            stdin when file names are being printed. There is no short form.
3102    
3103        (k) Re-factored the options decoding logic because we are going to add
3104            two more options that take data. Such options can now be given in four
3105            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3106    
3107        (l) Added the -A, -B, and -C options for requesting that lines of context
3108            around matches be printed.
3109    
3110        (m) Added the -L option to print the names of files that do not contain
3111            any matching lines, that is, the complement of -l.
3112    
3113        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3114            continue to scan other files.
3115    
3116        (o) The -s option was incorrectly implemented. For compatibility with other
3117            greps, it now suppresses the error message for a non-existent or non-
3118            accessible file (but not the return code). There is a new option called
3119            -q that suppresses the output of matching lines, which was what -s was
3120            previously doing.
3121    
3122        (p) Added --include and --exclude options to specify files for inclusion
3123            and exclusion when recursing.
3124    
3125    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3126        Hopefully, it now does.
3127    
3128    12. Missing cast in pcre_study().
3129    
3130    13. Added an "uninstall" target to the makefile.
3131    
3132    14. Replaced "extern" in the function prototypes in Makefile.in with
3133        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3134        world, but is set differently for Windows.
3135    
3136    15. Added a second compiling function called pcre_compile2(). The only
3137        difference is that it has an extra argument, which is a pointer to an
3138        integer error code. When there is a compile-time failure, this is set
3139        non-zero, in addition to the error test pointer being set to point to an
3140        error message. The new argument may be NULL if no error number is required
3141        (but then you may as well call pcre_compile(), which is now just a
3142        wrapper). This facility is provided because some applications need a
3143        numeric error indication, but it has also enabled me to tidy up the way
3144        compile-time errors are handled in the POSIX wrapper.
3145    
3146    16. Added VPATH=.libs to the makefile; this should help when building with one
3147        prefix path and installing with another. (Or so I'm told by someone who
3148        knows more about this stuff than I do.)
3149    
3150    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3151        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3152        match everything, including newlines. This is not POSIX-compatible, but
3153        somebody wanted the feature. From pcretest it can be activated by using
3154        both the P and the s flags.
3155    
3156    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3157    
3158    19. libpcre.pc was being incorrectly installed as executable.
3159    
3160    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3161        it now also looks for '\r' so that it will work unmodified on Windows.
3162    
3163    21. Added Google's contributed C++ wrapper to the distribution.
3164    
3165    22. Added some untidy missing memory free() calls in pcretest, to keep
3166        Electric Fence happy when testing.
3167    
3168    
3169    
3170    Version 5.0 13-Sep-04
3171    ---------------------
3172    
3173     1. Internal change: literal characters are no longer packed up into items
3174        containing multiple characters in a single byte-string. Each character
3175        is now matched using a separate opcode. However, there may be more than one
3176        byte in the character in UTF-8 mode.
3177    
3178     2. The pcre_callout_block structure has two new fields: pattern_position and
3179        next_item_length. These contain the offset in the pattern to the next match
3180        item, and its length, respectively.
3181    
3182     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3183        insertion of callouts before each pattern item. Added the /C option to
3184        pcretest to make use of this.
3185    
3186     4. On the advice of a Windows user, the lines
3187    
3188          #if defined(_WIN32) || defined(WIN32)
3189          _setmode( _fileno( stdout ), 0x8000 );
3190          #endif  /* defined(_WIN32) || defined(WIN32) */
3191    
3192        have been added to the source of pcretest. This apparently does useful
3193        magic in relation to line terminators.
3194    
3195     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3196        for the benefit of those environments where the "b" makes a difference.
3197    
3198     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3199        to know about it. I have put a hack into configure.in that adds in code
3200        to set GCC=yes if CC=icc. This seems to end up at a point in the
3201        generated configure script that is early enough to affect the setting of
3202        compiler options, which is what is needed, but I have no means of testing
3203        whether it really works. (The user who reported this had patched the
3204        generated configure script, which of course I cannot do.)
3205    
3206        LATER: After change 22 below (new libtool files), the configure script
3207        seems to know about icc (and also ecc). Therefore, I have commented out
3208        this hack in configure.in.
3209    
3210     7. Added support for pkg-config (2 patches were sent in).
3211    
3212     8. Negated POSIX character classes that used a combination of internal tables
3213        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3214        [[:^ascii]]. Typically, they would match almost any characters. The other
3215        POSIX classes were not broken in this way.
3216    
3217     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3218        to find the match, as PCRE was deluded into thinking that the match had to
3219        start at the start point or following a newline. The same bug applied to
3220        patterns with negative forward assertions or any backward assertions
3221        preceding ".*" at the start, unless the pattern required a fixed first
3222        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3223    
3224    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3225        starting at the last subject character, bytes beyond the end of the subject
3226        string were read.
3227    
3228    11. Renamed the variable "class" as "classbits" to make life easier for C++
3229        users. (Previously there was a macro definition, but it apparently wasn't
3230        enough.)
3231    
3232    12. Added the new field "tables" to the extra data so that tables can be passed
3233        in at exec time, or the internal tables can be re-selected. This allows
3234        a compiled regex to be saved and re-used at a later time by a different
3235        program that might have everything at different addresses.
3236    
3237    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3238        -R library as well as a -L library.
3239    
3240    14. The debugging options of pcretest (-d on the command line or D on a
3241        pattern) showed incorrect output for anything following an extended class
3242        that contained multibyte characters and which was followed by a quantifier.
3243    
3244    15. Added optional support for general category Unicode character properties
3245        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3246        support. It adds about 90K to the size of the library. The meanings of the
3247        inbuilt class escapes such as \d and \s have NOT been changed.
3248    
3249    16. Updated pcredemo.c to include calls to free() to release the memory for the
3250        compiled pattern.
3251    
3252    17. The generated file chartables.c was being created in the source directory
3253        instead of in the building directory. This caused the build to fail if the
3254        source directory was different from the building directory, and was
3255        read-only.
3256    
3257    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3258        file. No doubt somebody will tell me if they don't make sense... Also added
3259        Dan Mooney's comments about building on OpenVMS.
3260    
3261    19. Added support for partial matching via the PCRE_PARTIAL option for
3262        pcre_exec() and the \P data escape in pcretest.
3263    
3264    20. Extended pcretest with 3 new pattern features:
3265    
3266        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3267              write the compiled pattern to the file whose name is "rest-of-line".
3268              This is a straight binary dump of the data, with the saved pointer to
3269              the character tables forced to be NULL. The study data, if any, is
3270              written too. After writing, pcretest reads a new pattern.
3271    
3272        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3273              compiled pattern from the given file. There must not be any
3274              occurrences of "<" in the file name (pretty unlikely); if there are,
3275              pcretest will instead treat the initial "<" as a pattern delimiter.
3276              After reading in the pattern, pcretest goes on to read data lines as
3277              usual.
3278    
3279        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3280              and 16-bit fields in a compiled pattern, to simulate a pattern that
3281              was compiled on a host of opposite endianness.
3282    
3283    21. The pcre-exec() function can now cope with patterns that were compiled on
3284        hosts of opposite endianness, with this restriction:
3285    
3286          As for any compiled expression that is saved and used later, the tables
3287          pointer field cannot be preserved; the extra_data field in the arguments
3288          to pcre_exec() should be used to pass in a tables address if a value
3289          other than the default internal tables were used at compile time.
3290    
3291    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3292        now diagnosed as an error. Previously, most of the time, a negative number
3293        would have been treated as zero, but if in addition "ovector" was passed as
3294        NULL, a crash could occur.
3295    
3296    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3297        new versions from the libtool 1.5 distribution (the last one is a copy of
3298        a file called libtool.m4). This seems to have fixed the need to patch
3299        "configure" to support Darwin 1.3 (which I used to do). However, I still
3300        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3301        workstation).
3302    
3303    24. Changed the PCRE licence to be the more standard "BSD" licence.
3304    
3305    
3306    Version 4.5 01-Dec-03
3307    ---------------------
3308    
3309     1. There has been some re-arrangement of the code for the match() function so
3310        that it can be compiled in a version that does not call itself recursively.
3311        Instead, it keeps those local variables that need separate instances for
3312        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3313        needs to "recurse". Keeping track of where control must go is done by means
3314        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3315        hide most of the details from the main code, and operates only if
3316        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3317        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3318        operating.
3319    
3320        To make it easier for callers to provide specially tailored get/free
3321        functions for this usage, two new functions, pcre_stack_malloc, and
3322        pcre_stack_free, are used. They are always called in strict stacking order,
3323        and the size of block requested is always the same.
3324    
3325        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3326        PCRE has been compiled to use the stack or the heap for recursion. The
3327        -C option of pcretest uses this to show which version is compiled.
3328    
3329        A new data escape \S, is added to pcretest; it causes the amounts of store
3330        obtained and freed by both kinds of malloc/free at match time to be added
3331        to the output.
3332    
3333     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3334        what's available on my current Linux desktop machine.
3335    
3336     3. When matching a UTF-8 string, the test for a valid string at the start has
3337        been extended. If start_offset is not zero, PCRE now checks that it points
3338        to a byte that is the start of a UTF-8 character. If not, it returns
3339        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3340        this is necessary because there may be backward assertions in the pattern.
3341        When matching the same subject several times, it may save resources to use
3342        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3343    
3344     4. The code for checking the validity of UTF-8 strings has been tightened so
3345        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3346        containing "overlong sequences".
3347    
3348     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3349        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3350        should have been "&", but it just so happened that all the cases this let
3351        through by mistake were picked up later in the function.
3352    
3353     6. I had used a variable called "isblank" - this is a C99 function, causing
3354        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3355    
3356     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3357        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3358        so the version is shown; (c) stop "make test" echoing "./RunTest".
3359    
3360     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3361    
3362     9. The prototype for memmove() for systems that don't have it was using
3363        size_t, but the inclusion of the header that defines size_t was later. I've
3364        moved the #includes for the C headers earlier to avoid this.
3365    
3366    10. Added some adjustments to the code to make it easier to compiler on certain
3367        special systems:
3368    
3369          (a) Some "const" qualifiers were missing.
3370          (b) Added the macro EXPORT before all exported functions; by default this
3371              is defined to be empty.
3372          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3373              that it reads its output file name as an argument instead of writing
3374              to the standard output and assuming this can be redirected.
3375    
3376    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3377        class containing characters with values greater than 255, PCRE compilation
3378        went into a loop.
3379    
3380    12. A recursive reference to a subpattern that was within another subpattern
3381        that had a minimum quantifier of zero caused PCRE to crash. For example,
3382        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3383        recursion. If the recursively-called subpattern itself had a zero repeat,
3384        that was OK.
3385    
3386    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3387        buffer into which it was copied (for escape processing) was still set at
3388        1024, so long lines caused crashes.
3389    
3390    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3391        "internal error: code overflow...". This applied to any character class
3392        that was followed by a possessive quantifier.
3393    
3394    15. Modified the Makefile to add libpcre.la as a prerequisite for
3395        libpcreposix.la because I was told this is needed for a parallel build to
3396        work.
3397    
3398    16. If a pattern that contained .* following optional items at the start was
3399        studied, the wrong optimizing data was generated, leading to matching
3400        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3401        matching string must start with a or b or c. The correct conclusion for
3402        this pattern is that a match can start with any character.
3403    
3404    
3405    Version 4.4 13-Aug-03
3406    ---------------------
3407    
3408     1. In UTF-8 mode, a character class containing characters with values between
3409        127 and 255 was not handled correctly if the compiled pattern was studied.
3410        In fixing this, I have also improved the studying algorithm for such
3411        classes (slightly).
3412    
3413     2. Three internal functions had redundant arguments passed to them. Removal
3414        might give a very teeny performance improvement.
3415    
3416     3. Documentation bug: the value of the capture_top field in a callout is *one
3417        more than* the number of the hightest numbered captured substring.
3418    
3419     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3420        in incorrectly linking with a previously installed version. They now link
3421        explicitly with libpcre.la.
3422    
3423     5. configure.in no longer needs to recognize Cygwin specially.
3424    
3425     6. A problem in pcre.in for Windows platforms is fixed.
3426    
3427     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3428        pcretest, it used to include the size of the study block as part of its
3429        output. Unfortunately, the structure contains a field that has a different
3430        size on different hardware architectures. This meant that the tests that
3431        showed this size failed. As the block is currently always of a fixed size,
3432        this information isn't actually particularly useful in pcretest output, so
3433        I have just removed it.
3434    
3435     8. Three pre-processor statements accidentally did not start in column 1.
3436        Sadly, there are *still* compilers around that complain, even though
3437        standard C has not required this for well over a decade. Sigh.
3438    
3439     9. In pcretest, the code for checking callouts passed small integers in the
3440        callout_data field, which is a void * field. However, some picky compilers
3441        complained about the casts involved for this on 64-bit systems. Now
3442        pcretest passes the address of the small integer instead, which should get
3443        rid of the warnings.
3444    
3445    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3446        both compile and run time, and gives an error if an invalid UTF-8 sequence
3447        is found. There is a option for disabling this check in cases where the
3448        string is known to be correct and/or the maximum performance is wanted.
3449    
3450    11. In response to a bug report, I changed one line in Makefile.in from
3451    
3452            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3453        to
3454            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3455    
3456        to look similar to other lines, but I have no way of telling whether this
3457        is the right thing to do, as I do not use Windows. No doubt I'll get told
3458        if it's wrong...
3459    
3460    
3461    Version 4.3 21-May-03
3462    ---------------------
3463    
3464    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3465       Makefile.
3466    
3467    2. Some refactoring to improve the quality of the code:
3468    
3469       (i)   The utf8_table... variables are now declared "const".
3470    
3471       (ii)  The code for \cx, which used the "case flipping" table to upper case
3472             lower case letters, now just substracts 32. This is ASCII-specific,
3473             but the whole concept of \cx is ASCII-specific, so it seems
3474             reasonable.
3475    
3476       (iii) PCRE was using its character types table to recognize decimal and
3477             hexadecimal digits in the pattern. This is silly, because it handles
3478             only 0-9, a-f, and A-F, but the character types table is locale-
3479             specific, which means strange things might happen. A private
3480             table is now used for this - though it costs 256 bytes, a table is
3481             much faster than multiple explicit tests. Of course, the standard
3482             character types table is still used for matching digits in subject
3483             strings against \d.
3484    
3485       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3486             ending in _t are). So I've renamed it as ESC_tee.
3487    
3488    3. The first argument for regexec() in the POSIX wrapper should have been
3489       defined as "const".
3490    
3491    4. Changed pcretest to use malloc() for its buffers so that they can be
3492       Electric Fenced for debugging.
3493    
3494    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3495       to read one or more bytes before the start of the subject string. Often this
3496       had no effect on PCRE's behaviour, but in some circumstances it could
3497       provoke a segmentation fault.
3498    
3499    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3500       to try to read one or more bytes before the start of the subject string.
3501    
3502    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3503       UTF-8 support could misbehave in various ways if the subject string
3504       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3505       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3506       back over UTF-8 characters.)
3507    
3508    
3509    Version 4.2 14-Apr-03
3510    ---------------------
3511    
3512    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3513    
3514    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3515         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3516         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3517         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3518         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3519           and BUILD_EXEEXT
3520         Note: automatic setting of the BUILD variables is not yet working
3521         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3522           compile-time but not at link-time
3523         [LINK]: use for linking executables only
3524         make different versions for Windows and non-Windows
3525         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3526           libraries
3527         [LINK_FOR_BUILD]: new variable
3528         [OBJEXT]: use throughout
3529         [EXEEXT]: use throughout
3530         <winshared>: new target
3531         <wininstall>: new target
3532         <dftables.o>: use native compiler
3533         <dftables>: use native linker
3534         <install>: handle Windows platform correctly
3535         <clean>: ditto
3536         <check>: ditto
3537         copy DLL to top builddir before testing
3538    
3539       As part of these changes, -no-undefined was removed again. This was reported