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

Diff of /code/trunk/ChangeLog

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

revision 67 by nigel, Sat Feb 24 21:40:13 2007 UTC revision 964 by ph10, Fri May 4 13:03:39 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    9.  (*MARK) control verb is now supported by the JIT compiler.
46    
47    10. The command "./RunTest list" lists the available tests without actually
48        running any of them. (Because I keep forgetting what they all are.)
49    
50    11. Add PCRE_INFO_MAXLOOKBEHIND.
51    
52    12. Applied a (slightly modified) user-supplied patch that improves performance
53        when the heap is used for recursion (compiled with --disable-stack-for-
54        recursion). Instead of malloc and free for each heap frame each time a
55        logical recursion happens, frames are retained on a chain and re-used where
56        possible. This sometimes gives as much as 30% improvement.
57    
58    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59        call.
60    
61    14. As documented, (*COMMIT) is now confined to within a positive assertion.
62    
63    15. It is now possible to link pcretest with libedit as an alternative to
64        libreadline.
65    
66    16. (*COMMIT) control verb is now supported by the JIT compiler.
67    
68    17. The Unicode data tables have been updated to Unicode 6.1.0.
69    
70    18. Added --file-list option to pcregrep.
71    
72    19. Added binary file support to pcregrep, including the -a, --binary-files,
73        -I, and --text options.
74    
75    20. The madvise function is renamed for posix_madvise for QNX compatibility
76        reasons. Fixed by Giuseppe D'Angelo.
77    
78    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79    
80    22. Changed the option for creating symbolic links for 16-bit man pages from
81        -s to -sf so that re-installing does not cause issues.
82    
83    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84    
85    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86        matches in certain environments (the workspace was not being correctly
87        retained). Also added to pcre_dfa_exec() a simple plausibility check on
88        some of the workspace data at the beginning of a restart.
89    
90    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91        was not doing so when it should - probably a typo introduced by SVN 528
92        (change 8.10/14).
93    
94    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    
115    Version 8.30 04-February-2012
116    -----------------------------
117    
118    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
119        name is defined in ctype.h.
120    
121    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
122        only in quite long subpatterns.
123    
124    3.  Removed the function pcre_info(), which has been obsolete and deprecated
125        since it was replaced by pcre_fullinfo() in February 2000.
126    
127    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
128        match a (*MARK), and the match failed at the start of the subject, a
129        reference to memory before the start of the subject could occur. This bug
130        was introduced by fix 17 of release 8.21.
131    
132    5.  A reference to an unset group with zero minimum repetition was giving
133        totally wrong answers (in non-JavaScript-compatibility mode). For example,
134        /(another)?(\1?)test/ matched against "hello world test". This bug was
135        introduced in release 8.13.
136    
137    6.  Add support for 16-bit character strings (a large amount of work involving
138        many changes and refactorings).
139    
140    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
141        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
142        from a file.
143    
144    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
145        rounding is not applied in this particular case).
146    
147    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
148        if they appear, or are escaped, in patterns.
149    
150    10. Get rid of a number of -Wunused-but-set-variable warnings.
151    
152    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
153        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
154        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
155        also returns the mark "x". This bug applied to capturing parentheses,
156        non-capturing parentheses, and atomic parentheses. It also applied to some
157        assertions.
158    
159    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
160        information out of configure.ac instead of relying on pcre.h.generic, which
161        is not stored in the repository.
162    
163    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
164        -lreadline.
165    
166    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
167    
168    15. Applied Graycode's patch to put the top-level frame on the stack rather
169        than the heap when not using the stack for recursion. This gives a
170        performance improvement in many cases when recursion is not deep.
171    
172    16. Experimental code added to "pcretest -C" to output the stack frame size.
173    
174    
175    Version 8.21 12-Dec-2011
176    ------------------------
177    
178    1.  Updating the JIT compiler.
179    
180    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
181        are added as well.
182    
183    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
184        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
185        calling _pcre_jit_exec. Some extra comments are added.
186    
187    4.  (*MARK) settings inside atomic groups that do not contain any capturing
188        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
189        was introduced by change 18 for 8.20.
190    
191    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
192        ECMA-262 standard.
193    
194    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
195        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
196        This bug was probably introduced by change 9 of 8.13.
197    
198    7.  While fixing 6 above, I noticed that a number of other items were being
199        incorrectly rejected as "not fixed length". This arose partly because newer
200        opcodes had not been added to the fixed-length checking code. I have (a)
201        corrected the bug and added tests for these items, and (b) arranged for an
202        error to occur if an unknown opcode is encountered while checking for fixed
203        length instead of just assuming "not fixed length". The items that were
204        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
205        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
206        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
207    
208    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
209        being incorrectly compiled and would have given unpredicatble results.
210    
211    9.  A possessively repeated subpattern with minimum repeat count greater than
212        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
213        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
214        the first (A) could occur when it should not.
215    
216    10. Add a cast and remove a redundant test from the code.
217    
218    11. JIT should use pcre_malloc/pcre_free for allocation.
219    
220    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
221        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
222        is anything other than /usr, -L is still shown).
223    
224    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
225    
226    14. Perl does not support \N without a following name in a [] class; PCRE now
227        also gives an error.
228    
229    15. If a forward reference was repeated with an upper limit of around 2000,
230        it caused the error "internal error: overran compiling workspace". The
231        maximum number of forward references (including repeats) was limited by the
232        internal workspace, and dependent on the LINK_SIZE. The code has been
233        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
234        the default depends on LINK_SIZE. There is a new upper limit (for safety)
235        of around 200,000 forward references. While doing this, I also speeded up
236        the filling in of repeated forward references.
237    
238    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
239        incorrectly expecting the subject to contain another "a" after the start.
240    
241    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
242        in the match, the SKIP should be ignored. This was not happening; instead
243        the SKIP was being treated as NOMATCH. For patterns such as
244        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
245        tested.
246    
247    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
248        now much more compatible with Perl, in particular in cases where the result
249        is a non-match for a non-anchored pattern. For example, if
250        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
251        "m", where previously it did not return a name. A side effect of this
252        change is that for partial matches, the last encountered mark name is
253        returned, as for non matches. A number of tests that were previously not
254        Perl-compatible have been moved into the Perl-compatible test files. The
255        refactoring has had the pleasing side effect of removing one argument from
256        the match() function, thus reducing its stack requirements.
257    
258    19. If the /S+ option was used in pcretest to study a pattern using JIT,
259        subsequent uses of /S (without +) incorrectly behaved like /S+.
260    
261    21. Retrieve executable code size support for the JIT compiler and fixing
262        some warnings.
263    
264    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
265        not work when the shorter character appeared right at the end of the
266        subject string.
267    
268    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
269        systems.
270    
271    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
272        output it when the /M option is used in pcretest.
273    
274    25. The CheckMan script was not being included in the distribution. Also, added
275        an explicit "perl" to run Perl scripts from the PrepareRelease script
276        because this is reportedly needed in Windows.
277    
278    26. If study data was being save in a file and studying had not found a set of
279        "starts with" bytes for the pattern, the data written to the file (though
280        never used) was taken from uninitialized memory and so caused valgrind to
281        complain.
282    
283    27. Updated RunTest.bat as provided by Sheri Pierce.
284    
285    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
286    
287    29. Computation of memory usage for the table of capturing group names was
288        giving an unnecessarily large value.
289    
290    
291    Version 8.20 21-Oct-2011
292    ------------------------
293    
294    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
295        a POSIX class. After further experiments with Perl, which convinced me that
296        Perl has bugs and confusions, a closing square bracket is no longer allowed
297        in a POSIX name. This bug also affected patterns with classes that started
298        with full stops.
299    
300    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
301        captured substring, but while checking the failing first alternative,
302        substring 1 is temporarily captured. If the output vector supplied to
303        pcre_exec() was not big enough for this capture, the yield of the function
304        was still zero ("insufficient space for captured substrings"). This cannot
305        be totally fixed without adding another stack variable, which seems a lot
306        of expense for a edge case. However, I have improved the situation in cases
307        such as /(a)(b)x|abc/ matched against "abc", where the return code
308        indicates that fewer than the maximum number of slots in the ovector have
309        been set.
310    
311    3.  Related to (2) above: when there are more back references in a pattern than
312        slots in the output vector, pcre_exec() uses temporary memory during
313        matching, and copies in the captures as far as possible afterwards. It was
314        using the entire output vector, but this conflicts with the specification
315        that only 2/3 is used for passing back captured substrings. Now it uses
316        only the first 2/3, for compatibility. This is, of course, another edge
317        case.
318    
319    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
320        main code base, and can be used by building with --enable-jit. When this is
321        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
322        runtime --no-jit option is given.
323    
324    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
325        ovector, the return from the function was zero, implying that there were
326        other matches that did not fit. The correct "exactly full" value is now
327        returned.
328    
329    6.  If a subpattern that was called recursively or as a subroutine contained
330        (*PRUNE) or any other control that caused it to give a non-standard return,
331        invalid errors such as "Error -26 (nested recursion at the same subject
332        position)" or even infinite loops could occur.
333    
334    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
335        computing the minimum length on reaching *ACCEPT, and so ended up with the
336        wrong value of 1 rather than 0. Further investigation indicates that
337        computing a minimum subject length in the presence of *ACCEPT is difficult
338        (think back references, subroutine calls), and so I have changed the code
339        so that no minimum is registered for a pattern that contains *ACCEPT.
340    
341    8.  If (*THEN) was present in the first (true) branch of a conditional group,
342        it was not handled as intended. [But see 16 below.]
343    
344    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
345        Sheri Pierce.
346    
347    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
348        the first byte in a match must be "a".
349    
350    11. Change 17 for 8.13 increased the recursion depth for patterns like
351        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
352        pattern contains any instances of (*THEN). If it does not, the old
353        optimizations are restored. It would be nice to do this on a per-group
354        basis, but at the moment that is not feasible.
355    
356    12. In some environments, the output of pcretest -C is CRLF terminated. This
357        broke RunTest's code that checks for the link size. A single white space
358        character after the value is now allowed for.
359    
360    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
361        For "fr", it uses the Windows-specific input and output files.
362    
363    14. If (*THEN) appeared in a group that was called recursively or as a
364        subroutine, it did not work as intended. [But see next item.]
365    
366    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
367        pattern fragments (but not containing any | characters). If A and B are
368        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
369        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
370        D. In other words, Perl considers parentheses that do not contain any |
371        characters to be part of a surrounding alternative, whereas PCRE was
372        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
373        differently. PCRE now behaves in the same way as Perl, except in the case
374        of subroutine/recursion calls such as (?1) which have in any case always
375        been different (but PCRE had them first :-).
376    
377    16. Related to 15 above: Perl does not treat the | in a conditional group as
378        creating alternatives. Such a group is treated in the same way as an
379        ordinary group without any | characters when processing (*THEN). PCRE has
380        been changed to match Perl's behaviour.
381    
382    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
383        RunGrepTest script failed.
384    
385    18. Change 22 for version 13 caused atomic groups to use more stack. This is
386        inevitable for groups that contain captures, but it can lead to a lot of
387        stack use in large patterns. The old behaviour has been restored for atomic
388        groups that do not contain any capturing parentheses.
389    
390    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
391        suppress the check for a minimum subject length at run time. (If it was
392        given to pcre_exec() or pcre_dfa_exec() it did work.)
393    
394    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
395        fail to work when decoding hex characters in data strings in EBCDIC
396        environments.
397    
398    21. It appears that in at least one Mac OS environment, the isxdigit() function
399        is implemented as a macro that evaluates to its argument more than once,
400        contravening the C 90 Standard (I haven't checked a later standard). There
401        was an instance in pcretest which caused it to go wrong when processing
402        \x{...} escapes in subject strings. The has been rewritten to avoid using
403        things like p++ in the argument of isxdigit().
404    
405    
406    Version 8.13 16-Aug-2011
407    ------------------------
408    
409    1.  The Unicode data tables have been updated to Unicode 6.0.0.
410    
411    2.  Two minor typos in pcre_internal.h have been fixed.
412    
413    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
414        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
415        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
416    
417    4.  There were a number of related bugs in the code for matching backrefences
418        caselessly in UTF-8 mode when codes for the characters concerned were
419        different numbers of bytes. For example, U+023A and U+2C65 are an upper
420        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
421        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
422        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
423        2-byte code at the end of the subject (it thought there wasn't enough data
424        left).
425    
426    5.  Comprehensive information about what went wrong is now returned by
427        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
428        as the output vector has at least 2 elements. The offset of the start of
429        the failing character and a reason code are placed in the vector.
430    
431    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
432        now returned is for the first byte of the failing character, instead of the
433        last byte inspected. This is an incompatible change, but I hope it is small
434        enough not to be a problem. It makes the returned offset consistent with
435        pcre_exec() and pcre_dfa_exec().
436    
437    7.  pcretest now gives a text phrase as well as the error number when
438        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
439        failure, the offset and reason code are output.
440    
441    8.  When \R was used with a maximizing quantifier it failed to skip backwards
442        over a \r\n pair if the subsequent match failed. Instead, it just skipped
443        back over a single character (\n). This seems wrong (because it treated the
444        two characters as a single entity when going forwards), conflicts with the
445        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
446        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
447        has been changed.
448    
449    9.  Some internal refactoring has changed the processing so that the handling
450        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
451        time (the PCRE_DOTALL option was changed this way some time ago: version
452        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
453        which was always a bit of a fudge. It also means that there is one less
454        argument for the match() function, which reduces its stack requirements
455        slightly. This change also fixes an incompatibility with Perl: the pattern
456        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
457    
458    10. More internal refactoring has drastically reduced the number of recursive
459        calls to match() for possessively repeated groups such as (abc)++ when
460        using pcre_exec().
461    
462    11. While implementing 10, a number of bugs in the handling of groups were
463        discovered and fixed:
464    
465        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
466        (a|)*(?1) gave a compile-time internal error.
467        ((a|)+)+  did not notice that the outer group could match an empty string.
468        (^a|^)+   was not marked as anchored.
469        (.*a|.*)+ was not marked as matching at start or after a newline.
470    
471    12. Yet more internal refactoring has removed another argument from the match()
472        function. Special calls to this function are now indicated by setting a
473        value in a variable in the "match data" data block.
474    
475    13. Be more explicit in pcre_study() instead of relying on "default" for
476        opcodes that mean there is no starting character; this means that when new
477        ones are added and accidentally left out of pcre_study(), testing should
478        pick them up.
479    
480    14. The -s option of pcretest has been documented for ages as being an old
481        synonym of -m (show memory usage). I have changed it to mean "force study
482        for every regex", that is, assume /S for every regex. This is similar to -i
483        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
484        using it. It makes it easier to run collections of tests with and without
485        study enabled, and thereby test pcre_study() more easily. All the standard
486        tests are now run with and without -s (but some patterns can be marked as
487        "never study" - see 20 below).
488    
489    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
490        restoration of the capturing data to the outer values was not happening
491        correctly.
492    
493    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
494        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
495        pattern had matched an empty string, and so incorrectly returned a no
496        match.
497    
498    17. There was optimizing code for the last branch of non-capturing parentheses,
499        and also for the obeyed branch of a conditional subexpression, which used
500        tail recursion to cut down on stack usage. Unfortunately, now that there is
501        the possibility of (*THEN) occurring in these branches, tail recursion is
502        no longer possible because the return has to be checked for (*THEN). These
503        two optimizations have therefore been removed. [But see 8.20/11 above.]
504    
505    18. If a pattern containing \R was studied, it was assumed that \R always
506        matched two bytes, thus causing the minimum subject length to be
507        incorrectly computed because \R can also match just one byte.
508    
509    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
510        was incorrectly computed.
511    
512    20. If /S is present twice on a test pattern in pcretest input, it now
513        *disables* studying, thereby overriding the use of -s on the command line
514        (see 14 above). This is necessary for one or two tests to keep the output
515        identical in both cases.
516    
517    21. When (*ACCEPT) was used in an assertion that matched an empty string and
518        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
519    
520    22. When an atomic group that contained a capturing parenthesis was
521        successfully matched, but the branch in which it appeared failed, the
522        capturing was not being forgotten if a higher numbered group was later
523        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
524        group 1 to "a", when in fact it should be unset. This applied to multi-
525        branched capturing and non-capturing groups, repeated or not, and also to
526        positive assertions (capturing in negative assertions does not happen
527        in PCRE) and also to nested atomic groups.
528    
529    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
530        subject after a captured substring, to make it easier to tell which of a
531        number of identical substrings has been captured.
532    
533    24. The way atomic groups are processed by pcre_exec() has been changed so that
534        if they are repeated, backtracking one repetition now resets captured
535        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
536        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
537        "aaa". Previously, it would have been "a". As part of this code
538        refactoring, the way recursive calls are handled has also been changed.
539    
540    25. If an assertion condition captured any substrings, they were not passed
541        back unless some other capturing happened later. For example, if
542        (?(?=(a))a) was matched against "a", no capturing was returned.
543    
544    26. When studying a pattern that contained subroutine calls or assertions,
545        the code for finding the minimum length of a possible match was handling
546        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
547        group 1 called group 2 while simultaneously a separate group 2 called group
548        1). A stack overflow occurred in this case. I have fixed this by limiting
549        the recursion depth to 10.
550    
551    27. Updated RunTest.bat in the distribution to the version supplied by Tom
552        Fortmann. This supports explicit test numbers on the command line, and has
553        argument validation and error reporting.
554    
555    28. An instance of \X with an unlimited repeat could fail if at any point the
556        first character it looked at was a mark character.
557    
558    29. Some minor code refactoring concerning Unicode properties and scripts
559        should reduce the stack requirement of match() slightly.
560    
561    30. Added the '=' option to pcretest to check the setting of unused capturing
562        slots at the end of the pattern, which are documented as being -1, but are
563        not included in the return count.
564    
565    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
566        compiled something random. Now it gives a compile-time error (as does
567        Perl).
568    
569    32. A *MARK encountered during the processing of a positive assertion is now
570        recorded and passed back (compatible with Perl).
571    
572    33. If --only-matching or --colour was set on a pcregrep call whose pattern
573        had alternative anchored branches, the search for a second match in a line
574        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
575        matched the line "0102" twice. The same bug affected patterns that started
576        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
577        twice.
578    
579    34. Previously, PCRE did not allow quantification of assertions. However, Perl
580        does, and because of capturing effects, quantifying parenthesized
581        assertions may at times be useful. Quantifiers are now allowed for
582        parenthesized assertions.
583    
584    35. A minor code tidy in pcre_compile() when checking options for \R usage.
585    
586    36. \g was being checked for fancy things in a character class, when it should
587        just be a literal "g".
588    
589    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
590        appearance of a nested POSIX class supersedes an apparent external class.
591        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
592        unescaped square brackets may also appear as part of class names. For
593        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
594        more like Perl. (But see 8.20/1 above.)
595    
596    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
597        was because it thought it was \N{name}, which is not supported).
598    
599    39. Add minix to OS list not supporting the -S option in pcretest.
600    
601    40. PCRE tries to detect cases of infinite recursion at compile time, but it
602        cannot analyze patterns in sufficient detail to catch mutual recursions
603        such as ((?1))((?2)). There is now a runtime test that gives an error if a
604        subgroup is called recursively as a subpattern for a second time at the
605        same position in the subject string. In previous releases this might have
606        been caught by the recursion limit, or it might have run out of stack.
607    
608    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
609        happen only once. PCRE was, however incorrectly giving a compile time error
610        "recursive call could loop indefinitely" because it cannot analyze the
611        pattern in sufficient detail. The compile time test no longer happens when
612        PCRE is compiling a conditional subpattern, but actual runaway loops are
613        now caught at runtime (see 40 above).
614    
615    42. It seems that Perl allows any characters other than a closing parenthesis
616        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
617        has been changed to be the same.
618    
619    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
620        as not to get warnings when autogen.sh is called. Also changed
621        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
622    
623    44. To help people who use pcregrep to scan files containing exceedingly long
624        lines, the following changes have been made:
625    
626        (a) The default value of the buffer size parameter has been increased from
627            8K to 20K. (The actual buffer used is three times this size.)
628    
629        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
630            PCRE is built.
631    
632        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
633            to be set at run time.
634    
635        (d) Numerical values in pcregrep options can be followed by K or M, for
636            example --buffer-size=50K.
637    
638        (e) If a line being scanned overflows pcregrep's buffer, an error is now
639            given and the return code is set to 2.
640    
641    45. Add a pointer to the latest mark to the callout data block.
642    
643    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
644        partial match of an empty string instead of no match. This was specific to
645        the use of ".".
646    
647    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
648        complete match instead of a partial match. This bug was dependent on both
649        the PCRE_UTF8 and PCRE_DOTALL options being set.
650    
651    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
652        starting byte set, because \b was not being ignored.
653    
654    
655    Version 8.12 15-Jan-2011
656    ------------------------
657    
658    1.  Fixed some typos in the markup of the man pages, and wrote a script that
659        checks for such things as part of the documentation building process.
660    
661    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
662        --match-limit and --recursion-limit options (added for 8.11). In
663        particular, this made one of the standard tests fail. (The integer value
664        went into the wrong half of a long int.)
665    
666    3.  If the --colour option was given to pcregrep with -v (invert match), it
667        did strange things, either producing crazy output, or crashing. It should,
668        of course, ignore a request for colour when reporting lines that do not
669        match.
670    
671    4.  Another pcregrep bug caused similar problems if --colour was specified with
672        -M (multiline) and the pattern match finished with a line ending.
673    
674    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
675        matched in multiline mode, the following line was shown as part of the
676        match. This seems wrong, so I have changed it.
677    
678    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
679        the check for further matches in the same line (so they could be coloured)
680        to overrun the end of the current line. If another match was found, it was
681        incorrectly shown (and then shown again when found in the next line).
682    
683    7.  If pcregrep was compiled under Windows, there was a reference to the
684        function pcregrep_exit() before it was defined. I am assuming this was
685        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
686        reported by a user. I've moved the definition above the reference.
687    
688    
689    Version 8.11 10-Dec-2010
690    ------------------------
691    
692    1.  (*THEN) was not working properly if there were untried alternatives prior
693        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
694        backtracked to try for "b" instead of moving to the next alternative branch
695        at the same level (in this case, to look for "c"). The Perl documentation
696        is clear that when (*THEN) is backtracked onto, it goes to the "next
697        alternative in the innermost enclosing group".
698    
699    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
700        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
701        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
702        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
703        (*THEN).
704    
705    3.  If \s appeared in a character class, it removed the VT character from
706        the class, even if it had been included by some previous item, for example
707        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
708        of \s, but is part of the POSIX "space" class.)
709    
710    4.  A partial match never returns an empty string (because you can always
711        match an empty string at the end of the subject); however the checking for
712        an empty string was starting at the "start of match" point. This has been
713        changed to the "earliest inspected character" point, because the returned
714        data for a partial match starts at this character. This means that, for
715        example, /(?<=abc)def/ gives a partial match for the subject "abc"
716        (previously it gave "no match").
717    
718    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
719        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
720        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
721        has an implication that the given string is incomplete (because a partial
722        match is preferred over a full match). For this reason, these items now
723        give a partial match in this situation. [Aside: previously, the one case
724        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
725        match rather than a full match, which was wrong by the old rules, but is
726        now correct.]
727    
728    6.  There was a bug in the handling of #-introduced comments, recognized when
729        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
730        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
731        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
732        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
733        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
734        places in pcre_compile().
735    
736    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
737        comments when looking ahead for named forward references to subpatterns,
738        the only newline sequence it recognized was NL. It now handles newlines
739        according to the set newline convention.
740    
741    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
742        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
743        cater for a lack of strerror(). These oversights have been fixed.
744    
745    9.  Added --match-limit and --recursion-limit to pcregrep.
746    
747    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
748    
749    11. When the -o option was used, pcregrep was setting a return code of 1, even
750        when matches were found, and --line-buffered was not being honoured.
751    
752    12. Added an optional parentheses number to the -o and --only-matching options
753        of pcregrep.
754    
755    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
756        can match an empty string. The code to do it in pcretest and pcredemo
757        needed fixing:
758    
759        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
760            only one byte after an empty string match just before CRLF (this case
761            just got forgotten; "any" and "anycrlf" were OK).
762    
763        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
764            mode when an empty string match preceded an ASCII character followed by
765            a non-ASCII character. (The code for advancing by one character rather
766            than one byte was nonsense.)
767    
768        (c) The pcredemo.c sample program did not have any code at all to handle
769            the cases when CRLF is a valid newline sequence.
770    
771    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
772        as a starting offset was within the subject string. There is now a new
773        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
774        negative or greater than the length of the string. In order to test this,
775        pcretest is extended to allow the setting of negative starting offsets.
776    
777    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
778        starting offset points to the beginning of a UTF-8 character was
779        unnecessarily clumsy. I tidied it up.
780    
781    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
782        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
783    
784    17. Nobody had reported that the --include_dir option, which was added in
785        release 7.7 should have been called --include-dir (hyphen, not underscore)
786        for compatibility with GNU grep. I have changed it to --include-dir, but
787        left --include_dir as an undocumented synonym, and the same for
788        --exclude-dir, though that is not available in GNU grep, at least as of
789        release 2.5.4.
790    
791    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
792        characters from a string of bytes) have been redefined so as not to use
793        loops, in order to improve performance in some environments. At the same
794        time, I abstracted some of the common code into auxiliary macros to save
795        repetition (this should not affect the compiled code).
796    
797    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
798        compile-time error is now given if \c is not followed by an ASCII
799        character, that is, a byte less than 128. (In EBCDIC mode, the code is
800        different, and any byte value is allowed.)
801    
802    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
803        START_OPTIMIZE option, which is now allowed at compile time - but just
804        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
805        to pcregrep and other applications that have no direct access to PCRE
806        options. The new /Y option in pcretest sets this option when calling
807        pcre_compile().
808    
809    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
810        back references. Groups containing recursive back references were forced to
811        be atomic by that change, but in the case of named groups, the amount of
812        memory required was incorrectly computed, leading to "Failed: internal
813        error: code overflow". This has been fixed.
814    
815    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
816        pcretest.c, to avoid build problems in some Borland environments.
817    
818    
819    Version 8.10 25-Jun-2010
820    ------------------------
821    
822    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
823        THEN.
824    
825    2.  (*ACCEPT) was not working when inside an atomic group.
826    
827    3.  Inside a character class, \B is treated as a literal by default, but
828        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
829        causes the error). The code is unchanged, but I tidied the documentation.
830    
831    4.  Inside a character class, PCRE always treated \R and \X as literals,
832        whereas Perl faults them if its -w option is set. I have changed PCRE so
833        that it faults them when PCRE_EXTRA is set.
834    
835    5.  Added support for \N, which always matches any character other than
836        newline. (It is the same as "." when PCRE_DOTALL is not set.)
837    
838    6.  When compiling pcregrep with newer versions of gcc which may have
839        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
840        declared with attribute warn_unused_result" were given. Just casting the
841        result to (void) does not stop the warnings; a more elaborate fudge is
842        needed. I've used a macro to implement this.
843    
844    7.  Minor change to pcretest.c to avoid a compiler warning.
845    
846    8.  Added four artifical Unicode properties to help with an option to make
847        \s etc use properties (see next item). The new properties are: Xan
848        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
849    
850    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
851        use Unicode properties. (*UCP) at the start of a pattern can be used to set
852        this option. Modified pcretest to add /W to test this facility. Added
853        REG_UCP to make it available via the POSIX interface.
854    
855    10. Added --line-buffered to pcregrep.
856    
857    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
858        studied, and the match started with a letter with a code point greater than
859        127 whose first byte was different to the first byte of the other case of
860        the letter, the other case of this starting letter was not recognized
861        (#976).
862    
863    12. If a pattern that was studied started with a repeated Unicode property
864        test, for example, \p{Nd}+, there was the theoretical possibility of
865        setting up an incorrect bitmap of starting bytes, but fortunately it could
866        not have actually happened in practice until change 8 above was made (it
867        added property types that matched character-matching opcodes).
868    
869    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
870        possible starting bytes for non-anchored patterns.
871    
872    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
873        \R, and also a number of cases that involve Unicode properties, both
874        explicit and implicit when PCRE_UCP is set.
875    
876    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
877        input, it could crash or give wrong results if characters with values
878        greater than 0xc0 were present in the subject string. (Detail: it assumed
879        UTF-8 input when processing these items.)
880    
881    16. Added a lot of (int) casts to avoid compiler warnings in systems where
882        size_t is 64-bit (#991).
883    
884    17. Added a check for running out of memory when PCRE is compiled with
885        --disable-stack-for-recursion (#990).
886    
887    18. If the last data line in a file for pcretest does not have a newline on
888        the end, a newline was missing in the output.
889    
890    19. The default pcre_chartables.c file recognizes only ASCII characters (values
891        less than 128) in its various bitmaps. However, there is a facility for
892        generating tables according to the current locale when PCRE is compiled. It
893        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
894        characters, are recognized by isspace() and therefore were getting set in
895        these tables, and indeed these tables seem to approximate to ISO 8859. This
896        caused a problem in UTF-8 mode when pcre_study() was used to create a list
897        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
898        which of course cannot start UTF-8 characters. I have changed the code so
899        that only real ASCII characters (less than 128) and the correct starting
900        bytes for UTF-8 encodings are set for characters greater than 127 when in
901        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
902        altogether.)
903    
904    20. Added the /T option to pcretest so as to be able to run tests with non-
905        standard character tables, thus making it possible to include the tests
906        used for 19 above in the standard set of tests.
907    
908    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
909        reference to a subpattern the other side of a comment that contains an
910        opening parenthesis caused either an internal compiling error, or a
911        reference to the wrong subpattern.
912    
913    
914    Version 8.02 19-Mar-2010
915    ------------------------
916    
917    1.  The Unicode data tables have been updated to Unicode 5.2.0.
918    
919    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
920        configured.
921    
922    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
923        original author of that file, following a query about its status.
924    
925    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
926        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
927    
928    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
929        quantifier applied to a forward-referencing subroutine call, could compile
930        incorrect code or give the error "internal error: previously-checked
931        referenced subpattern not found".
932    
933    6.  Both MS Visual Studio and Symbian OS have problems with initializing
934        variables to point to external functions. For these systems, therefore,
935        pcre_malloc etc. are now initialized to local functions that call the
936        relevant global functions.
937    
938    7.  There were two entries missing in the vectors called coptable and poptable
939        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
940        I've fixed the data, and added a kludgy way of testing at compile time that
941        the lengths are correct (equal to the number of opcodes).
942    
943    8.  Following on from 7, I added a similar kludge to check the length of the
944        eint vector in pcreposix.c.
945    
946    9.  Error texts for pcre_compile() are held as one long string to avoid too
947        much relocation at load time. To find a text, the string is searched,
948        counting zeros. There was no check for running off the end of the string,
949        which could happen if a new error number was added without updating the
950        string.
951    
952    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
953    
954    11. \K was not working if it appeared in an atomic group or in a group that
955        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
956        \K is "not well defined" if used in an assertion. PCRE now accepts it if
957        the assertion is positive, but not if it is negative.
958    
959    12. Change 11 fortuitously reduced the size of the stack frame used in the
960        "match()" function of pcre_exec.c by one pointer. Forthcoming
961        implementation of support for (*MARK) will need an extra pointer on the
962        stack; I have reserved it now, so that the stack frame size does not
963        decrease.
964    
965    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
966        item in branch that calls a recursion is a subroutine call - as in the
967        second branch in the above example - was incorrectly given the compile-
968        time error "recursive call could loop indefinitely" because pcre_compile()
969        was not correctly checking the subroutine for matching a non-empty string.
970    
971    14. The checks for overrunning compiling workspace could trigger after an
972        overrun had occurred. This is a "should never occur" error, but it can be
973        triggered by pathological patterns such as hundreds of nested parentheses.
974        The checks now trigger 100 bytes before the end of the workspace.
975    
976    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
977    
978    
979    Version 8.01 19-Jan-2010
980    ------------------------
981    
982    1.  If a pattern contained a conditional subpattern with only one branch (in
983        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
984        computed the wrong minimum data length (which is of course zero for such
985        subpatterns). This could cause incorrect "no match" results.
986    
987    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
988        the pattern is reset in the first branch, pcre_compile() failed with
989        "internal error: code overflow at offset...". This happened only when
990        the reset was to the original external option setting. (An optimization
991        abstracts leading options settings into an external setting, which was the
992        cause of this.)
993    
994    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
995        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
996        assertion pattern did not match (meaning that the assertion was true), it
997        was incorrectly treated as false if the SKIP had been reached during the
998        matching. This also applied to assertions used as conditions.
999    
1000    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1001        assertion subpattern, including such a pattern used as a condition,
1002        unpredictable results occurred, instead of the error return
1003        PCRE_ERROR_DFA_UITEM.
1004    
1005    5.  The C++ GlobalReplace function was not working like Perl for the special
1006        situation when an empty string is matched. It now does the fancy magic
1007        stuff that is necessary.
1008    
1009    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1010        removed. (These were left over from very, very early versions of PCRE.)
1011    
1012    7.  Some cosmetic changes to the code to make life easier when compiling it
1013        as part of something else:
1014    
1015        (a) Change DEBUG to PCRE_DEBUG.
1016    
1017        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1018            called "current" as "current_branch", to prevent a collision with the
1019            Linux macro when compiled as a kernel module.
1020    
1021        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1022            prevent a collision with the Linux macro when compiled as a kernel
1023            module.
1024    
1025    8.  In pcre_compile() there are some checks for integer overflows that used to
1026        cast potentially large values to (double). This has been changed to that
1027        when building, a check for int64_t is made, and if it is found, it is used
1028        instead, thus avoiding the use of floating point arithmetic. (There is no
1029        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1030        double.
1031    
1032    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1033        2005 (difference between two addresses compared to an unsigned value).
1034    
1035    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1036        custom one, because of the following reported problem in Windows:
1037    
1038          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1039              under Win32.
1040          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1041              therefore missing the function definition.
1042          - The compiler thus generates a "C" signature for the test function.
1043          - The linker fails to find the "C" function.
1044          - PCRE fails to configure if asked to do so against libbz2.
1045    
1046    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1047        messages were output:
1048    
1049          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1050          rerunning libtoolize, to keep the correct libtool macros in-tree.
1051          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1052    
1053        I have done both of these things.
1054    
1055    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1056        most of the time, it *can* run out if it is given a pattern that contains a
1057        runaway infinite recursion. I updated the discussion in the pcrestack man
1058        page.
1059    
1060    13. Now that we have gone to the x.xx style of version numbers, the minor
1061        version may start with zero. Using 08 or 09 is a bad idea because users
1062        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1063        interpreted as invalid octal numbers. I've updated the previous comment in
1064        configure.ac, and also added a check that gives an error if 08 or 09 are
1065        used.
1066    
1067    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1068        causing partial matching to fail when the end of the subject matched \W
1069        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1070    
1071    15. There were some discrepancies between the declarations in pcre_internal.h
1072        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1073        their definitions. The declarations used "const uschar *" and the
1074        definitions used USPTR. Even though USPTR is normally defined as "const
1075        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1076        reported that: "This difference in casting confuses some C++ compilers, for
1077        example, SunCC recognizes above declarations as different functions and
1078        generates broken code for hbpcre." I have changed the declarations to use
1079        USPTR.
1080    
1081    16. GNU libtool is named differently on some systems. The autogen.sh script now
1082        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1083        (FreeBSD).
1084    
1085    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1086        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1087        comment: "Figure out how to create a longlong from a string: strtoll and
1088        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1089        instance, but it only takes 2 args instead of 3!"
1090    
1091    18. A subtle bug concerned with back references has been fixed by a change of
1092        specification, with a corresponding code fix. A pattern such as
1093        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1094        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1095        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1096        same bug. Such groups have to be quantified to be useful, or contained
1097        inside another quantified group. (If there's no repetition, the reference
1098        can never match.) The problem arises because, having left the group and
1099        moved on to the rest of the pattern, a later failure that backtracks into
1100        the group uses the captured value from the final iteration of the group
1101        rather than the correct earlier one. I have fixed this in PCRE by forcing
1102        any group that contains a reference to itself to be an atomic group; that
1103        is, there cannot be any backtracking into it once it has completed. This is
1104        similar to recursive and subroutine calls.
1105    
1106    
1107    Version 8.00 19-Oct-09
1108    ----------------------
1109    
1110    1.  The table for translating pcre_compile() error codes into POSIX error codes
1111        was out-of-date, and there was no check on the pcre_compile() error code
1112        being within the table. This could lead to an OK return being given in
1113        error.
1114    
1115    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1116        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1117        in a Windows environment.
1118    
1119    3.  The pcregrep --count option prints the count for each file even when it is
1120        zero, as does GNU grep. However, pcregrep was also printing all files when
1121        --files-with-matches was added. Now, when both options are given, it prints
1122        counts only for those files that have at least one match. (GNU grep just
1123        prints the file name in this circumstance, but including the count seems
1124        more useful - otherwise, why use --count?) Also ensured that the
1125        combination -clh just lists non-zero counts, with no names.
1126    
1127    4.  The long form of the pcregrep -F option was incorrectly implemented as
1128        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1129        but it seems right to fix it, and I didn't think it was worth preserving
1130        the old behaviour.
1131    
1132    5.  The command line items --regex=pattern and --regexp=pattern were not
1133        recognized by pcregrep, which required --regex pattern or --regexp pattern
1134        (with a space rather than an '='). The man page documented the '=' forms,
1135        which are compatible with GNU grep; these now work.
1136    
1137    6.  No libpcreposix.pc file was created for pkg-config; there was just
1138        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1139    
1140    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1141        when UCP support is not needed, by modifying the Python script that
1142        generates it from Unicode data files. This should not matter if the module
1143        is correctly used as a library, but I received one complaint about 50K of
1144        unwanted data. My guess is that the person linked everything into his
1145        program rather than using a library. Anyway, it does no harm.
1146    
1147    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1148        was a minimum greater than 1 for a wide character in a possessive
1149        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1150        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1151        character. Chaos in the form of incorrect output or a compiling loop could
1152        result.
1153    
1154    9.  The restrictions on what a pattern can contain when partial matching is
1155        requested for pcre_exec() have been removed. All patterns can now be
1156        partially matched by this function. In addition, if there are at least two
1157        slots in the offset vector, the offset of the earliest inspected character
1158        for the match and the offset of the end of the subject are set in them when
1159        PCRE_ERROR_PARTIAL is returned.
1160    
1161    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1162        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1163        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1164        and may be more useful for multi-segment matching.
1165    
1166    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1167        used to be given if ever the end of the subject was reached; now it is
1168        given only if matching could not proceed because another character was
1169        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1170        string "Z", which now yields "no match" instead of "partial match". In the
1171        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1172        final character ended with (*FAIL).
1173    
1174    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1175        if the pattern had a "must contain" character that was already found in the
1176        earlier partial match, unless partial matching was again requested. For
1177        example, with the pattern /dog.(body)?/, the "must contain" character is
1178        "g". If the first part-match was for the string "dog", restarting with
1179        "sbody" failed. This bug has been fixed.
1180    
1181    13. The string returned by pcre_dfa_exec() after a partial match has been
1182        changed so that it starts at the first inspected character rather than the
1183        first character of the match. This makes a difference only if the pattern
1184        starts with a lookbehind assertion or \b or \B (\K is not supported by
1185        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1186        matching functions compatible, and I think it's the right thing to do.
1187    
1188    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1189        so that the demonstration program is easily available in environments where
1190        PCRE has not been installed from source.
1191    
1192    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1193        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1194        library.
1195    
1196    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1197        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1198        is not the first non-POSIX option to be added. Clearly some people find
1199        these options useful.
1200    
1201    17. If a caller to the POSIX matching function regexec() passes a non-zero
1202        value for nmatch with a NULL value for pmatch, the value of
1203        nmatch is forced to zero.
1204    
1205    18. RunGrepTest did not have a test for the availability of the -u option of
1206        the diff command, as RunTest does. It now checks in the same way as
1207        RunTest, and also checks for the -b option.
1208    
1209    19. If an odd number of negated classes containing just a single character
1210        interposed, within parentheses, between a forward reference to a named
1211        subpattern and the definition of the subpattern, compilation crashed with
1212        an internal error, complaining that it could not find the referenced
1213        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1214        [The bug was that it was starting one character too far in when skipping
1215        over the character class, thus treating the ] as data rather than
1216        terminating the class. This meant it could skip too much.]
1217    
1218    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1219        /g option in pcretest when the pattern contains \K, which makes it possible
1220        to have an empty string match not at the start, even when the pattern is
1221        anchored. Updated pcretest and pcredemo to use this option.
1222    
1223    21. If the maximum number of capturing subpatterns in a recursion was greater
1224        than the maximum at the outer level, the higher number was returned, but
1225        with unset values at the outer level. The correct (outer level) value is
1226        now given.
1227    
1228    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1229        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1230        make it do so. The string so far is captured, making this feature
1231        compatible with Perl.
1232    
1233    23. The tests have been re-organized, adding tests 11 and 12, to make it
1234        possible to check the Perl 5.10 features against Perl 5.10.
1235    
1236    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1237        pattern matches a fixed length string. PCRE did not allow this; now it
1238        does. Neither allows recursion.
1239    
1240    25. I finally figured out how to implement a request to provide the minimum
1241        length of subject string that was needed in order to match a given pattern.
1242        (It was back references and recursion that I had previously got hung up
1243        on.) This code has now been added to pcre_study(); it finds a lower bound
1244        to the length of subject needed. It is not necessarily the greatest lower
1245        bound, but using it to avoid searching strings that are too short does give
1246        some useful speed-ups. The value is available to calling programs via
1247        pcre_fullinfo().
1248    
1249    26. While implementing 25, I discovered to my embarrassment that pcretest had
1250        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1251        study optimizations had never been tested with that matching function.
1252        Oops. What is worse, even when it was passed study data, there was a bug in
1253        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1254        were also very few tests of studied patterns with pcre_dfa_exec().
1255    
1256    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1257        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1258        on the other side of the coin, they are no longer allowed to have different
1259        names, because these cannot be distinguished in PCRE, and this has caused
1260        confusion. (This is a difference from Perl.)
1261    
1262    28. When duplicate subpattern names are present (necessarily with different
1263        numbers, as required by 27 above), and a test is made by name in a
1264        conditional pattern, either for a subpattern having been matched, or for
1265        recursion in such a pattern, all the associated numbered subpatterns are
1266        tested, and the overall condition is true if the condition is true for any
1267        one of them. This is the way Perl works, and is also more like the way
1268        testing by number works.
1269    
1270    
1271    Version 7.9 11-Apr-09
1272    ---------------------
1273    
1274    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1275        (pcretest), all targets were linked against these libraries. This included
1276        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1277        libraries. This caused unwanted dependencies to be created. This problem
1278        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1279        pcretest is linked with readline.
1280    
1281    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1282        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1283        moved outside it again, because FALSE and TRUE are already defined in AIX,
1284        but BOOL is not.
1285    
1286    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1287        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1288    
1289    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1290        hyphens) following file names and line numbers when outputting matching
1291        lines. This is not true; no spaces are inserted. I have also clarified the
1292        wording for the --colour (or --color) option.
1293    
1294    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1295        was not coloured; this is different to GNU grep, so I have changed it to be
1296        the same.
1297    
1298    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1299        each matching line was coloured. Now it goes on to look for further matches
1300        of any of the test patterns, which is the same behaviour as GNU grep.
1301    
1302    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1303        doesn't make sense to accept an empty string match in pcregrep, so I have
1304        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1305        seems to be how GNU grep behaves.
1306    
1307    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1308        start or after a newline", because the conditional assertion was not being
1309        correctly handled. The rule now is that both the assertion and what follows
1310        in the first alternative must satisfy the test.
1311    
1312    9.  If auto-callout was enabled in a pattern with a conditional group whose
1313        condition was an assertion, PCRE could crash during matching, both with
1314        pcre_exec() and pcre_dfa_exec().
1315    
1316    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1317        used for matching.
1318    
1319    11. Unicode property support in character classes was not working for
1320        characters (bytes) greater than 127 when not in UTF-8 mode.
1321    
1322    12. Added the -M command line option to pcretest.
1323    
1324    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1325    
1326    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1327    
1328    16. Added comments and documentation about mis-use of no_arg in the C++
1329        wrapper.
1330    
1331    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1332        from Martin Jerabek that uses macro names for all relevant character and
1333        string constants.
1334    
1335    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1336        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1337        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1338        these, but not everybody uses configure.
1339    
1340    19. A conditional group that had only one branch was not being correctly
1341        recognized as an item that could match an empty string. This meant that an
1342        enclosing group might also not be so recognized, causing infinite looping
1343        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1344        with the subject "ab", where knowledge that the repeated group can match
1345        nothing is needed in order to break the loop.
1346    
1347    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1348        exec(), but without supplying a callout function, matching went wrong.
1349    
1350    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1351        leak if the size of the offset vector was greater than 30. When the vector
1352        is smaller, the saved offsets during recursion go onto a local stack
1353        vector, but for larger vectors malloc() is used. It was failing to free
1354        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1355        error, in fact).
1356    
1357    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1358        heapframe that is used only when UTF-8 support is enabled. This caused no
1359        problem, but was untidy.
1360    
1361    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1362        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1363        included within another project.
1364    
1365    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1366        slightly modified by me:
1367    
1368          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1369              not building pcregrep.
1370    
1371          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1372              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1373    
1374    25. Forward references, both numeric and by name, in patterns that made use of
1375        duplicate group numbers, could behave incorrectly or give incorrect errors,
1376        because when scanning forward to find the reference group, PCRE was not
1377        taking into account the duplicate group numbers. A pattern such as
1378        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1379    
1380    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1381        the feature of a custom pointer more persuasive (as requested by a user).
1382    
1383    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1384        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1385        pre-defined.
1386    
1387    28. Added support for (*UTF8) at the start of a pattern.
1388    
1389    29. Arrange for flags added by the "release type" setting in CMake to be shown
1390        in the configuration summary.
1391    
1392    
1393    Version 7.8 05-Sep-08
1394    ---------------------
1395    
1396    1.  Replaced UCP searching code with optimized version as implemented for Ad
1397        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1398        stage table and inline lookup instead of a function, giving speed ups of 2
1399        to 5 times on some simple patterns that I tested. Permission was given to
1400        distribute the MultiStage2.py script that generates the tables (it's not in
1401        the tarball, but is in the Subversion repository).
1402    
1403    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1404        scripts.
1405    
1406    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1407        a group with a zero qualifier. The result of the study could be incorrect,
1408        or the function might crash, depending on the pattern.
1409    
1410    4.  Caseless matching was not working for non-ASCII characters in back
1411        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1412        It now works when Unicode Property Support is available.
1413    
1414    5.  In pcretest, an escape such as \x{de} in the data was always generating
1415        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1416        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1417        truncation.
1418    
1419    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1420    
1421    7.  Added two (int) casts to pcregrep when printing the difference of two
1422        pointers, in case they are 64-bit values.
1423    
1424    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1425        test 2 if it fails.
1426    
1427    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1428        and a #define of that name to empty if it is not externally set. This is to
1429        allow users of MSVC to set it if necessary.
1430    
1431    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1432        the convenience functions in the pcre_get.c source file.
1433    
1434    11. An option change at the start of a pattern that had top-level alternatives
1435        could cause overwriting and/or a crash. This command provoked a crash in
1436        some environments:
1437    
1438          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1439    
1440        This potential security problem was recorded as CVE-2008-2371.
1441    
1442    12. For a pattern where the match had to start at the beginning or immediately
1443        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1444        pcre_dfa_exec() could read past the end of the passed subject if there was
1445        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1446        pcretest so that it places the subject at the end of its malloc-ed buffer.
1447    
1448    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1449        exec() might read past the end of the data buffer in UTF-8 mode.
1450    
1451    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1452        the data contained the byte 0x85 as part of a UTF-8 character within its
1453        first line. This applied both to normal and DFA matching.
1454    
1455    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1456        /^[^d]*?$/8 failed to match "abc".
1457    
1458    16. Added a missing copyright notice to pcrecpp_internal.h.
1459    
1460    17. Make it more clear in the documentation that values returned from
1461        pcre_exec() in ovector are byte offsets, not character counts.
1462    
1463    18. Tidied a few places to stop certain compilers from issuing warnings.
1464    
1465    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1466        supplied by Stefan Weber. I made a further small update for 7.8 because
1467        there is a change of source arrangements: the pcre_searchfuncs.c module is
1468        replaced by pcre_ucd.c.
1469    
1470    
1471    Version 7.7 07-May-08
1472    ---------------------
1473    
1474    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1475        a string to a long long, pretend we don't even have a long long." This is
1476        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1477    
1478    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1479        pre-7.6 versions, which defined a global no_arg variable instead of putting
1480        it in the RE class. (See also #8 below.)
1481    
1482    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1483        Lopes.
1484    
1485    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1486    
1487        (1) The include/exclude patterns were being applied to the whole pathnames
1488            of files, instead of just to the final components.
1489    
1490        (2) If there was more than one level of directory, the subdirectories were
1491            skipped unless they satisfied the include/exclude conditions. This is
1492            inconsistent with GNU grep (and could even be seen as contrary to the
1493            pcregrep specification - which I improved to make it absolutely clear).
1494            The action now is always to scan all levels of directory, and just
1495            apply the include/exclude patterns to regular files.
1496    
1497    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1498        --exclude_dir in the tests to avoid scanning .svn directories.
1499    
1500    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1501        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1502        doesn't support NULs in patterns.
1503    
1504    7.  Added some missing "const"s to declarations of static tables in
1505        pcre_compile.c and pcre_dfa_exec.c.
1506    
1507    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1508        caused by fix #2  above. (Subsequently also a second patch to fix the
1509        first patch. And a third patch - this was a messy problem.)
1510    
1511    9.  Applied Craig's patch to remove the use of push_back().
1512    
1513    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1514        matching function regexec().
1515    
1516    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1517        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1518        references. PCRE supports relative numbers with this syntax (I don't think
1519        Oniguruma does).
1520    
1521    12. Previously, a group with a zero repeat such as (...){0} was completely
1522        omitted from the compiled regex. However, this means that if the group
1523        was called as a subroutine from elsewhere in the pattern, things went wrong
1524        (an internal error was given). Such groups are now left in the compiled
1525        pattern, with a new opcode that causes them to be skipped at execution
1526        time.
1527    
1528    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1529        to the way PCRE behaves:
1530    
1531        (a) A lone ] character is dis-allowed (Perl treats it as data).
1532    
1533        (b) A back reference to an unmatched subpattern matches an empty string
1534            (Perl fails the current match path).
1535    
1536        (c) A data ] in a character class must be notated as \] because if the
1537            first data character in a class is ], it defines an empty class. (In
1538            Perl it is not possible to have an empty class.) The empty class []
1539            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1540            The negative empty class [^] matches any one character, independently
1541            of the DOTALL setting.
1542    
1543    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1544        non-existent subpattern following a character class starting with ']' and
1545        containing () gave an internal compiling error instead of "reference to
1546        non-existent subpattern". Fortunately, when the pattern did exist, the
1547        compiled code was correct. (When scanning forwards to check for the
1548        existencd of the subpattern, it was treating the data ']' as terminating
1549        the class, so got the count wrong. When actually compiling, the reference
1550        was subsequently set up correctly.)
1551    
1552    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1553        it was being rejected as not supported by pcre_dfa_exec(), even though
1554        other assertions are supported. I have made pcre_dfa_exec() support
1555        (*FAIL).
1556    
1557    16. The implementation of 13c above involved the invention of a new opcode,
1558        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1559        cannot be changed at match time, I realized I could make a small
1560        improvement to matching performance by compiling OP_ALLANY instead of
1561        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1562        on the OP_ANY path.
1563    
1564    17. Compiling pcretest on Windows with readline support failed without the
1565        following two fixes: (1) Make the unistd.h include conditional on
1566        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1567    
1568    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1569        ncurses library to be included for pcretest when ReadLine support is
1570        requested, but also to allow for it to be overridden. This patch came from
1571        Daniel Bergström.
1572    
1573    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1574        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1575        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1576        spotting this.
1577    
1578    
1579    Version 7.6 28-Jan-08
1580    ---------------------
1581    
1582    1.  A character class containing a very large number of characters with
1583        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1584        overflow.
1585    
1586    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1587        HAVE_LONG_LONG is not defined.
1588    
1589    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1590        bring them up to date and include new features. This patch includes:
1591    
1592        - Fixed PH's badly added libz and libbz2 support.
1593        - Fixed a problem with static linking.
1594        - Added pcredemo. [But later removed - see 7 below.]
1595        - Fixed dftables problem and added an option.
1596        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1597            HAVE_LONG_LONG.
1598        - Added readline support for pcretest.
1599        - Added an listing of the option settings after cmake has run.
1600    
1601    4.  A user submitted a patch to Makefile that makes it easy to create
1602        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1603        Makefile.am that cause it to include this special target, without
1604        affecting anything else. Note that the same mingw target plus all
1605        the other distribution libraries and programs are now supported
1606        when configuring with CMake (see 6 below) instead of with
1607        Configure/Make.
1608    
1609    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1610        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1611        exported in the Windows port". It has not yet been confirmed that the patch
1612        solves the problem, but it does no harm.
1613    
1614    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1615        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1616        with CMake, and also correct the comment about stack recursion.
1617    
1618    7.  Remove the automatic building of pcredemo from the ./configure system and
1619        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1620        of a program that users should build themselves after PCRE is installed, so
1621        building it automatically is not really right. What is more, it gave
1622        trouble in some build environments.
1623    
1624    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1625    
1626    
1627    Version 7.5 10-Jan-08
1628    ---------------------
1629    
1630    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1631        values in parens when parsing an RE using the C++ wrapper."
1632    
1633    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1634        Characters greater than 255 were excluded from the class instead of being
1635        included.
1636    
1637    3.  The same bug as (2) above applied to negated POSIX classes such as
1638        [:^space:].
1639    
1640    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1641        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1642        I have changed it.
1643    
1644    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1645        first named subpattern) and a construct such as (?&a) would reference the
1646        first named subpattern whose name started with "a" (in other words, the
1647        length check was missing). Both these problems are fixed. "Subpattern name
1648        expected" is now given for (?&) (a zero-length name), and this patch also
1649        makes it give the same error for \k'' (previously it complained that that
1650        was a reference to a non-existent subpattern).
1651    
1652    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1653        this is right because (?- can be followed by option settings as well as by
1654        digits. I have, however, made the messages clearer.
1655    
1656    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1657        than the number used in the conditional) now cause a compile-time error.
1658        This is actually not compatible with Perl, which accepts such patterns, but
1659        treats the conditional as always being FALSE (as PCRE used to), but it
1660        seems to me that giving a diagnostic is better.
1661    
1662    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1663        and messages.
1664    
1665    9.  Fix two occurrences of "backslash" in comments that should have been
1666        "backspace".
1667    
1668    10. Remove two redundant lines of code that can never be obeyed (their function
1669        was moved elsewhere).
1670    
1671    11. The program that makes PCRE's Unicode character property table had a bug
1672        which caused it to generate incorrect table entries for sequences of
1673        characters that have the same character type, but are in different scripts.
1674        It amalgamated them into a single range, with the script of the first of
1675        them. In other words, some characters were in the wrong script. There were
1676        thirteen such cases, affecting characters in the following ranges:
1677    
1678          U+002b0 - U+002c1
1679          U+0060c - U+0060d
1680          U+0061e - U+00612
1681          U+0064b - U+0065e
1682          U+0074d - U+0076d
1683          U+01800 - U+01805
1684          U+01d00 - U+01d77
1685          U+01d9b - U+01dbf
1686          U+0200b - U+0200f
1687          U+030fc - U+030fe
1688          U+03260 - U+0327f
1689          U+0fb46 - U+0fbb1
1690          U+10450 - U+1049d
1691    
1692    12. The -o option (show only the matching part of a line) for pcregrep was not
1693        compatible with GNU grep in that, if there was more than one match in a
1694        line, it showed only the first of them. It now behaves in the same way as
1695        GNU grep.
1696    
1697    13. If the -o and -v options were combined for pcregrep, it printed a blank
1698        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1699        does the same. The return code can be used to tell if there were any
1700        non-matching lines.
1701    
1702    14. Added --file-offsets and --line-offsets to pcregrep.
1703    
1704    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1705        infinitely looping recursion. The bug was that positive lookaheads were not
1706        being skipped when checking for a possible empty match (negative lookaheads
1707        and both kinds of lookbehind were skipped).
1708    
1709    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1710        inclusion of <windows.h> to before rather than after the definition of
1711        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1712    
1713    17. Specifying a possessive quantifier with a specific limit for a Unicode
1714        character property caused pcre_compile() to compile bad code, which led at
1715        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1716        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1717        caused the error; without that there was no problem.
1718    
1719    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1720    
1721    19. Added --enable-pcretest-libreadline.
1722    
1723    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1724        RE::GlobalReplace(). As a result, the number of replacements returned was
1725        double what it should be. I removed one of the increments, but Craig sent a
1726        later patch that removed the other one (the right fix) and added unit tests
1727        that check the return values (which was not done before).
1728    
1729    21. Several CMake things:
1730    
1731        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1732            the names libpcre and libpcreposix, not just pcre and pcreposix.
1733    
1734        (2) The above change means that pcretest and pcregrep are now correctly
1735            linked with the newly-built libraries, not previously installed ones.
1736    
1737        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1738    
1739    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1740        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1741        UTF-8 newline character). The key issue is that the pattern starts .*;
1742        this means that the match must be either at the beginning, or after a
1743        newline. The bug was in the code for advancing after a failed match and
1744        checking that the new position followed a newline. It was not taking
1745        account of UTF-8 characters correctly.
1746    
1747    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1748        character classes. PCRE was not treating the sequence [:...:] as a
1749        character class unless the ... were all letters. Perl, however, seems to
1750        allow any characters between [: and :], though of course it rejects as
1751        unknown any "names" that contain non-letters, because all the known class
1752        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1753        for example, whereas PCRE did not - it did not recognize a POSIX character
1754        class. This seemed a bit dangerous, so the code has been changed to be
1755        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1756        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1757        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1758        Perl does, and where it didn't before.
1759    
1760    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1761        Windows environments %n is disabled by default.
1762    
1763    
1764    Version 7.4 21-Sep-07
1765    ---------------------
1766    
1767    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1768        means that a class such as [\s] counted as "explicit reference to CR or
1769        LF". That isn't really right - the whole point of the change was to try to
1770        help when there was an actual mention of one of the two characters. So now
1771        the change happens only if \r or \n (or a literal CR or LF) character is
1772        encountered.
1773    
1774    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1775        of both had grown to the point where there were only 3 bits left.
1776        Fortunately, there was spare space in the data structure, and so I have
1777        moved the internal flags into a new 16-bit field to free up more option
1778        bits.
1779    
1780    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1781        but did not set the internal JCHANGED flag - either of these is enough to
1782        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1783        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1784        start sets both bits.
1785    
1786    4.  Added options (at build time, compile time, exec time) to change \R from
1787        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1788    
1789    5.  doc/pcresyntax.html was missing from the distribution.
1790    
1791    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1792        compatibility, even though it is no longer used.
1793    
1794    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1795        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1796        windows.h file is present (where different names are used). [This was
1797        reversed later after testing - see 16 below.]
1798    
1799    8.  Changed all #include <config.h> to #include "config.h". There were also
1800        some further <pcre.h> cases that I changed to "pcre.h".
1801    
1802    9.  When pcregrep was used with the --colour option, it missed the line ending
1803        sequence off the lines that it output.
1804    
1805    10. It was pointed out to me that arrays of string pointers cause lots of
1806        relocations when a shared library is dynamically loaded. A technique of
1807        using a single long string with a table of offsets can drastically reduce
1808        these. I have refactored PCRE in four places to do this. The result is
1809        dramatic:
1810    
1811          Originally:                          290
1812          After changing UCP table:            187
1813          After changing error message table:   43
1814          After changing table of "verbs"       36
1815          After changing table of Posix names   22
1816    
1817        Thanks to the folks working on Gregex for glib for this insight.
1818    
1819    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1820        unicode-properties was also set.
1821    
1822    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1823    
1824    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1825        checked only for CRLF.
1826    
1827    14. Added casts to pcretest.c to avoid compiler warnings.
1828    
1829    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1830    
1831    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1832        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1833        entirely. This removes changes made in 7 above.
1834    
1835    17. The CMake files have been updated, and there is now more information about
1836        building with CMake in the NON-UNIX-USE document.
1837    
1838    
1839    Version 7.3 28-Aug-07
1840    ---------------------
1841    
1842     1. In the rejigging of the build system that eventually resulted in 7.1, the
1843        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1844        brackets there is not right, since it causes compilers to look for an
1845        installed pcre.h, not the version that is in the source that is being
1846        compiled (which of course may be different). I have changed it back to:
1847    
1848          #include "pcre.h"
1849    
1850        I have a vague recollection that the change was concerned with compiling in
1851        different directories, but in the new build system, that is taken care of
1852        by the VPATH setting the Makefile.
1853    
1854     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1855        when the subject happened to end in the byte 0x85 (e.g. if the last
1856        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1857        characters but of course it shouldn't be taken as a newline when it is part
1858        of another character. The bug was that, for an unlimited repeat of . in
1859        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1860        characters when looking for a newline.
1861    
1862     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1863    
1864     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1865        in debug output.
1866    
1867     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1868        long printing in the pcrecpp unittest when running under MinGW.
1869    
1870     6. ESC_K was left out of the EBCDIC table.
1871    
1872     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1873        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1874        limit also applies to "virtual nesting" when a pattern is recursive, and in
1875        this case 1000 isn't so big. I have been able to remove this limit at the
1876        expense of backing off one optimization in certain circumstances. Normally,
1877        when pcre_exec() would call its internal match() function recursively and
1878        immediately return the result unconditionally, it uses a "tail recursion"
1879        feature to save stack. However, when a subpattern that can match an empty
1880        string has an unlimited repetition quantifier, it no longer makes this
1881        optimization. That gives it a stack frame in which to save the data for
1882        checking that an empty string has been matched. Previously this was taken
1883        from the 1000-entry workspace that had been reserved. So now there is no
1884        explicit limit, but more stack is used.
1885    
1886     8. Applied Daniel's patches to solve problems with the import/export magic
1887        syntax that is required for Windows, and which was going wrong for the
1888        pcreposix and pcrecpp parts of the library. These were overlooked when this
1889        problem was solved for the main library.
1890    
1891     9. There were some crude static tests to avoid integer overflow when computing
1892        the size of patterns that contain repeated groups with explicit upper
1893        limits. As the maximum quantifier is 65535, the maximum group length was
1894        set at 30,000 so that the product of these two numbers did not overflow a
1895        32-bit integer. However, it turns out that people want to use groups that
1896        are longer than 30,000 bytes (though not repeat them that many times).
1897        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1898        made it possible to implement the integer overflow checks in a much more
1899        dynamic way, which I have now done. The artificial limitation on group
1900        length has been removed - we now have only the limit on the total length of
1901        the compiled pattern, which depends on the LINK_SIZE setting.
1902    
1903    10. Fixed a bug in the documentation for get/copy named substring when
1904        duplicate names are permitted. If none of the named substrings are set, the
1905        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1906        empty string.
1907    
1908    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1909        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1910        because the ] is interpreted as the first data character and the
1911        terminating ] is not found. PCRE has been made compatible with Perl in this
1912        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1913        cause memory overwriting.
1914    
1915    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1916        string has been matched (to stop an infinite loop). It was not recognizing
1917        a conditional subpattern that could match an empty string if that
1918        subpattern was within another subpattern. For example, it looped when
1919        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1920        condition was not nested. This bug has been fixed.
1921    
1922    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1923        past the start of the subject in the presence of bytes with the top bit
1924        set, for example "\x8aBCD".
1925    
1926    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1927        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1928    
1929    14. Optimized (?!) to (*FAIL).
1930    
1931    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1932        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1933        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1934        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1935        does: it's just the validity check that is more restrictive.
1936    
1937    16. Inserted checks for integer overflows during escape sequence (backslash)
1938        processing, and also fixed erroneous offset values for syntax errors during
1939        backslash processing.
1940    
1941    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1942        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1943    
1944    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1945        caused an overrun.
1946    
1947    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1948        something other than just ASCII characters) inside a group that had an
1949        unlimited repeat caused a loop at compile time (while checking to see
1950        whether the group could match an empty string).
1951    
1952    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1953        [\P{Any}] did so. (Error in the code for printing property names.)
1954    
1955    21. An orphan \E inside a character class could cause a crash.
1956    
1957    22. A repeated capturing bracket such as (A)? could cause a wild memory
1958        reference during compilation.
1959    
1960    23. There are several functions in pcre_compile() that scan along a compiled
1961        expression for various reasons (e.g. to see if it's fixed length for look
1962        behind). There were bugs in these functions when a repeated \p or \P was
1963        present in the pattern. These operators have additional parameters compared
1964        with \d, etc, and these were not being taken into account when moving along
1965        the compiled data. Specifically:
1966    
1967        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1968            length.
1969    
1970        (b) An item such as \pL+ within a repeated group could cause crashes or
1971            loops.
1972    
1973        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1974            "reference to non-existent subpattern" error.
1975    
1976        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1977    
1978    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1979        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1980    
1981    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1982    
1983    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1984        character were causing crashes (broken optimization).
1985    
1986    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1987        \p or \P) caused a compile-time loop.
1988    
1989    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1990        break. For example, the unstudied pattern [\r\n]A does not match the string
1991        "\r\nA" because change 7.0/46 below moves the current point on by two
1992        characters after failing to match at the start. However, the pattern \nA
1993        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1994        the same is true. There doesn't seem any very clean way out of this, but
1995        what I have chosen to do makes the common cases work: PCRE now takes note
1996        of whether there can be an explicit match for \r or \n anywhere in the
1997        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1998        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1999        pattern has explicit CR or LF references.
2000    
2001    29. Added (*CR) etc for changing newline setting at start of pattern.
2002    
2003    
2004    Version 7.2 19-Jun-07
2005    ---------------------
2006    
2007     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2008        which is apparently normally available under Windows.
2009    
2010     2. Re-jig the pcregrep tests with different newline settings in an attempt
2011        to make them independent of the local environment's newline setting.
2012    
2013     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2014    
2015     4. Some of the "internals" tests were previously cut out when the link size
2016        was not 2, because the output contained actual offsets. The recent new
2017        "Z" feature of pcretest means that these can be cut out, making the tests
2018        usable with all link sizes.
2019    
2020     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2021        stack recursion. This gives a massive performance boost under BSD, but just
2022        a small improvement under Linux. However, it saves one field in the frame
2023        in all cases.
2024    
2025     6. Added more features from the forthcoming Perl 5.10:
2026    
2027        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2028            recursion call. It refers to the nth most recently opened parentheses.
2029    
2030        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2031            to be opened parentheses.
2032    
2033        (c) Conditions that refer to capturing parentheses can be specified
2034            relatively, for example, (?(-2)... or (?(+3)...
2035    
2036        (d) \K resets the start of the current match so that everything before
2037            is not part of it.
2038    
2039        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2040    
2041        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2042            reference syntax.
2043    
2044        (g) (?| introduces a group in which the numbering of parentheses in each
2045            alternative starts with the same number.
2046    
2047        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2048    
2049     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2050        PCRE_INFO_JCHANGED.
2051    
2052     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2053        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2054        for detecting groups that can match an empty string.
2055    
2056     9. A pattern with a very large number of alternatives (more than several
2057        hundred) was running out of internal workspace during the pre-compile
2058        phase, where pcre_compile() figures out how much memory will be needed. A
2059        bit of new cunning has reduced the workspace needed for groups with
2060        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2061        workspace instead of running out of the 4096 that are available.
2062    
2063    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2064    
2065    11. Applied patch from Google to remove an optimization that didn't quite work.
2066        The report of the bug said:
2067    
2068          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2069          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2070          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2071    
2072    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2073        it matched the wrong number of bytes.
2074    
2075    
2076    Version 7.1 24-Apr-07
2077    ---------------------
2078    
2079     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2080        that is more "standard", making use of automake and other Autotools. There
2081        is some re-arrangement of the files and adjustment of comments consequent
2082        on this.
2083    
2084     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2085        for recursive directory scanning broke on some systems because the files
2086        are not scanned in any specific order and on different systems the order
2087        was different. A call to "sort" has been inserted into RunGrepTest for the
2088        approprate test as a short-term fix. In the longer term there may be an
2089        alternative.
2090    
2091     3. I had an email from Eric Raymond about problems translating some of PCRE's
2092        man pages to HTML (despite the fact that I distribute HTML pages, some
2093        people do their own conversions for various reasons). The problems
2094        concerned the use of low-level troff macros .br and .in. I have therefore
2095        removed all such uses from the man pages (some were redundant, some could
2096        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2097        HTML has been updated to handle .nf/.fi and to complain if it encounters
2098        .br or .in.
2099    
2100     4. Updated comments in configure.ac that get placed in config.h.in and also
2101        arranged for config.h to be included in the distribution, with the name
2102        config.h.generic, for the benefit of those who have to compile without
2103        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2104    
2105     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2106        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2107        makevp.bat for the current PCRE, using the additional files
2108        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2109    
2110     6. A Windows user reported a minor discrepancy with test 2, which turned out
2111        to be caused by a trailing space on an input line that had got lost in his
2112        copy. The trailing space was an accident, so I've just removed it.
2113    
2114     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2115        that is needed.
2116    
2117     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2118        as "const" (a) because they are and (b) because it helps the PHP
2119        maintainers who have recently made a script to detect big data structures
2120        in the php code that should be moved to the .rodata section. I remembered
2121        to update Builducptable as well, so it won't revert if ucptable.h is ever
2122        re-created.
2123    
2124     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2125        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2126        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2127        support is not required. This saves 1.5-2K of code, which is important in
2128        some applications.
2129    
2130        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2131        so as not to refer to the tables, even though these functions will never be
2132        called when UTF-8 support is disabled. Otherwise there are problems with a
2133        shared library.
2134    
2135    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2136    
2137        (a) It was defining its arguments as char * instead of void *.
2138    
2139        (b) It was assuming that all moves were upwards in memory; this was true
2140            a long time ago when I wrote it, but is no longer the case.
2141    
2142        The emulated memove() is provided for those environments that have neither
2143        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2144        is clearly not the case, as these two bugs were recently reported.
2145    
2146    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2147        and Detrail to create the HTML documentation, the .txt form of the man
2148        pages, and it removes trailing spaces from listed files. It also creates
2149        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2150        case, it wraps all the #defines with #ifndefs. This script should be run
2151        before "make dist".
2152    
2153    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2154        with Unicode property support.
2155    
2156        (a) For a maximizing quantifier, if the two different cases of the
2157            character were of different lengths in their UTF-8 codings (there are
2158            some cases like this - I found 11), and the matching function had to
2159            back up over a mixture of the two cases, it incorrectly assumed they
2160            were both the same length.
2161    
2162        (b) When PCRE was configured to use the heap rather than the stack for
2163            recursion during matching, it was not correctly preserving the data for
2164            the other case of a UTF-8 character when checking ahead for a match
2165            while processing a minimizing repeat. If the check also involved
2166            matching a wide character, but failed, corruption could cause an
2167            erroneous result when trying to check for a repeat of the original
2168            character.
2169    
2170    13. Some tidying changes to the testing mechanism:
2171    
2172        (a) The RunTest script now detects the internal link size and whether there
2173            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2174            values substituted by "configure". (The RunGrepTest script already did
2175            this for UTF-8.) The configure.ac script no longer substitutes the
2176            relevant variables.
2177    
2178        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2179            with length and offset values. This means that the output is different
2180            for different internal link sizes. Test 2 is skipped for link sizes
2181            other than 2 because of this, bypassing the problem. Unfortunately,
2182            there was also a test in test 3 (the locale tests) that used /B and
2183            failed for link sizes other than 2. Rather than cut the whole test out,
2184            I have added a new /Z option to pcretest that replaces the length and
2185            offset values with spaces. This is now used to make test 3 independent
2186            of link size. (Test 2 will be tidied up later.)
2187    
2188    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2189        segmentation fault instead of returning the appropriate error message.
2190    
2191    15. In multiline mode when the newline sequence was set to "any", the pattern
2192        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2193        This doesn't seem right; it now treats the CRLF combination as the line
2194        ending, and so does not match in that case. It's only a pattern such as ^$
2195        that would hit this one: something like ^ABC$ would have failed after \r
2196        and then tried again after \r\n.
2197    
2198    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2199        in an attempt to make files that differ only in their line terminators
2200        compare equal. This works on Linux.
2201    
2202    17. Under certain error circumstances pcregrep might try to free random memory
2203        as it exited. This is now fixed, thanks to valgrind.
2204    
2205    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2206        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2207        was because its rules for how to advance for /g after matching an empty
2208        string at the end of a line did not allow for this case. They now check for
2209        it specially.
2210    
2211    20. pcretest is supposed to handle patterns and data of any length, by
2212        extending its buffers when necessary. It was getting this wrong when the
2213        buffer for a data line had to be extended.
2214    
2215    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2216        CRLF as a newline sequence.
2217    
2218    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2219        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2220        I have nevertheless tidied it up.
2221    
2222    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2223    
2224    24. Added a man page for pcre-config.
2225    
2226    
2227    Version 7.0 19-Dec-06
2228    ---------------------
2229    
2230     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2231        moving to gcc 4.1.1.
2232    
2233     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2234        sys/time.h, which is documented as needed for this function. It doesn't
2235        seem to matter on Linux, but it showed up on some releases of OS X.
2236    
2237     3. It seems that there are systems where bytes whose values are greater than
2238        127 match isprint() in the "C" locale. The "C" locale should be the
2239        default when a C program starts up. In most systems, only ASCII printing
2240        characters match isprint(). This difference caused the output from pcretest
2241        to vary, making some of the tests fail. I have changed pcretest so that:
2242    
2243        (a) When it is outputting text in the compiled version of a pattern, bytes
2244            other than 32-126 are always shown as hex escapes.
2245    
2246        (b) When it is outputting text that is a matched part of a subject string,
2247            it does the same, unless a different locale has been set for the match
2248            (using the /L modifier). In this case, it uses isprint() to decide.
2249    
2250     4. Fixed a major bug that caused incorrect computation of the amount of memory
2251        required for a compiled pattern when options that changed within the
2252        pattern affected the logic of the preliminary scan that determines the
2253        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2254        that the computed length was too small. The symptoms of this bug were
2255        either the PCRE error "internal error: code overflow" from pcre_compile(),
2256        or a glibc crash with a message such as "pcretest: free(): invalid next
2257        size (fast)". Examples of patterns that provoked this bug (shown in
2258        pcretest format) are:
2259    
2260          /(?-x: )/x
2261          /(?x)(?-x: \s*#\s*)/
2262          /((?i)[\x{c0}])/8
2263          /(?i:[\x{c0}])/8
2264    
2265        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2266        is now done differently.
2267    
2268     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2269        wrapper classes; (b) implement a new function in the C++ scanner that is
2270        more efficient than the old way of doing things because it avoids levels of
2271        recursion in the regex matching; (c) add a paragraph to the documentation
2272        for the FullMatch() function.
2273    
2274     6. The escape sequence \n was being treated as whatever was defined as
2275        "newline". Not only was this contrary to the documentation, which states
2276        that \n is character 10 (hex 0A), but it also went horribly wrong when
2277        "newline" was defined as CRLF. This has been fixed.
2278    
2279     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2280        was being set to -1 for the "end of line" case (supposedly a value that no
2281        character can have). Though this value is never used (the check for end of
2282        line is "zero bytes in current character"), it caused compiler complaints.
2283        I've changed it to 0xffffffff.
2284    
2285     8. In pcre_version.c, the version string was being built by a sequence of
2286        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2287        string (as it is for production releases) called a macro with an empty
2288        argument. The C standard says the result of this is undefined. The gcc
2289        compiler treats it as an empty string (which was what was wanted) but it is
2290        reported that Visual C gives an error. The source has been hacked around to
2291        avoid this problem.
2292    
2293     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2294        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2295        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2296        of them did).
2297    
2298    10. Originally, pcretest opened its input and output without "b"; then I was
2299        told that "b" was needed in some environments, so it was added for release
2300        5.0 to both the input and output. (It makes no difference on Unix-like
2301        systems.) Later I was told that it is wrong for the input on Windows. I've
2302        now abstracted the modes into two macros, to make it easier to fiddle with
2303        them, and removed "b" from the input mode under Windows.
2304    
2305    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2306    
2307    12. Added -help and --help to pcretest as an official way of being reminded
2308        of the options.
2309    
2310    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2311        and pcrecpp.cc because they annoy compilers at high warning levels.
2312    
2313    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2314    
2315    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2316        scripts are not C programs :-) and which was not noticed because it works
2317        on Linux.
2318    
2319    16. pcretest is supposed to handle any length of pattern and data line (as one
2320        line or as a continued sequence of lines) by extending its input buffer if
2321        necessary. This feature was broken for very long pattern lines, leading to
2322        a string of junk being passed to pcre_compile() if the pattern was longer
2323        than about 50K.
2324    
2325    17. I have done a major re-factoring of the way pcre_compile() computes the
2326        amount of memory needed for a compiled pattern. Previously, there was code
2327        that made a preliminary scan of the pattern in order to do this. That was
2328        OK when PCRE was new, but as the facilities have expanded, it has become
2329        harder and harder to keep it in step with the real compile phase, and there
2330        have been a number of bugs (see for example, 4 above). I have now found a
2331        cunning way of running the real compile function in a "fake" mode that
2332        enables it to compute how much memory it would need, while actually only
2333        ever using a few hundred bytes of working memory and without too many
2334        tests of the mode. This should make future maintenance and development
2335        easier. A side effect of this work is that the limit of 200 on the nesting
2336        depth of parentheses has been removed (though this was never a serious
2337        limitation, I suspect). However, there is a downside: pcre_compile() now
2338        runs more slowly than before (30% or more, depending on the pattern). I
2339        hope this isn't a big issue. There is no effect on runtime performance.
2340    
2341    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2342        newline (only possible for the last line of a file) and it was a
2343        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2344    
2345    19. Added additional timing features to pcretest. (1) The -tm option now times
2346        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2347        separate command line item, by a number that specifies the number of
2348        repeats to use when timing. The default is 50000; this gives better
2349        precision, but takes uncomfortably long for very large patterns.
2350    
2351    20. Extended pcre_study() to be more clever in cases where a branch of a
2352        subpattern has no definite first character. For example, (a*|b*)[cd] would
2353        previously give no result from pcre_study(). Now it recognizes that the
2354        first character must be a, b, c, or d.
2355    
2356    21. There was an incorrect error "recursive call could loop indefinitely" if
2357        a subpattern (or the entire pattern) that was being tested for matching an
2358        empty string contained only one non-empty item after a nested subpattern.
2359        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2360        incorrectly, because the \d was being skipped in the check.
2361    
2362    22. The pcretest program now has a new pattern option /B and a command line
2363        option -b, which is equivalent to adding /B to every pattern. This causes
2364        it to show the compiled bytecode, without the additional information that
2365        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2366        is the same as /B/I).
2367    
2368    23. A new optimization is now able automatically to treat some sequences such
2369        as a*b as a*+b. More specifically, if something simple (such as a character
2370        or a simple class like \d) has an unlimited quantifier, and is followed by
2371        something that cannot possibly match the quantified thing, the quantifier
2372        is automatically "possessified".
2373    
2374    24. A recursive reference to a subpattern whose number was greater than 39
2375        went wrong under certain circumstances in UTF-8 mode. This bug could also
2376        have affected the operation of pcre_study().
2377    
2378    25. Realized that a little bit of performance could be had by replacing
2379        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2380    
2381    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2382    
2383    27. Possessive quantifiers such as a++ were previously implemented by turning
2384        them into atomic groups such as ($>a+). Now they have their own opcodes,
2385        which improves performance. This includes the automatically created ones
2386        from 23 above.
2387    
2388    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2389        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2390        the first matched character to be a colon. This applied both to named and
2391        numbered groups.
2392    
2393    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2394    
2395    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2396        building PCRE on Windows easier, so I have included it in the distribution.
2397    
2398    31. There is now a check in pcretest against a ridiculously large number being
2399        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2400        loop, the loop is abandoned.
2401    
2402    32. Forward references to subpatterns in conditions such as (?(2)...) where
2403        subpattern 2 is defined later cause pcre_compile() to search forwards in
2404        the pattern for the relevant set of parentheses. This search went wrong
2405        when there were unescaped parentheses in a character class, parentheses
2406        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2407    
2408    33. "Subroutine" calls and backreferences were previously restricted to
2409        referencing subpatterns earlier in the regex. This restriction has now
2410        been removed.
2411    
2412    34. Added a number of extra features that are going to be in Perl 5.10. On the
2413        whole, these are just syntactic alternatives for features that PCRE had
2414        previously implemented using the Python syntax or my own invention. The
2415        other formats are all retained for compatibility.
2416    
2417        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2418            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2419            also .NET compatible.
2420    
2421        (b) A recursion or subroutine call to a named group can now be defined as
2422            (?&name) as well as (?P>name).
2423    
2424        (c) A backreference to a named group can now be defined as \k<name> or
2425            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2426            5.10, are also .NET compatible.
2427    
2428        (d) A conditional reference to a named group can now use the syntax
2429            (?(<name>) or (?('name') as well as (?(name).
2430    
2431        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2432            groups (named and numbered) that are never evaluated inline, but can be
2433            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2434            is always false. There may be only one alternative in such a group.
2435    
2436        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2437            as the simple (?(R). The condition is true only if the most recent
2438            recursion is that of the given number or name. It does not search out
2439            through the entire recursion stack.
2440    
2441        (g) The escape \gN or \g{N} has been added, where N is a positive or
2442            negative number, specifying an absolute or relative reference.
2443    
2444    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2445        some "unreachable code" warnings.
2446    
2447    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2448        things, this adds five new scripts.
2449    
2450    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2451        There were also incompatibilities regarding the handling of \Q..\E inside
2452        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2453        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2454    
2455    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2456        matches an empty string, and forcibly breaks the loop. There were bugs in
2457        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2458        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2459        separate and independent bugs (that affected different cases) have been
2460        fixed.
2461    
2462    39. Refactored the code to abolish the use of different opcodes for small
2463        capturing bracket numbers. This is a tidy that I avoided doing when I
2464        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2465        The new approach is not only tidier, it makes it possible to reduce the
2466        memory needed to fix the previous bug (38).
2467    
2468    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2469        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2470        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2471        mode.
2472    
2473    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2474        report.
2475    
2476    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2477        copy construction and assignment in the C++ wrapper.
2478    
2479    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2480        couple of bugs that could have given wrong results in the "--newline=crlf"
2481        case.
2482    
2483    44. Added a number of casts and did some reorganization of signed/unsigned int
2484        variables following suggestions from Dair Grant. Also renamed the variable
2485        "this" as "item" because it is a C++ keyword.
2486    
2487    45. Arranged for dftables to add
2488    
2489          #include "pcre_internal.h"
2490    
2491        to pcre_chartables.c because without it, gcc 4.x may remove the array
2492        definition from the final binary if PCRE is built into a static library and
2493        dead code stripping is activated.
2494    
2495    46. For an unanchored pattern, if a match attempt fails at the start of a
2496        newline sequence, and the newline setting is CRLF or ANY, and the next two
2497        characters are CRLF, advance by two characters instead of one.
2498    
2499    
2500    Version 6.7 04-Jul-06
2501    ---------------------
2502    
2503     1. In order to handle tests when input lines are enormously long, pcretest has
2504        been re-factored so that it automatically extends its buffers when
2505        necessary. The code is crude, but this _is_ just a test program. The
2506        default size has been increased from 32K to 50K.
2507    
2508     2. The code in pcre_study() was using the value of the re argument before
2509        testing it for NULL. (Of course, in any sensible call of the function, it
2510        won't be NULL.)
2511    
2512     3. The memmove() emulation function in pcre_internal.h, which is used on
2513        systems that lack both memmove() and bcopy() - that is, hardly ever -
2514        was missing a "static" storage class specifier.
2515    
2516     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2517        containing an extended class (one that cannot be represented by a bitmap
2518        because it contains high-valued characters or Unicode property items, e.g.
2519        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2520        pattern, but PCRE should not loop if you do not (it no longer does).
2521        [Detail: two cases were found: (a) a repeated subpattern containing an
2522        extended class; (b) a recursive reference to a subpattern that followed a
2523        previous extended class. It wasn't skipping over the extended class
2524        correctly when UTF-8 mode was not set.]
2525    
2526     5. A negated single-character class was not being recognized as fixed-length
2527        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2528        compile error "lookbehind assertion is not fixed length".
2529    
2530     6. The RunPerlTest auxiliary script was showing an unexpected difference
2531        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2532        write a Perl script that can interpret lines of an input file either as
2533        byte characters or as UTF-8, which is what "perltest" was being required to
2534        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2535        can't do is switch easily at run time between having the "use utf8;" pragma
2536        or not. In the end, I fudged it by using the RunPerlTest script to insert
2537        "use utf8;" explicitly for the UTF-8 tests.
2538    
2539     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2540        the end of the subject string, contrary to the documentation and to what
2541        Perl does. This was true of both matching functions. Now it matches only at
2542        the start of the subject and immediately after *internal* newlines.
2543    
2544     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2545        a pointer to an int instead of a pointer to an unsigned long int. This
2546        caused problems on 64-bit systems.
2547    
2548     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2549        instance of the 'standard' template library not being so standard".
2550    
2551    10. There was no check on the number of named subpatterns nor the maximum
2552        length of a subpattern name. The product of these values is used to compute
2553        the size of the memory block for a compiled pattern. By supplying a very
2554        long subpattern name and a large number of named subpatterns, the size
2555        computation could be caused to overflow. This is now prevented by limiting
2556        the length of names to 32 characters, and the number of named subpatterns
2557        to 10,000.
2558    
2559    11. Subpatterns that are repeated with specific counts have to be replicated in
2560        the compiled pattern. The size of memory for this was computed from the
2561        length of the subpattern and the repeat count. The latter is limited to
2562        65535, but there was no limit on the former, meaning that integer overflow
2563        could in principle occur. The compiled length of a repeated subpattern is
2564        now limited to 30,000 bytes in order to prevent this.
2565    
2566    12. Added the optional facility to have named substrings with the same name.
2567    
2568    13. Added the ability to use a named substring as a condition, using the
2569        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2570        are numbers (not recommended). Forward references are permitted.
2571    
2572    14. Added forward references in named backreferences (if you see what I mean).
2573    
2574    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2575        pattern could run off the end of the subject. For example, the pattern
2576        "(?s)(.{1,5})"8 did this with the subject "ab".
2577    
2578    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2579        PCRE_CASELESS was set when matching characters that were quantified with ?
2580        or *.
2581    
2582    17. A character class other than a single negated character that had a minimum
2583        but no maximum quantifier - for example [ab]{6,} - was not handled
2584        correctly by pce_dfa_exec(). It would match only one character.
2585    
2586    18. A valid (though odd) pattern that looked like a POSIX character
2587        class but used an invalid character after [ (for example [[,abc,]]) caused
2588        pcre_compile() to give the error "Failed: internal error: code overflow" or
2589        in some cases to crash with a glibc free() error. This could even happen if
2590        the pattern terminated after [[ but there just happened to be a sequence of
2591        letters, a binary zero, and a closing ] in the memory that followed.
2592    
2593    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2594        over the years. Originally (before any Unicode support), just the bottom 8
2595        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2596        output from "man perlunicode" includes this:
2597    
2598          The regular expression compiler produces polymorphic opcodes.  That
2599          is, the pattern adapts to the data and automatically switches to
2600          the Unicode character scheme when presented with Unicode data--or
2601          instead uses a traditional byte scheme when presented with byte
2602          data.
2603    
2604        Sadly, a wide octal escape does not cause a switch, and in a string with
2605        no other multibyte characters, these octal escapes are treated as before.
2606        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2607        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2608        Unicode string.
2609    
2610        I have not perpetrated such confusion in PCRE. Up till now, it took just
2611        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2612        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2613        translate to the appropriate multibyte character.
2614    
2615    29. Applied some refactoring to reduce the number of warnings from Microsoft
2616        and Borland compilers. This has included removing the fudge introduced
2617        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2618        a warning about an unused variable.
2619    
2620    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2621        characters since release 4.0, because Perl (from release 5.004) does not.
2622        [Or at least, is documented not to: some releases seem to be in conflict
2623        with the documentation.] However, when a pattern was studied with
2624        pcre_study() and all its branches started with \s, PCRE still included VT
2625        as a possible starting character. Of course, this did no harm; it just
2626        caused an unnecessary match attempt.
2627    
2628    22. Removed a now-redundant internal flag bit that recorded the fact that case
2629        dependency changed within the pattern. This was once needed for "required
2630        byte" processing, but is no longer used. This recovers a now-scarce options
2631        bit. Also moved the least significant internal flag bit to the most-
2632        significant bit of the word, which was not previously used (hangover from
2633        the days when it was an int rather than a uint) to free up another bit for
2634        the future.
2635    
2636    23. Added support for CRLF line endings as well as CR and LF. As well as the
2637        default being selectable at build time, it can now be changed at runtime
2638        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2639        specify that it is scanning data with non-default line endings.
2640    
2641    24. Changed the definition of CXXLINK to make it agree with the definition of
2642        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2643    
2644    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2645        recursions. This makes a big different to stack usage for some patterns.
2646    
2647    26. If a subpattern containing a named recursion or subroutine reference such
2648        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2649        the space required for the compiled pattern went wrong and gave too small a
2650        value. Depending on the environment, this could lead to "Failed: internal
2651        error: code overflow at offset 49" or "glibc detected double free or
2652        corruption" errors.
2653    
2654    27. Applied patches from Google (a) to support the new newline modes and (b) to
2655        advance over multibyte UTF-8 characters in GlobalReplace.
2656    
2657    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2658        difference for some implementation of PCRE in some Windows version.
2659    
2660    29. Added some extra testing facilities to pcretest:
2661    
2662        \q<number>   in a data line sets the "match limit" value
2663        \Q<number>   in a data line sets the "match recursion limt" value
2664        -S <number>  sets the stack size, where <number> is in megabytes
2665    
2666        The -S option isn't available for Windows.
2667    
2668    
2669    Version 6.6 06-Feb-06
2670    ---------------------
2671    
2672     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2673        in pcreposix.h. I have copied the definition from pcre.h.
2674    
2675     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2676        because pcre.h is no longer a built file.
2677    
2678     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2679        not normally included in the compiled code.
2680    
2681    
2682    Version 6.5 01-Feb-06
2683    ---------------------
2684    
2685     1. When using the partial match feature with pcre_dfa_exec(), it was not
2686        anchoring the second and subsequent partial matches at the new starting
2687        point. This could lead to incorrect results. For example, with the pattern
2688        /1234/, partially matching against "123" and then "a4" gave a match.
2689    
2690     2. Changes to pcregrep:
2691    
2692        (a) All non-match returns from pcre_exec() were being treated as failures
2693            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2694            error message is output. Some extra information is given for the
2695            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2696            probably the only errors that are likely to be caused by users (by
2697            specifying a regex that has nested indefinite repeats, for instance).
2698            If there are more than 20 of these errors, pcregrep is abandoned.
2699    
2700        (b) A binary zero was treated as data while matching, but terminated the
2701            output line if it was written out. This has been fixed: binary zeroes
2702            are now no different to any other data bytes.
2703    
2704        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2705            used to set a locale for matching. The --locale=xxxx long option has
2706            been added (no short equivalent) to specify a locale explicitly on the
2707            pcregrep command, overriding the environment variables.
2708    
2709        (d) When -B was used with -n, some line numbers in the output were one less
2710            than they should have been.
2711    
2712        (e) Added the -o (--only-matching) option.
2713    
2714        (f) If -A or -C was used with -c (count only), some lines of context were
2715            accidentally printed for the final match.
2716    
2717        (g) Added the -H (--with-filename) option.
2718    
2719        (h) The combination of options -rh failed to suppress file names for files
2720            that were found from directory arguments.
2721    
2722        (i) Added the -D (--devices) and -d (--directories) options.
2723    
2724        (j) Added the -F (--fixed-strings) option.
2725    
2726        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2727    
2728        (l) Added the --colo(u)r option.
2729    
2730        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2731            is not present by default.
2732    
2733     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2734        items such as (?R) or (?1), when the recursion could match a number of
2735        alternatives. If it matched one of the alternatives, but subsequently,
2736        outside the recursion, there was a failure, the code tried to back up into
2737        the recursion. However, because of the way PCRE is implemented, this is not
2738        possible, and the result was an incorrect result from the match.
2739    
2740        In order to prevent this happening, the specification of recursion has
2741        been changed so that all such subpatterns are automatically treated as
2742        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2743    
2744     4. I had overlooked the fact that, in some locales, there are characters for
2745        which isalpha() is true but neither isupper() nor islower() are true. In
2746        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2747        and ordfeminine) are like this. This affected the treatment of \w and \W
2748        when they appeared in character classes, but not when they appeared outside
2749        a character class. The bit map for "word" characters is now created
2750        separately from the results of isalnum() instead of just taking it from the
2751        upper, lower, and digit maps. (Plus the underscore character, of course.)
2752    
2753     5. The above bug also affected the handling of POSIX character classes such as
2754        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2755        permanent tables. Instead, the bit maps for such a class were previously
2756        created as the appropriate unions of the upper, lower, and digit bitmaps.
2757        Now they are created by subtraction from the [[:word:]] class, which has
2758        its own bitmap.
2759    
2760     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2761        It is created by subtracting the vertical space characters (\x09, \x0a,
2762        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2763        subtraction was done in the overall bitmap for a character class, meaning
2764        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2765        be recognized. This bug has been fixed.
2766    
2767     7. Patches from the folks at Google:
2768    
2769          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2770          real life, but is still worth protecting against".
2771    
2772          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2773          regular expressions".
2774    
2775          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2776          have it.
2777    
2778          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2779          "configure" and the latter not, in order to fix a problem somebody had
2780          with compiling the Arg class on HP-UX.
2781    
2782          (e) Improve the error-handling of the C++ wrapper a little bit.
2783    
2784          (f) New tests for checking recursion limiting.
2785    
2786     8. The pcre_memmove() function, which is used only if the environment does not
2787        have a standard memmove() function (and is therefore rarely compiled),
2788        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2789        returning a result (though PCRE never actually uses the result).
2790    
2791     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2792        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2793        returned instead of calling malloc() with an overflowing number that would
2794        most likely cause subsequent chaos.
2795    
2796    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2797    
2798    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2799        with this option is matched, the nmatch and pmatch options of regexec() are
2800        ignored.
2801    
2802    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2803        provided in case anyone wants to the the POSIX interface with UTF-8
2804        strings.
2805    
2806    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2807        C++ linking (needed for some HP-UX environments).
2808    
2809    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2810        (unused parameter) and in the pcre_printint() function (omitted "default"
2811        switch label when the default is to do nothing).
2812    
2813    15. Added some code to make it possible, when PCRE is compiled as a C++
2814        library, to replace subject pointers for pcre_exec() with a smart pointer
2815        class, thus making it possible to process discontinuous strings.
2816    
2817    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2818        much the same function. They were added by different people who were trying
2819        to make PCRE easy to compile on non-Unix systems. It has been suggested
2820        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2821        for compiling on Windows systems. I have therefore replaced it with
2822        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2823        defaults to "extern" for C or "extern C" for C++, which works fine on
2824        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2825        SCOPE with something explicit in config.h. In addition:
2826    
2827        (a) pcreposix.h still had just "extern" instead of either of these macros;
2828            I have replaced it with PCRE_DATA_SCOPE.
2829    
2830        (b) Functions such as _pcre_xclass(), which are internal to the library,
2831            but external in the C sense, all had PCRE_EXPORT in their definitions.
2832            This is apparently wrong for the Windows case, so I have removed it.
2833            (It makes no difference on Unix-like systems.)
2834    
2835    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2836        of recursive calls to match(). This is different to MATCH_LIMIT because
2837        that limits the total number of calls to match(), not all of which increase
2838        the depth of recursion. Limiting the recursion depth limits the amount of
2839        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2840        when PCRE is compiled, and changed at run time. A patch from Google adds
2841        this functionality to the C++ interface.
2842    
2843    18. Changes to the handling of Unicode character properties:
2844    
2845        (a) Updated the table to Unicode 4.1.0.
2846    
2847        (b) Recognize characters that are not in the table as "Cn" (undefined).
2848    
2849        (c) I revised the way the table is implemented to a much improved format
2850            which includes recognition of ranges. It now supports the ranges that
2851            are defined in UnicodeData.txt, and it also amalgamates other
2852            characters into ranges. This has reduced the number of entries in the
2853            table from around 16,000 to around 3,000, thus reducing its size
2854            considerably. I realized I did not need to use a tree structure after
2855            all - a binary chop search is just as efficient. Having reduced the
2856            number of entries, I extended their size from 6 bytes to 8 bytes to
2857            allow for more data.
2858    
2859        (d) Added support for Unicode script names via properties such as \p{Han}.
2860    
2861    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2862        matching that character.
2863    
2864    20. When matching a repeated Unicode property with a minimum greater than zero,
2865        (for example \pL{2,}), PCRE could look past the end of the subject if it
2866        reached it while seeking the minimum number of characters. This could
2867        happen only if some of the characters were more than one byte long, because
2868        there is a check for at least the minimum number of bytes.
2869    
2870    21. Refactored the implementation of \p and \P so as to be more general, to
2871        allow for more different types of property in future. This has changed the
2872        compiled form incompatibly. Anybody with saved compiled patterns that use
2873        \p or \P will have to recompile them.
2874    
2875    22. Added "Any" and "L&" to the supported property types.
2876    
2877    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2878        but give a compile time error if the value is greater than 0xff.
2879    
2880    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2881        accidentally not being installed or uninstalled.
2882    
2883    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2884        made were to insert the current release number. This seemed silly, because
2885        it made things harder for people building PCRE on systems that don't run
2886        "configure". I have turned pcre.h into a distributed file, no longer built
2887        by "configure", with the version identification directly included. There is
2888        no longer a pcre.h.in file.
2889    
2890        However, this change necessitated a change to the pcre-config script as
2891        well. It is built from pcre-config.in, and one of the substitutions was the
2892        release number. I have updated configure.ac so that ./configure now finds
2893        the release number by grepping pcre.h.
2894    
2895    26. Added the ability to run the tests under valgrind.
2896    
2897    
2898    Version 6.4 05-Sep-05
2899    ---------------------
2900    
2901     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2902        "--" to be printed when multiple files were scanned, even when none of the
2903        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2904        consider it to be a bug, and have restored the previous behaviour.
2905    
2906     2. A couple of code tidies to get rid of compiler warnings.
2907    
2908     3. The pcretest program used to cheat by referring to symbols in the library
2909        whose names begin with _pcre_. These are internal symbols that are not
2910        really supposed to be visible externally, and in some environments it is
2911        possible to suppress them. The cheating is now confined to including
2912        certain files from the library's source, which is a bit cleaner.
2913    
2914     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2915        file's purpose clearer.
2916    
2917     5. Reorganized pcre_ucp_findchar().
2918    
2919    
2920    Version 6.3 15-Aug-05
2921    ---------------------
2922    
2923     1. The file libpcre.pc.in did not have general read permission in the tarball.
2924    
2925     2. There were some problems when building without C++ support:
2926    
2927        (a) If C++ support was not built, "make install" and "make test" still
2928            tried to test it.
2929    
2930        (b) There were problems when the value of CXX was explicitly set. Some
2931            changes have been made to try to fix these, and ...
2932    
2933        (c) --disable-cpp can now be used to explicitly disable C++ support.
2934    
2935        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2936            backslash in a target when C++ was disabled. This confuses some
2937            versions of "make", apparently. Using an intermediate variable solves
2938            this. (Same for CPP_LOBJ.)
2939    
2940     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2941        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2942        necessary on certain architectures.
2943    
2944     4. Added a setting of -export-symbols-regex to the link command to remove
2945        those symbols that are exported in the C sense, but actually are local
2946        within the library, and not documented. Their names all begin with
2947        "_pcre_". This is not a perfect job, because (a) we have to except some
2948        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2949        available (and never for static libraries). I have made a note to try to
2950        find a way round (a) in the future.
2951    
2952    
2953    Version 6.2 01-Aug-05
2954    ---------------------
2955    
2956     1. There was no test for integer overflow of quantifier values. A construction
2957        such as {1111111111111111} would give undefined results. What is worse, if
2958        a minimum quantifier for a parenthesized subpattern overflowed and became
2959        negative, the calculation of the memory size went wrong. This could have
2960        led to memory overwriting.
2961    
2962     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2963    
2964     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2965        operating environments where this matters.
2966    
2967     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2968        PCRE options from within the C++ wrapper.
2969    
2970     5. Named capturing subpatterns were not being correctly counted when a pattern
2971        was compiled. This caused two problems: (a) If there were more than 100
2972        such subpatterns, the calculation of the memory needed for the whole
2973        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2974        back references of the form \12, where the number was greater than 9, were
2975        not recognized as back references, even though there were sufficient
2976        previous subpatterns.
2977    
2978     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2979        versions of gcc, e.g. 2.95.4.
2980    
2981    
2982    Version 6.1 21-Jun-05
2983    ---------------------
2984    
2985     1. There was one reference to the variable "posix" in pcretest.c that was not
2986        surrounded by "#if !defined NOPOSIX".
2987    
2988     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2989        the cross-check on the old pcre_info() function, for the benefit of the
2990        cut-down version of PCRE that is currently imported into Exim.
2991    
2992     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2993        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2994        patterns that start (?i) but I don't think that matters. The use of (?i) is
2995        just an example; this all applies to the other options as well.
2996    
2997     4. Since libtool seems to echo the compile commands it is issuing, the output
2998        from "make" can be reduced a bit by putting "@" in front of each libtool
2999        compile command.
3000    
3001     5. Patch from the folks at Google for configure.in to be a bit more thorough
3002        in checking for a suitable C++ installation before trying to compile the
3003        C++ stuff. This should fix a reported problem when a compiler was present,
3004        but no suitable headers.
3005    
3006     6. The man pages all had just "PCRE" as their title. I have changed them to
3007        be the relevant file name. I have also arranged that these names are
3008        retained in the file doc/pcre.txt, which is a concatenation in text format
3009        of all the man pages except the little individual ones for each function.
3010    
3011     7. The NON-UNIX-USE file had not been updated for the different set of source
3012        files that come with release 6. I also added a few comments about the C++
3013        wrapper.
3014    
3015    
3016    Version 6.0 07-Jun-05
3017    ---------------------
3018    
3019     1. Some minor internal re-organization to help with my DFA experiments.
3020    
3021     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3022        didn't matter for the library itself when fully configured, but did matter
3023        when compiling without UCP support, or within Exim, where the ucp files are
3024        not imported.
3025    
3026     3. Refactoring of the library code to split up the various functions into
3027        different source modules. The addition of the new DFA matching code (see
3028        below) to a single monolithic source would have made it really too
3029        unwieldy, quite apart from causing all the code to be include in a
3030        statically linked application, when only some functions are used. This is
3031        relevant even without the DFA addition now that patterns can be compiled in
3032        one application and matched in another.
3033    
3034        The downside of splitting up is that there have to be some external
3035        functions and data tables that are used internally in different modules of
3036        the library but which are not part of the API. These have all had their
3037        names changed to start with "_pcre_" so that they are unlikely to clash
3038        with other external names.
3039    
3040     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3041        a different (DFA) algorithm. Although it is slower than the original
3042        function, it does have some advantages for certain types of matching
3043        problem.
3044    
3045     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3046        including restarting after a partial match.
3047    
3048     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3049        defined when compiling for Windows was sent to me. I have put it into the
3050        code, though I have no means of testing or verifying it.
3051    
3052     7. Added the pcre_refcount() auxiliary function.
3053    
3054     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3055        match before or at the first newline in the subject string. In pcretest,
3056        the /f option on a pattern can be used to set this.
3057    
3058     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3059        would behave wrongly. This has been present in PCRE since release 4.0.
3060    
3061    10. A number of changes to the pcregrep command:
3062    
3063        (a) Refactored how -x works; insert ^(...)$ instead of setting
3064            PCRE_ANCHORED and checking the length, in preparation for adding
3065            something similar for -w.
3066    
3067        (b) Added the -w (match as a word) option.
3068    
3069        (c) Refactored the way lines are read and buffered so as to have more
3070            than one at a time available.
3071    
3072        (d) Implemented a pcregrep test script.
3073    
3074        (e) Added the -M (multiline match) option. This allows patterns to match
3075            over several lines of the subject. The buffering ensures that at least
3076            8K, or the rest of the document (whichever is the shorter) is available
3077            for matching (and similarly the previous 8K for lookbehind assertions).
3078    
3079        (f) Changed the --help output so that it now says
3080    
3081              -w, --word-regex(p)
3082    
3083            instead of two lines, one with "regex" and the other with "regexp"
3084            because that confused at least one person since the short forms are the
3085            same. (This required a bit of code, as the output is generated
3086            automatically from a table. It wasn't just a text change.)
3087    
3088        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3089            option but starts with a hyphen. Could be a pattern or a path name
3090            starting with a hyphen, for instance.
3091    
3092        (h) "-" can be given as a file name to represent stdin.
3093    
3094        (i) When file names are being printed, "(standard input)" is used for
3095            the standard input, for compatibility with GNU grep. Previously
3096            "<stdin>" was used.
3097    
3098        (j) The option --label=xxx can be used to supply a name to be used for
3099            stdin when file names are being printed. There is no short form.
3100    
3101        (k) Re-factored the options decoding logic because we are going to add
3102            two more options that take data. Such options can now be given in four
3103            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3104    
3105        (l) Added the -A, -B, and -C options for requesting that lines of context
3106            around matches be printed.
3107    
3108        (m) Added the -L option to print the names of files that do not contain
3109            any matching lines, that is, the complement of -l.
3110    
3111        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3112            continue to scan other files.
3113    
3114        (o) The -s option was incorrectly implemented. For compatibility with other
3115            greps, it now suppresses the error message for a non-existent or non-
3116            accessible file (but not the return code). There is a new option called
3117            -q that suppresses the output of matching lines, which was what -s was
3118            previously doing.
3119    
3120        (p) Added --include and --exclude options to specify files for inclusion
3121            and exclusion when recursing.
3122    
3123    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3124        Hopefully, it now does.
3125    
3126    12. Missing cast in pcre_study().
3127    
3128    13. Added an "uninstall" target to the makefile.
3129    
3130    14. Replaced "extern" in the function prototypes in Makefile.in with
3131        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3132        world, but is set differently for Windows.
3133    
3134    15. Added a second compiling function called pcre_compile2(). The only
3135        difference is that it has an extra argument, which is a pointer to an
3136        integer error code. When there is a compile-time failure, this is set
3137        non-zero, in addition to the error test pointer being set to point to an
3138        error message. The new argument may be NULL if no error number is required
3139        (but then you may as well call pcre_compile(), which is now just a
3140        wrapper). This facility is provided because some applications need a
3141        numeric error indication, but it has also enabled me to tidy up the way
3142        compile-time errors are handled in the POSIX wrapper.
3143    
3144    16. Added VPATH=.libs to the makefile; this should help when building with one
3145        prefix path and installing with another. (Or so I'm told by someone who
3146        knows more about this stuff than I do.)
3147    
3148    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3149        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3150        match everything, including newlines. This is not POSIX-compatible, but
3151        somebody wanted the feature. From pcretest it can be activated by using
3152        both the P and the s flags.
3153    
3154    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3155    
3156    19. libpcre.pc was being incorrectly installed as executable.
3157    
3158    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3159        it now also looks for '\r' so that it will work unmodified on Windows.
3160    
3161    21. Added Google's contributed C++ wrapper to the distribution.
3162    
3163    22. Added some untidy missing memory free() calls in pcretest, to keep
3164        Electric Fence happy when testing.
3165    
3166    
3167    
3168    Version 5.0 13-Sep-04
3169    ---------------------
3170    
3171     1. Internal change: literal characters are no longer packed up into items
3172        containing multiple characters in a single byte-string. Each character
3173        is now matched using a separate opcode. However, there may be more than one
3174        byte in the character in UTF-8 mode.
3175    
3176     2. The pcre_callout_block structure has two new fields: pattern_position and
3177        next_item_length. These contain the offset in the pattern to the next match
3178        item, and its length, respectively.
3179    
3180     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3181        insertion of callouts before each pattern item. Added the /C option to
3182        pcretest to make use of this.
3183    
3184     4. On the advice of a Windows user, the lines
3185    
3186          #if defined(_WIN32) || defined(WIN32)
3187          _setmode( _fileno( stdout ), 0x8000 );
3188          #endif  /* defined(_WIN32) || defined(WIN32) */
3189    
3190        have been added to the source of pcretest. This apparently does useful
3191        magic in relation to line terminators.
3192    
3193     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3194        for the benefit of those environments where the "b" makes a difference.
3195    
3196     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3197        to know about it. I have put a hack into configure.in that adds in code
3198        to set GCC=yes if CC=icc. This seems to end up at a point in the
3199        generated configure script that is early enough to affect the setting of
3200        compiler options, which is what is needed, but I have no means of testing
3201        whether it really works. (The user who reported this had patched the
3202        generated configure script, which of course I cannot do.)
3203    
3204        LATER: After change 22 below (new libtool files), the configure script
3205        seems to know about icc (and also ecc). Therefore, I have commented out
3206        this hack in configure.in.
3207    
3208     7. Added support for pkg-config (2 patches were sent in).
3209    
3210     8. Negated POSIX character classes that used a combination of internal tables
3211        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3212        [[:^ascii]]. Typically, they would match almost any characters. The other
3213        POSIX classes were not broken in this way.
3214    
3215     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3216        to find the match, as PCRE was deluded into thinking that the match had to
3217        start at the start point or following a newline. The same bug applied to
3218        patterns with negative forward assertions or any backward assertions
3219        preceding ".*" at the start, unless the pattern required a fixed first
3220        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3221    
3222    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3223        starting at the last subject character, bytes beyond the end of the subject
3224        string were read.
3225    
3226    11. Renamed the variable "class" as "classbits" to make life easier for C++
3227        users. (Previously there was a macro definition, but it apparently wasn't
3228        enough.)
3229    
3230    12. Added the new field "tables" to the extra data so that tables can be passed
3231        in at exec time, or the internal tables can be re-selected. This allows
3232        a compiled regex to be saved and re-used at a later time by a different
3233        program that might have everything at different addresses.
3234    
3235    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3236        -R library as well as a -L library.
3237    
3238    14. The debugging options of pcretest (-d on the command line or D on a
3239        pattern) showed incorrect output for anything following an extended class
3240        that contained multibyte characters and which was followed by a quantifier.
3241    
3242    15. Added optional support for general category Unicode character properties
3243        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3244        support. It adds about 90K to the size of the library. The meanings of the
3245        inbuilt class escapes such as \d and \s have NOT been changed.
3246    
3247    16. Updated pcredemo.c to include calls to free() to release the memory for the
3248        compiled pattern.
3249    
3250    17. The generated file chartables.c was being created in the source directory
3251        instead of in the building directory. This caused the build to fail if the
3252        source directory was different from the building directory, and was
3253        read-only.
3254    
3255    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3256        file. No doubt somebody will tell me if they don't make sense... Also added
3257        Dan Mooney's comments about building on OpenVMS.
3258    
3259    19. Added support for partial matching via the PCRE_PARTIAL option for
3260        pcre_exec() and the \P data escape in pcretest.
3261    
3262    20. Extended pcretest with 3 new pattern features:
3263    
3264        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3265              write the compiled pattern to the file whose name is "rest-of-line".
3266              This is a straight binary dump of the data, with the saved pointer to
3267              the character tables forced to be NULL. The study data, if any, is
3268              written too. After writing, pcretest reads a new pattern.
3269    
3270        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3271              compiled pattern from the given file. There must not be any
3272              occurrences of "<" in the file name (pretty unlikely); if there are,
3273              pcretest will instead treat the initial "<" as a pattern delimiter.
3274              After reading in the pattern, pcretest goes on to read data lines as
3275              usual.
3276    
3277        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3278              and 16-bit fields in a compiled pattern, to simulate a pattern that
3279              was compiled on a host of opposite endianness.
3280    
3281    21. The pcre-exec() function can now cope with patterns that were compiled on
3282        hosts of opposite endianness, with this restriction:
3283    
3284          As for any compiled expression that is saved and used later, the tables
3285          pointer field cannot be preserved; the extra_data field in the arguments
3286          to pcre_exec() should be used to pass in a tables address if a value
3287          other than the default internal tables were used at compile time.
3288    
3289    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3290        now diagnosed as an error. Previously, most of the time, a negative number
3291        would have been treated as zero, but if in addition "ovector" was passed as
3292        NULL, a crash could occur.
3293    
3294    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3295        new versions from the libtool 1.5 distribution (the last one is a copy of
3296        a file called libtool.m4). This seems to have fixed the need to patch
3297        "configure" to support Darwin 1.3 (which I used to do). However, I still
3298        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3299        workstation).
3300    
3301    24. Changed the PCRE licence to be the more standard "BSD" licence.
3302    
3303    
3304    Version 4.5 01-Dec-03
3305    ---------------------
3306    
3307     1. There has been some re-arrangement of the code for the match() function so
3308        that it can be compiled in a version that does not call itself recursively.
3309        Instead, it keeps those local variables that need separate instances for
3310        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3311        needs to "recurse". Keeping track of where control must go is done by means
3312        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3313        hide most of the details from the main code, and operates only if
3314        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3315        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3316        operating.
3317    
3318        To make it easier for callers to provide specially tailored get/free
3319        functions for this usage, two new functions, pcre_stack_malloc, and
3320        pcre_stack_free, are used. They are always called in strict stacking order,
3321        and the size of block requested is always the same.
3322    
3323        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3324        PCRE has been compiled to use the stack or the heap for recursion. The
3325        -C option of pcretest uses this to show which version is compiled.
3326    
3327        A new data escape \S, is added to pcretest; it causes the amounts of store
3328        obtained and freed by both kinds of malloc/free at match time to be added
3329        to the output.
3330    
3331     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3332        what's available on my current Linux desktop machine.
3333    
3334     3. When matching a UTF-8 string, the test for a valid string at the start has
3335        been extended. If start_offset is not zero, PCRE now checks that it points
3336        to a byte that is the start of a UTF-8 character. If not, it returns
3337        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3338        this is necessary because there may be backward assertions in the pattern.
3339        When matching the same subject several times, it may save resources to use
3340        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3341    
3342     4. The code for checking the validity of UTF-8 strings has been tightened so
3343        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3344        containing "overlong sequences".
3345    
3346     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3347        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3348        should have been "&", but it just so happened that all the cases this let
3349        through by mistake were picked up later in the function.
3350    
3351     6. I had used a variable called "isblank" - this is a C99 function, causing
3352        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3353    
3354     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3355        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3356        so the version is shown; (c) stop "make test" echoing "./RunTest".
3357    
3358     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3359    
3360     9. The prototype for memmove() for systems that don't have it was using
3361        size_t, but the inclusion of the header that defines size_t was later. I've
3362        moved the #includes for the C headers earlier to avoid this.
3363    
3364    10. Added some adjustments to the code to make it easier to compiler on certain
3365        special systems:
3366    
3367          (a) Some "const" qualifiers were missing.
3368          (b) Added the macro EXPORT before all exported functions; by default this
3369              is defined to be empty.
3370          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3371              that it reads its output file name as an argument instead of writing
3372              to the standard output and assuming this can be redirected.
3373    
3374    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3375        class containing characters with values greater than 255, PCRE compilation
3376        went into a loop.
3377    
3378    12. A recursive reference to a subpattern that was within another subpattern
3379        that had a minimum quantifier of zero caused PCRE to crash. For example,
3380        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3381        recursion. If the recursively-called subpattern itself had a zero repeat,
3382        that was OK.
3383    
3384    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3385        buffer into which it was copied (for escape processing) was still set at
3386        1024, so long lines caused crashes.
3387    
3388    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3389        "internal error: code overflow...". This applied to any character class
3390        that was followed by a possessive quantifier.
3391    
3392    15. Modified the Makefile to add libpcre.la as a prerequisite for
3393        libpcreposix.la because I was told this is needed for a parallel build to
3394        work.
3395    
3396    16. If a pattern that contained .* following optional items at the start was
3397        studied, the wrong optimizing data was generated, leading to matching
3398        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3399        matching string must start with a or b or c. The correct conclusion for
3400        this pattern is that a match can start with any character.
3401    
3402    
3403    Version 4.4 13-Aug-03
3404    ---------------------
3405    
3406     1. In UTF-8 mode, a character class containing characters with values between
3407        127 and 255 was not handled correctly if the compiled pattern was studied.
3408        In fixing this, I have also improved the studying algorithm for such
3409        classes (slightly).
3410    
3411     2. Three internal functions had redundant arguments passed to them. Removal
3412        might give a very teeny performance improvement.
3413    
3414     3. Documentation bug: the value of the capture_top field in a callout is *one
3415        more than* the number of the hightest numbered captured substring.
3416    
3417     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3418        in incorrectly linking with a previously installed version. They now link
3419        explicitly with libpcre.la.
3420    
3421     5. configure.in no longer needs to recognize Cygwin specially.
3422    
3423     6. A problem in pcre.in for Windows platforms is fixed.
3424    
3425     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3426        pcretest, it used to include the size of the study block as part of its
3427        output. Unfortunately, the structure contains a field that has a different
3428        size on different hardware architectures. This meant that the tests that
3429        showed this size failed. As the block is currently always of a fixed size,
3430        this information isn't actually particularly useful in pcretest output, so
3431        I have just removed it.
3432    
3433     8. Three pre-processor statements accidentally did not start in column 1.
3434        Sadly, there are *still* compilers around that complain, even though
3435        standard C has not required this for well over a decade. Sigh.
3436    
3437     9. In pcretest, the code for checking callouts passed small integers in the
3438        callout_data field, which is a void * field. However, some picky compilers
3439        complained about the casts involved for this on 64-bit systems. Now
3440        pcretest passes the address of the small integer instead, which should get
3441        rid of the warnings.
3442    
3443    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3444        both compile and run time, and gives an error if an invalid UTF-8 sequence
3445        is found. There is a option for disabling this check in cases where the
3446        string is known to be correct and/or the maximum performance is wanted.
3447    
3448    11. In response to a bug report, I changed one line in Makefile.in from
3449    
3450            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3451        to
3452            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3453    
3454        to look similar to other lines, but I have no way of telling whether this
3455        is the right thing to do, as I do not use Windows. No doubt I'll get told
3456        if it's wrong...
3457    
3458    
3459    Version 4.3 21-May-03
3460    ---------------------
3461    
3462    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3463       Makefile.
3464    
3465    2. Some refactoring to improve the quality of the code:
3466    
3467       (i)   The utf8_table... variables are now declared "const".
3468    
3469       (ii)  The code for \cx, which used the "case flipping" table to upper case
3470             lower case letters, now just substracts 32. This is ASCII-specific,
3471             but the whole concept of \cx is ASCII-specific, so it seems
3472             reasonable.
3473    
3474       (iii) PCRE was using its character types table to recognize decimal and
3475             hexadecimal digits in the pattern. This is silly, because it handles
3476             only 0-9, a-f, and A-F, but the character types table is locale-
3477             specific, which means strange things might happen. A private
3478             table is now used for this - though it costs 256 bytes, a table is
3479             much faster than multiple explicit tests. Of course, the standard
3480             character types table is still used for matching digits in subject
3481             strings against \d.
3482    
3483       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3484             ending in _t are). So I've renamed it as ESC_tee.
3485    
3486    3. The first argument for regexec() in the POSIX wrapper should have been
3487       defined as "const".
3488    
3489    4. Changed pcretest to use malloc() for its buffers so that they can be
3490       Electric Fenced for debugging.
3491    
3492    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3493       to read one or more bytes before the start of the subject string. Often this
3494       had no effect on PCRE's behaviour, but in some circumstances it could
3495       provoke a segmentation fault.
3496    
3497    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3498       to try to read one or more bytes before the start of the subject string.
3499    
3500    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3501       UTF-8 support could misbehave in various ways if the subject string
3502       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3503       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3504       back over UTF-8 characters.)
3505    
3506    
3507  Version 4.2 14-Apr-03  Version 4.2 14-Apr-03
3508  ---------------------  ---------------------
3509    

Legend:
Removed from v.67  
changed lines
  Added in v.964

  ViewVC Help
Powered by ViewVC 1.1.5