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

Diff of /code/trunk/ChangeLog

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

revision 98 by ph10, Mon Mar 5 15:55:02 2007 UTC revision 991 by zherczeg, Sun Jul 8 16:44:39 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 7.1 05-Mar-07  Version 8.32
5    ------------
6    
7    1.  Improved JIT compiler optimizations for first character search and single
8        character iterators.
9    
10    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11        Patch by Daniel Richard G.
12    
13    3.  Single character iterator optimizations in the JIT compiler.
14    
15    4.  Improved JIT compiler optimizations for character ranges.
16    
17    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18        Reported by Giuseppe D'Angelo.
19    
20    
21    Version 8.31 06-July-2012
22    -------------------------
23    
24    1.  Fixing a wrong JIT test case and some compiler warnings.
25    
26    2.  Removed a bashism from the RunTest script.
27    
28    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
29        to unsigned type, result still unsigned" that was given by an MS compiler
30        on encountering the code "-sizeof(xxx)".
31    
32    4.  Partial matching support is added to the JIT compiler.
33    
34    5.  Fixed several bugs concerned with partial matching of items that consist
35        of more than one character:
36    
37        (a) /^(..)\1/ did not partially match "aba" because checking references was
38            done on an "all or nothing" basis. This also applied to repeated
39            references.
40    
41        (b) \R did not give a hard partial match if \r was found at the end of the
42            subject.
43    
44        (c) \X did not give a hard partial match after matching one or more
45            characters at the end of the subject.
46    
47        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
48            a partial match for the string "\r".
49    
50        (e) When newline was set to CRLF, the metacharacter "." did not recognize
51            a partial match for a CR character at the end of the subject string.
52    
53    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
54        running pcretest, the text "(JIT)" added to the output whenever JIT is
55        actually used to run the match.
56    
57    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
58        or /S+[+] with a digit between 1 and 7.
59    
60    8.  OP_NOT now supports any UTF character not just single-byte ones.
61    
62    9.  (*MARK) control verb is now supported by the JIT compiler.
63    
64    10. The command "./RunTest list" lists the available tests without actually
65        running any of them. (Because I keep forgetting what they all are.)
66    
67    11. Add PCRE_INFO_MAXLOOKBEHIND.
68    
69    12. Applied a (slightly modified) user-supplied patch that improves performance
70        when the heap is used for recursion (compiled with --disable-stack-for-
71        recursion). Instead of malloc and free for each heap frame each time a
72        logical recursion happens, frames are retained on a chain and re-used where
73        possible. This sometimes gives as much as 30% improvement.
74    
75    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
76        call.
77    
78    14. As documented, (*COMMIT) is now confined to within a positive assertion.
79    
80    15. It is now possible to link pcretest with libedit as an alternative to
81        libreadline.
82    
83    16. (*COMMIT) control verb is now supported by the JIT compiler.
84    
85    17. The Unicode data tables have been updated to Unicode 6.1.0.
86    
87    18. Added --file-list option to pcregrep.
88    
89    19. Added binary file support to pcregrep, including the -a, --binary-files,
90        -I, and --text options.
91    
92    20. The madvise function is renamed for posix_madvise for QNX compatibility
93        reasons. Fixed by Giuseppe D'Angelo.
94    
95    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
96    
97    22. Changed the option for creating symbolic links for 16-bit man pages from
98        -s to -sf so that re-installing does not cause issues.
99    
100    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
101    
102    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
103        matches in certain environments (the workspace was not being correctly
104        retained). Also added to pcre_dfa_exec() a simple plausibility check on
105        some of the workspace data at the beginning of a restart.
106    
107    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
108        was not doing so when it should - probably a typo introduced by SVN 528
109        (change 8.10/14).
110    
111    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
112        \w+ when the character tables indicated that \x{c4} was a word character.
113        There were several related cases, all because the tests for doing a table
114        lookup were testing for characters less than 127 instead of 255.
115    
116    27. If a pattern contains capturing parentheses that are not used in a match,
117        their slots in the ovector are set to -1. For those that are higher than
118        any matched groups, this happens at the end of processing. In the case when
119        there were back references that the ovector was too small to contain
120        (causing temporary malloc'd memory to be used during matching), and the
121        highest capturing number was not used, memory off the end of the ovector
122        was incorrectly being set to -1. (It was using the size of the temporary
123        memory instead of the true size.)
124    
125    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
126        ovector size, it uses memory at the end of the block that it has got.
127    
128    29. Check for an overlong MARK name and give an error at compile time. The
129        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
130    
131    30. JIT compiler update.
132    
133    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
134        Rill for the patch.
135    
136    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
137    
138    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
139    
140    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
141        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
142        was enabled.
143    
144    35. Improve JIT code generation for greedy plus quantifier.
145    
146    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
147        1 to "aa" instead of to an empty string. The bug affected repeated groups
148        that could potentially match an empty string.
149    
150    37. Optimizing single character iterators in JIT.
151    
152    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
153        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
154        codepoints that are too big for the mode are faulted, and in a UTF mode,
155        disallowed codepoints are also faulted.
156    
157    39. If PCRE was compiled with UTF support, in three places in the DFA
158        matcher there was code that should only have been obeyed in UTF mode, but
159        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
160        processing when bytes with values greater than 127 were present. In 16-bit
161        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
162        both cases the values are those that cannot be the first data item in a UTF
163        character. The three items that might have provoked this were recursions,
164        possessively repeated groups, and atomic groups.
165    
166    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
167        and pcregrep, because some OS require shared objects to be explicitly
168        passed to ld, causing the link step to fail if they are not.
169    
170    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
171        mode, patterns that started with \h* or \R* might be incorrectly matched.
172    
173    
174    Version 8.30 04-February-2012
175    -----------------------------
176    
177    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
178        name is defined in ctype.h.
179    
180    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
181        only in quite long subpatterns.
182    
183    3.  Removed the function pcre_info(), which has been obsolete and deprecated
184        since it was replaced by pcre_fullinfo() in February 2000.
185    
186    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
187        match a (*MARK), and the match failed at the start of the subject, a
188        reference to memory before the start of the subject could occur. This bug
189        was introduced by fix 17 of release 8.21.
190    
191    5.  A reference to an unset group with zero minimum repetition was giving
192        totally wrong answers (in non-JavaScript-compatibility mode). For example,
193        /(another)?(\1?)test/ matched against "hello world test". This bug was
194        introduced in release 8.13.
195    
196    6.  Add support for 16-bit character strings (a large amount of work involving
197        many changes and refactorings).
198    
199    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
200        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
201        from a file.
202    
203    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
204        rounding is not applied in this particular case).
205    
206    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
207        if they appear, or are escaped, in patterns.
208    
209    10. Get rid of a number of -Wunused-but-set-variable warnings.
210    
211    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
212        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
213        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
214        also returns the mark "x". This bug applied to capturing parentheses,
215        non-capturing parentheses, and atomic parentheses. It also applied to some
216        assertions.
217    
218    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
219        information out of configure.ac instead of relying on pcre.h.generic, which
220        is not stored in the repository.
221    
222    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
223        -lreadline.
224    
225    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
226    
227    15. Applied Graycode's patch to put the top-level frame on the stack rather
228        than the heap when not using the stack for recursion. This gives a
229        performance improvement in many cases when recursion is not deep.
230    
231    16. Experimental code added to "pcretest -C" to output the stack frame size.
232    
233    
234    Version 8.21 12-Dec-2011
235    ------------------------
236    
237    1.  Updating the JIT compiler.
238    
239    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
240        are added as well.
241    
242    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
243        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
244        calling _pcre_jit_exec. Some extra comments are added.
245    
246    4.  (*MARK) settings inside atomic groups that do not contain any capturing
247        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
248        was introduced by change 18 for 8.20.
249    
250    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
251        ECMA-262 standard.
252    
253    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
254        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
255        This bug was probably introduced by change 9 of 8.13.
256    
257    7.  While fixing 6 above, I noticed that a number of other items were being
258        incorrectly rejected as "not fixed length". This arose partly because newer
259        opcodes had not been added to the fixed-length checking code. I have (a)
260        corrected the bug and added tests for these items, and (b) arranged for an
261        error to occur if an unknown opcode is encountered while checking for fixed
262        length instead of just assuming "not fixed length". The items that were
263        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
264        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
265        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
266    
267    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
268        being incorrectly compiled and would have given unpredicatble results.
269    
270    9.  A possessively repeated subpattern with minimum repeat count greater than
271        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
272        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
273        the first (A) could occur when it should not.
274    
275    10. Add a cast and remove a redundant test from the code.
276    
277    11. JIT should use pcre_malloc/pcre_free for allocation.
278    
279    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
280        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
281        is anything other than /usr, -L is still shown).
282    
283    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
284    
285    14. Perl does not support \N without a following name in a [] class; PCRE now
286        also gives an error.
287    
288    15. If a forward reference was repeated with an upper limit of around 2000,
289        it caused the error "internal error: overran compiling workspace". The
290        maximum number of forward references (including repeats) was limited by the
291        internal workspace, and dependent on the LINK_SIZE. The code has been
292        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
293        the default depends on LINK_SIZE. There is a new upper limit (for safety)
294        of around 200,000 forward references. While doing this, I also speeded up
295        the filling in of repeated forward references.
296    
297    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
298        incorrectly expecting the subject to contain another "a" after the start.
299    
300    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
301        in the match, the SKIP should be ignored. This was not happening; instead
302        the SKIP was being treated as NOMATCH. For patterns such as
303        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
304        tested.
305    
306    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
307        now much more compatible with Perl, in particular in cases where the result
308        is a non-match for a non-anchored pattern. For example, if
309        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
310        "m", where previously it did not return a name. A side effect of this
311        change is that for partial matches, the last encountered mark name is
312        returned, as for non matches. A number of tests that were previously not
313        Perl-compatible have been moved into the Perl-compatible test files. The
314        refactoring has had the pleasing side effect of removing one argument from
315        the match() function, thus reducing its stack requirements.
316    
317    19. If the /S+ option was used in pcretest to study a pattern using JIT,
318        subsequent uses of /S (without +) incorrectly behaved like /S+.
319    
320    21. Retrieve executable code size support for the JIT compiler and fixing
321        some warnings.
322    
323    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
324        not work when the shorter character appeared right at the end of the
325        subject string.
326    
327    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
328        systems.
329    
330    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
331        output it when the /M option is used in pcretest.
332    
333    25. The CheckMan script was not being included in the distribution. Also, added
334        an explicit "perl" to run Perl scripts from the PrepareRelease script
335        because this is reportedly needed in Windows.
336    
337    26. If study data was being save in a file and studying had not found a set of
338        "starts with" bytes for the pattern, the data written to the file (though
339        never used) was taken from uninitialized memory and so caused valgrind to
340        complain.
341    
342    27. Updated RunTest.bat as provided by Sheri Pierce.
343    
344    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
345    
346    29. Computation of memory usage for the table of capturing group names was
347        giving an unnecessarily large value.
348    
349    
350    Version 8.20 21-Oct-2011
351    ------------------------
352    
353    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
354        a POSIX class. After further experiments with Perl, which convinced me that
355        Perl has bugs and confusions, a closing square bracket is no longer allowed
356        in a POSIX name. This bug also affected patterns with classes that started
357        with full stops.
358    
359    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
360        captured substring, but while checking the failing first alternative,
361        substring 1 is temporarily captured. If the output vector supplied to
362        pcre_exec() was not big enough for this capture, the yield of the function
363        was still zero ("insufficient space for captured substrings"). This cannot
364        be totally fixed without adding another stack variable, which seems a lot
365        of expense for a edge case. However, I have improved the situation in cases
366        such as /(a)(b)x|abc/ matched against "abc", where the return code
367        indicates that fewer than the maximum number of slots in the ovector have
368        been set.
369    
370    3.  Related to (2) above: when there are more back references in a pattern than
371        slots in the output vector, pcre_exec() uses temporary memory during
372        matching, and copies in the captures as far as possible afterwards. It was
373        using the entire output vector, but this conflicts with the specification
374        that only 2/3 is used for passing back captured substrings. Now it uses
375        only the first 2/3, for compatibility. This is, of course, another edge
376        case.
377    
378    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
379        main code base, and can be used by building with --enable-jit. When this is
380        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
381        runtime --no-jit option is given.
382    
383    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
384        ovector, the return from the function was zero, implying that there were
385        other matches that did not fit. The correct "exactly full" value is now
386        returned.
387    
388    6.  If a subpattern that was called recursively or as a subroutine contained
389        (*PRUNE) or any other control that caused it to give a non-standard return,
390        invalid errors such as "Error -26 (nested recursion at the same subject
391        position)" or even infinite loops could occur.
392    
393    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
394        computing the minimum length on reaching *ACCEPT, and so ended up with the
395        wrong value of 1 rather than 0. Further investigation indicates that
396        computing a minimum subject length in the presence of *ACCEPT is difficult
397        (think back references, subroutine calls), and so I have changed the code
398        so that no minimum is registered for a pattern that contains *ACCEPT.
399    
400    8.  If (*THEN) was present in the first (true) branch of a conditional group,
401        it was not handled as intended. [But see 16 below.]
402    
403    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
404        Sheri Pierce.
405    
406    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
407        the first byte in a match must be "a".
408    
409    11. Change 17 for 8.13 increased the recursion depth for patterns like
410        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
411        pattern contains any instances of (*THEN). If it does not, the old
412        optimizations are restored. It would be nice to do this on a per-group
413        basis, but at the moment that is not feasible.
414    
415    12. In some environments, the output of pcretest -C is CRLF terminated. This
416        broke RunTest's code that checks for the link size. A single white space
417        character after the value is now allowed for.
418    
419    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
420        For "fr", it uses the Windows-specific input and output files.
421    
422    14. If (*THEN) appeared in a group that was called recursively or as a
423        subroutine, it did not work as intended. [But see next item.]
424    
425    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
426        pattern fragments (but not containing any | characters). If A and B are
427        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
428        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
429        D. In other words, Perl considers parentheses that do not contain any |
430        characters to be part of a surrounding alternative, whereas PCRE was
431        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
432        differently. PCRE now behaves in the same way as Perl, except in the case
433        of subroutine/recursion calls such as (?1) which have in any case always
434        been different (but PCRE had them first :-).
435    
436    16. Related to 15 above: Perl does not treat the | in a conditional group as
437        creating alternatives. Such a group is treated in the same way as an
438        ordinary group without any | characters when processing (*THEN). PCRE has
439        been changed to match Perl's behaviour.
440    
441    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
442        RunGrepTest script failed.
443    
444    18. Change 22 for version 13 caused atomic groups to use more stack. This is
445        inevitable for groups that contain captures, but it can lead to a lot of
446        stack use in large patterns. The old behaviour has been restored for atomic
447        groups that do not contain any capturing parentheses.
448    
449    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
450        suppress the check for a minimum subject length at run time. (If it was
451        given to pcre_exec() or pcre_dfa_exec() it did work.)
452    
453    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
454        fail to work when decoding hex characters in data strings in EBCDIC
455        environments.
456    
457    21. It appears that in at least one Mac OS environment, the isxdigit() function
458        is implemented as a macro that evaluates to its argument more than once,
459        contravening the C 90 Standard (I haven't checked a later standard). There
460        was an instance in pcretest which caused it to go wrong when processing
461        \x{...} escapes in subject strings. The has been rewritten to avoid using
462        things like p++ in the argument of isxdigit().
463    
464    
465    Version 8.13 16-Aug-2011
466    ------------------------
467    
468    1.  The Unicode data tables have been updated to Unicode 6.0.0.
469    
470    2.  Two minor typos in pcre_internal.h have been fixed.
471    
472    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
473        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
474        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
475    
476    4.  There were a number of related bugs in the code for matching backrefences
477        caselessly in UTF-8 mode when codes for the characters concerned were
478        different numbers of bytes. For example, U+023A and U+2C65 are an upper
479        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
480        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
481        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
482        2-byte code at the end of the subject (it thought there wasn't enough data
483        left).
484    
485    5.  Comprehensive information about what went wrong is now returned by
486        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
487        as the output vector has at least 2 elements. The offset of the start of
488        the failing character and a reason code are placed in the vector.
489    
490    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
491        now returned is for the first byte of the failing character, instead of the
492        last byte inspected. This is an incompatible change, but I hope it is small
493        enough not to be a problem. It makes the returned offset consistent with
494        pcre_exec() and pcre_dfa_exec().
495    
496    7.  pcretest now gives a text phrase as well as the error number when
497        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
498        failure, the offset and reason code are output.
499    
500    8.  When \R was used with a maximizing quantifier it failed to skip backwards
501        over a \r\n pair if the subsequent match failed. Instead, it just skipped
502        back over a single character (\n). This seems wrong (because it treated the
503        two characters as a single entity when going forwards), conflicts with the
504        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
505        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
506        has been changed.
507    
508    9.  Some internal refactoring has changed the processing so that the handling
509        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
510        time (the PCRE_DOTALL option was changed this way some time ago: version
511        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
512        which was always a bit of a fudge. It also means that there is one less
513        argument for the match() function, which reduces its stack requirements
514        slightly. This change also fixes an incompatibility with Perl: the pattern
515        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
516    
517    10. More internal refactoring has drastically reduced the number of recursive
518        calls to match() for possessively repeated groups such as (abc)++ when
519        using pcre_exec().
520    
521    11. While implementing 10, a number of bugs in the handling of groups were
522        discovered and fixed:
523    
524        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
525        (a|)*(?1) gave a compile-time internal error.
526        ((a|)+)+  did not notice that the outer group could match an empty string.
527        (^a|^)+   was not marked as anchored.
528        (.*a|.*)+ was not marked as matching at start or after a newline.
529    
530    12. Yet more internal refactoring has removed another argument from the match()
531        function. Special calls to this function are now indicated by setting a
532        value in a variable in the "match data" data block.
533    
534    13. Be more explicit in pcre_study() instead of relying on "default" for
535        opcodes that mean there is no starting character; this means that when new
536        ones are added and accidentally left out of pcre_study(), testing should
537        pick them up.
538    
539    14. The -s option of pcretest has been documented for ages as being an old
540        synonym of -m (show memory usage). I have changed it to mean "force study
541        for every regex", that is, assume /S for every regex. This is similar to -i
542        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
543        using it. It makes it easier to run collections of tests with and without
544        study enabled, and thereby test pcre_study() more easily. All the standard
545        tests are now run with and without -s (but some patterns can be marked as
546        "never study" - see 20 below).
547    
548    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
549        restoration of the capturing data to the outer values was not happening
550        correctly.
551    
552    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
553        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
554        pattern had matched an empty string, and so incorrectly returned a no
555        match.
556    
557    17. There was optimizing code for the last branch of non-capturing parentheses,
558        and also for the obeyed branch of a conditional subexpression, which used
559        tail recursion to cut down on stack usage. Unfortunately, now that there is
560        the possibility of (*THEN) occurring in these branches, tail recursion is
561        no longer possible because the return has to be checked for (*THEN). These
562        two optimizations have therefore been removed. [But see 8.20/11 above.]
563    
564    18. If a pattern containing \R was studied, it was assumed that \R always
565        matched two bytes, thus causing the minimum subject length to be
566        incorrectly computed because \R can also match just one byte.
567    
568    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
569        was incorrectly computed.
570    
571    20. If /S is present twice on a test pattern in pcretest input, it now
572        *disables* studying, thereby overriding the use of -s on the command line
573        (see 14 above). This is necessary for one or two tests to keep the output
574        identical in both cases.
575    
576    21. When (*ACCEPT) was used in an assertion that matched an empty string and
577        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
578    
579    22. When an atomic group that contained a capturing parenthesis was
580        successfully matched, but the branch in which it appeared failed, the
581        capturing was not being forgotten if a higher numbered group was later
582        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
583        group 1 to "a", when in fact it should be unset. This applied to multi-
584        branched capturing and non-capturing groups, repeated or not, and also to
585        positive assertions (capturing in negative assertions does not happen
586        in PCRE) and also to nested atomic groups.
587    
588    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
589        subject after a captured substring, to make it easier to tell which of a
590        number of identical substrings has been captured.
591    
592    24. The way atomic groups are processed by pcre_exec() has been changed so that
593        if they are repeated, backtracking one repetition now resets captured
594        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
595        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
596        "aaa". Previously, it would have been "a". As part of this code
597        refactoring, the way recursive calls are handled has also been changed.
598    
599    25. If an assertion condition captured any substrings, they were not passed
600        back unless some other capturing happened later. For example, if
601        (?(?=(a))a) was matched against "a", no capturing was returned.
602    
603    26. When studying a pattern that contained subroutine calls or assertions,
604        the code for finding the minimum length of a possible match was handling
605        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
606        group 1 called group 2 while simultaneously a separate group 2 called group
607        1). A stack overflow occurred in this case. I have fixed this by limiting
608        the recursion depth to 10.
609    
610    27. Updated RunTest.bat in the distribution to the version supplied by Tom
611        Fortmann. This supports explicit test numbers on the command line, and has
612        argument validation and error reporting.
613    
614    28. An instance of \X with an unlimited repeat could fail if at any point the
615        first character it looked at was a mark character.
616    
617    29. Some minor code refactoring concerning Unicode properties and scripts
618        should reduce the stack requirement of match() slightly.
619    
620    30. Added the '=' option to pcretest to check the setting of unused capturing
621        slots at the end of the pattern, which are documented as being -1, but are
622        not included in the return count.
623    
624    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
625        compiled something random. Now it gives a compile-time error (as does
626        Perl).
627    
628    32. A *MARK encountered during the processing of a positive assertion is now
629        recorded and passed back (compatible with Perl).
630    
631    33. If --only-matching or --colour was set on a pcregrep call whose pattern
632        had alternative anchored branches, the search for a second match in a line
633        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
634        matched the line "0102" twice. The same bug affected patterns that started
635        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
636        twice.
637    
638    34. Previously, PCRE did not allow quantification of assertions. However, Perl
639        does, and because of capturing effects, quantifying parenthesized
640        assertions may at times be useful. Quantifiers are now allowed for
641        parenthesized assertions.
642    
643    35. A minor code tidy in pcre_compile() when checking options for \R usage.
644    
645    36. \g was being checked for fancy things in a character class, when it should
646        just be a literal "g".
647    
648    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
649        appearance of a nested POSIX class supersedes an apparent external class.
650        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
651        unescaped square brackets may also appear as part of class names. For
652        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
653        more like Perl. (But see 8.20/1 above.)
654    
655    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
656        was because it thought it was \N{name}, which is not supported).
657    
658    39. Add minix to OS list not supporting the -S option in pcretest.
659    
660    40. PCRE tries to detect cases of infinite recursion at compile time, but it
661        cannot analyze patterns in sufficient detail to catch mutual recursions
662        such as ((?1))((?2)). There is now a runtime test that gives an error if a
663        subgroup is called recursively as a subpattern for a second time at the
664        same position in the subject string. In previous releases this might have
665        been caught by the recursion limit, or it might have run out of stack.
666    
667    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
668        happen only once. PCRE was, however incorrectly giving a compile time error
669        "recursive call could loop indefinitely" because it cannot analyze the
670        pattern in sufficient detail. The compile time test no longer happens when
671        PCRE is compiling a conditional subpattern, but actual runaway loops are
672        now caught at runtime (see 40 above).
673    
674    42. It seems that Perl allows any characters other than a closing parenthesis
675        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
676        has been changed to be the same.
677    
678    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
679        as not to get warnings when autogen.sh is called. Also changed
680        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
681    
682    44. To help people who use pcregrep to scan files containing exceedingly long
683        lines, the following changes have been made:
684    
685        (a) The default value of the buffer size parameter has been increased from
686            8K to 20K. (The actual buffer used is three times this size.)
687    
688        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
689            PCRE is built.
690    
691        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
692            to be set at run time.
693    
694        (d) Numerical values in pcregrep options can be followed by K or M, for
695            example --buffer-size=50K.
696    
697        (e) If a line being scanned overflows pcregrep's buffer, an error is now
698            given and the return code is set to 2.
699    
700    45. Add a pointer to the latest mark to the callout data block.
701    
702    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
703        partial match of an empty string instead of no match. This was specific to
704        the use of ".".
705    
706    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
707        complete match instead of a partial match. This bug was dependent on both
708        the PCRE_UTF8 and PCRE_DOTALL options being set.
709    
710    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
711        starting byte set, because \b was not being ignored.
712    
713    
714    Version 8.12 15-Jan-2011
715    ------------------------
716    
717    1.  Fixed some typos in the markup of the man pages, and wrote a script that
718        checks for such things as part of the documentation building process.
719    
720    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
721        --match-limit and --recursion-limit options (added for 8.11). In
722        particular, this made one of the standard tests fail. (The integer value
723        went into the wrong half of a long int.)
724    
725    3.  If the --colour option was given to pcregrep with -v (invert match), it
726        did strange things, either producing crazy output, or crashing. It should,
727        of course, ignore a request for colour when reporting lines that do not
728        match.
729    
730    4.  Another pcregrep bug caused similar problems if --colour was specified with
731        -M (multiline) and the pattern match finished with a line ending.
732    
733    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
734        matched in multiline mode, the following line was shown as part of the
735        match. This seems wrong, so I have changed it.
736    
737    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
738        the check for further matches in the same line (so they could be coloured)
739        to overrun the end of the current line. If another match was found, it was
740        incorrectly shown (and then shown again when found in the next line).
741    
742    7.  If pcregrep was compiled under Windows, there was a reference to the
743        function pcregrep_exit() before it was defined. I am assuming this was
744        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
745        reported by a user. I've moved the definition above the reference.
746    
747    
748    Version 8.11 10-Dec-2010
749    ------------------------
750    
751    1.  (*THEN) was not working properly if there were untried alternatives prior
752        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
753        backtracked to try for "b" instead of moving to the next alternative branch
754        at the same level (in this case, to look for "c"). The Perl documentation
755        is clear that when (*THEN) is backtracked onto, it goes to the "next
756        alternative in the innermost enclosing group".
757    
758    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
759        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
760        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
761        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
762        (*THEN).
763    
764    3.  If \s appeared in a character class, it removed the VT character from
765        the class, even if it had been included by some previous item, for example
766        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
767        of \s, but is part of the POSIX "space" class.)
768    
769    4.  A partial match never returns an empty string (because you can always
770        match an empty string at the end of the subject); however the checking for
771        an empty string was starting at the "start of match" point. This has been
772        changed to the "earliest inspected character" point, because the returned
773        data for a partial match starts at this character. This means that, for
774        example, /(?<=abc)def/ gives a partial match for the subject "abc"
775        (previously it gave "no match").
776    
777    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
778        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
779        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
780        has an implication that the given string is incomplete (because a partial
781        match is preferred over a full match). For this reason, these items now
782        give a partial match in this situation. [Aside: previously, the one case
783        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
784        match rather than a full match, which was wrong by the old rules, but is
785        now correct.]
786    
787    6.  There was a bug in the handling of #-introduced comments, recognized when
788        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
789        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
790        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
791        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
792        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
793        places in pcre_compile().
794    
795    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
796        comments when looking ahead for named forward references to subpatterns,
797        the only newline sequence it recognized was NL. It now handles newlines
798        according to the set newline convention.
799    
800    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
801        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
802        cater for a lack of strerror(). These oversights have been fixed.
803    
804    9.  Added --match-limit and --recursion-limit to pcregrep.
805    
806    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
807    
808    11. When the -o option was used, pcregrep was setting a return code of 1, even
809        when matches were found, and --line-buffered was not being honoured.
810    
811    12. Added an optional parentheses number to the -o and --only-matching options
812        of pcregrep.
813    
814    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
815        can match an empty string. The code to do it in pcretest and pcredemo
816        needed fixing:
817    
818        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
819            only one byte after an empty string match just before CRLF (this case
820            just got forgotten; "any" and "anycrlf" were OK).
821    
822        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
823            mode when an empty string match preceded an ASCII character followed by
824            a non-ASCII character. (The code for advancing by one character rather
825            than one byte was nonsense.)
826    
827        (c) The pcredemo.c sample program did not have any code at all to handle
828            the cases when CRLF is a valid newline sequence.
829    
830    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
831        as a starting offset was within the subject string. There is now a new
832        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
833        negative or greater than the length of the string. In order to test this,
834        pcretest is extended to allow the setting of negative starting offsets.
835    
836    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
837        starting offset points to the beginning of a UTF-8 character was
838        unnecessarily clumsy. I tidied it up.
839    
840    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
841        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
842    
843    17. Nobody had reported that the --include_dir option, which was added in
844        release 7.7 should have been called --include-dir (hyphen, not underscore)
845        for compatibility with GNU grep. I have changed it to --include-dir, but
846        left --include_dir as an undocumented synonym, and the same for
847        --exclude-dir, though that is not available in GNU grep, at least as of
848        release 2.5.4.
849    
850    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
851        characters from a string of bytes) have been redefined so as not to use
852        loops, in order to improve performance in some environments. At the same
853        time, I abstracted some of the common code into auxiliary macros to save
854        repetition (this should not affect the compiled code).
855    
856    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
857        compile-time error is now given if \c is not followed by an ASCII
858        character, that is, a byte less than 128. (In EBCDIC mode, the code is
859        different, and any byte value is allowed.)
860    
861    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
862        START_OPTIMIZE option, which is now allowed at compile time - but just
863        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
864        to pcregrep and other applications that have no direct access to PCRE
865        options. The new /Y option in pcretest sets this option when calling
866        pcre_compile().
867    
868    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
869        back references. Groups containing recursive back references were forced to
870        be atomic by that change, but in the case of named groups, the amount of
871        memory required was incorrectly computed, leading to "Failed: internal
872        error: code overflow". This has been fixed.
873    
874    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
875        pcretest.c, to avoid build problems in some Borland environments.
876    
877    
878    Version 8.10 25-Jun-2010
879    ------------------------
880    
881    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
882        THEN.
883    
884    2.  (*ACCEPT) was not working when inside an atomic group.
885    
886    3.  Inside a character class, \B is treated as a literal by default, but
887        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
888        causes the error). The code is unchanged, but I tidied the documentation.
889    
890    4.  Inside a character class, PCRE always treated \R and \X as literals,
891        whereas Perl faults them if its -w option is set. I have changed PCRE so
892        that it faults them when PCRE_EXTRA is set.
893    
894    5.  Added support for \N, which always matches any character other than
895        newline. (It is the same as "." when PCRE_DOTALL is not set.)
896    
897    6.  When compiling pcregrep with newer versions of gcc which may have
898        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
899        declared with attribute warn_unused_result" were given. Just casting the
900        result to (void) does not stop the warnings; a more elaborate fudge is
901        needed. I've used a macro to implement this.
902    
903    7.  Minor change to pcretest.c to avoid a compiler warning.
904    
905    8.  Added four artifical Unicode properties to help with an option to make
906        \s etc use properties (see next item). The new properties are: Xan
907        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
908    
909    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
910        use Unicode properties. (*UCP) at the start of a pattern can be used to set
911        this option. Modified pcretest to add /W to test this facility. Added
912        REG_UCP to make it available via the POSIX interface.
913    
914    10. Added --line-buffered to pcregrep.
915    
916    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
917        studied, and the match started with a letter with a code point greater than
918        127 whose first byte was different to the first byte of the other case of
919        the letter, the other case of this starting letter was not recognized
920        (#976).
921    
922    12. If a pattern that was studied started with a repeated Unicode property
923        test, for example, \p{Nd}+, there was the theoretical possibility of
924        setting up an incorrect bitmap of starting bytes, but fortunately it could
925        not have actually happened in practice until change 8 above was made (it
926        added property types that matched character-matching opcodes).
927    
928    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
929        possible starting bytes for non-anchored patterns.
930    
931    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
932        \R, and also a number of cases that involve Unicode properties, both
933        explicit and implicit when PCRE_UCP is set.
934    
935    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
936        input, it could crash or give wrong results if characters with values
937        greater than 0xc0 were present in the subject string. (Detail: it assumed
938        UTF-8 input when processing these items.)
939    
940    16. Added a lot of (int) casts to avoid compiler warnings in systems where
941        size_t is 64-bit (#991).
942    
943    17. Added a check for running out of memory when PCRE is compiled with
944        --disable-stack-for-recursion (#990).
945    
946    18. If the last data line in a file for pcretest does not have a newline on
947        the end, a newline was missing in the output.
948    
949    19. The default pcre_chartables.c file recognizes only ASCII characters (values
950        less than 128) in its various bitmaps. However, there is a facility for
951        generating tables according to the current locale when PCRE is compiled. It
952        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
953        characters, are recognized by isspace() and therefore were getting set in
954        these tables, and indeed these tables seem to approximate to ISO 8859. This
955        caused a problem in UTF-8 mode when pcre_study() was used to create a list
956        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
957        which of course cannot start UTF-8 characters. I have changed the code so
958        that only real ASCII characters (less than 128) and the correct starting
959        bytes for UTF-8 encodings are set for characters greater than 127 when in
960        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
961        altogether.)
962    
963    20. Added the /T option to pcretest so as to be able to run tests with non-
964        standard character tables, thus making it possible to include the tests
965        used for 19 above in the standard set of tests.
966    
967    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
968        reference to a subpattern the other side of a comment that contains an
969        opening parenthesis caused either an internal compiling error, or a
970        reference to the wrong subpattern.
971    
972    
973    Version 8.02 19-Mar-2010
974    ------------------------
975    
976    1.  The Unicode data tables have been updated to Unicode 5.2.0.
977    
978    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
979        configured.
980    
981    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
982        original author of that file, following a query about its status.
983    
984    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
985        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
986    
987    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
988        quantifier applied to a forward-referencing subroutine call, could compile
989        incorrect code or give the error "internal error: previously-checked
990        referenced subpattern not found".
991    
992    6.  Both MS Visual Studio and Symbian OS have problems with initializing
993        variables to point to external functions. For these systems, therefore,
994        pcre_malloc etc. are now initialized to local functions that call the
995        relevant global functions.
996    
997    7.  There were two entries missing in the vectors called coptable and poptable
998        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
999        I've fixed the data, and added a kludgy way of testing at compile time that
1000        the lengths are correct (equal to the number of opcodes).
1001    
1002    8.  Following on from 7, I added a similar kludge to check the length of the
1003        eint vector in pcreposix.c.
1004    
1005    9.  Error texts for pcre_compile() are held as one long string to avoid too
1006        much relocation at load time. To find a text, the string is searched,
1007        counting zeros. There was no check for running off the end of the string,
1008        which could happen if a new error number was added without updating the
1009        string.
1010    
1011    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1012    
1013    11. \K was not working if it appeared in an atomic group or in a group that
1014        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1015        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1016        the assertion is positive, but not if it is negative.
1017    
1018    12. Change 11 fortuitously reduced the size of the stack frame used in the
1019        "match()" function of pcre_exec.c by one pointer. Forthcoming
1020        implementation of support for (*MARK) will need an extra pointer on the
1021        stack; I have reserved it now, so that the stack frame size does not
1022        decrease.
1023    
1024    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1025        item in branch that calls a recursion is a subroutine call - as in the
1026        second branch in the above example - was incorrectly given the compile-
1027        time error "recursive call could loop indefinitely" because pcre_compile()
1028        was not correctly checking the subroutine for matching a non-empty string.
1029    
1030    14. The checks for overrunning compiling workspace could trigger after an
1031        overrun had occurred. This is a "should never occur" error, but it can be
1032        triggered by pathological patterns such as hundreds of nested parentheses.
1033        The checks now trigger 100 bytes before the end of the workspace.
1034    
1035    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1036    
1037    
1038    Version 8.01 19-Jan-2010
1039    ------------------------
1040    
1041    1.  If a pattern contained a conditional subpattern with only one branch (in
1042        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1043        computed the wrong minimum data length (which is of course zero for such
1044        subpatterns). This could cause incorrect "no match" results.
1045    
1046    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1047        the pattern is reset in the first branch, pcre_compile() failed with
1048        "internal error: code overflow at offset...". This happened only when
1049        the reset was to the original external option setting. (An optimization
1050        abstracts leading options settings into an external setting, which was the
1051        cause of this.)
1052    
1053    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1054        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1055        assertion pattern did not match (meaning that the assertion was true), it
1056        was incorrectly treated as false if the SKIP had been reached during the
1057        matching. This also applied to assertions used as conditions.
1058    
1059    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1060        assertion subpattern, including such a pattern used as a condition,
1061        unpredictable results occurred, instead of the error return
1062        PCRE_ERROR_DFA_UITEM.
1063    
1064    5.  The C++ GlobalReplace function was not working like Perl for the special
1065        situation when an empty string is matched. It now does the fancy magic
1066        stuff that is necessary.
1067    
1068    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1069        removed. (These were left over from very, very early versions of PCRE.)
1070    
1071    7.  Some cosmetic changes to the code to make life easier when compiling it
1072        as part of something else:
1073    
1074        (a) Change DEBUG to PCRE_DEBUG.
1075    
1076        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1077            called "current" as "current_branch", to prevent a collision with the
1078            Linux macro when compiled as a kernel module.
1079    
1080        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1081            prevent a collision with the Linux macro when compiled as a kernel
1082            module.
1083    
1084    8.  In pcre_compile() there are some checks for integer overflows that used to
1085        cast potentially large values to (double). This has been changed to that
1086        when building, a check for int64_t is made, and if it is found, it is used
1087        instead, thus avoiding the use of floating point arithmetic. (There is no
1088        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1089        double.
1090    
1091    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1092        2005 (difference between two addresses compared to an unsigned value).
1093    
1094    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1095        custom one, because of the following reported problem in Windows:
1096    
1097          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1098              under Win32.
1099          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1100              therefore missing the function definition.
1101          - The compiler thus generates a "C" signature for the test function.
1102          - The linker fails to find the "C" function.
1103          - PCRE fails to configure if asked to do so against libbz2.
1104    
1105    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1106        messages were output:
1107    
1108          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1109          rerunning libtoolize, to keep the correct libtool macros in-tree.
1110          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1111    
1112        I have done both of these things.
1113    
1114    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1115        most of the time, it *can* run out if it is given a pattern that contains a
1116        runaway infinite recursion. I updated the discussion in the pcrestack man
1117        page.
1118    
1119    13. Now that we have gone to the x.xx style of version numbers, the minor
1120        version may start with zero. Using 08 or 09 is a bad idea because users
1121        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1122        interpreted as invalid octal numbers. I've updated the previous comment in
1123        configure.ac, and also added a check that gives an error if 08 or 09 are
1124        used.
1125    
1126    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1127        causing partial matching to fail when the end of the subject matched \W
1128        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1129    
1130    15. There were some discrepancies between the declarations in pcre_internal.h
1131        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1132        their definitions. The declarations used "const uschar *" and the
1133        definitions used USPTR. Even though USPTR is normally defined as "const
1134        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1135        reported that: "This difference in casting confuses some C++ compilers, for
1136        example, SunCC recognizes above declarations as different functions and
1137        generates broken code for hbpcre." I have changed the declarations to use
1138        USPTR.
1139    
1140    16. GNU libtool is named differently on some systems. The autogen.sh script now
1141        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1142        (FreeBSD).
1143    
1144    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1145        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1146        comment: "Figure out how to create a longlong from a string: strtoll and
1147        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1148        instance, but it only takes 2 args instead of 3!"
1149    
1150    18. A subtle bug concerned with back references has been fixed by a change of
1151        specification, with a corresponding code fix. A pattern such as
1152        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1153        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1154        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1155        same bug. Such groups have to be quantified to be useful, or contained
1156        inside another quantified group. (If there's no repetition, the reference
1157        can never match.) The problem arises because, having left the group and
1158        moved on to the rest of the pattern, a later failure that backtracks into
1159        the group uses the captured value from the final iteration of the group
1160        rather than the correct earlier one. I have fixed this in PCRE by forcing
1161        any group that contains a reference to itself to be an atomic group; that
1162        is, there cannot be any backtracking into it once it has completed. This is
1163        similar to recursive and subroutine calls.
1164    
1165    
1166    Version 8.00 19-Oct-09
1167    ----------------------
1168    
1169    1.  The table for translating pcre_compile() error codes into POSIX error codes
1170        was out-of-date, and there was no check on the pcre_compile() error code
1171        being within the table. This could lead to an OK return being given in
1172        error.
1173    
1174    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1175        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1176        in a Windows environment.
1177    
1178    3.  The pcregrep --count option prints the count for each file even when it is
1179        zero, as does GNU grep. However, pcregrep was also printing all files when
1180        --files-with-matches was added. Now, when both options are given, it prints
1181        counts only for those files that have at least one match. (GNU grep just
1182        prints the file name in this circumstance, but including the count seems
1183        more useful - otherwise, why use --count?) Also ensured that the
1184        combination -clh just lists non-zero counts, with no names.
1185    
1186    4.  The long form of the pcregrep -F option was incorrectly implemented as
1187        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1188        but it seems right to fix it, and I didn't think it was worth preserving
1189        the old behaviour.
1190    
1191    5.  The command line items --regex=pattern and --regexp=pattern were not
1192        recognized by pcregrep, which required --regex pattern or --regexp pattern
1193        (with a space rather than an '='). The man page documented the '=' forms,
1194        which are compatible with GNU grep; these now work.
1195    
1196    6.  No libpcreposix.pc file was created for pkg-config; there was just
1197        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1198    
1199    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1200        when UCP support is not needed, by modifying the Python script that
1201        generates it from Unicode data files. This should not matter if the module
1202        is correctly used as a library, but I received one complaint about 50K of
1203        unwanted data. My guess is that the person linked everything into his
1204        program rather than using a library. Anyway, it does no harm.
1205    
1206    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1207        was a minimum greater than 1 for a wide character in a possessive
1208        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1209        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1210        character. Chaos in the form of incorrect output or a compiling loop could
1211        result.
1212    
1213    9.  The restrictions on what a pattern can contain when partial matching is
1214        requested for pcre_exec() have been removed. All patterns can now be
1215        partially matched by this function. In addition, if there are at least two
1216        slots in the offset vector, the offset of the earliest inspected character
1217        for the match and the offset of the end of the subject are set in them when
1218        PCRE_ERROR_PARTIAL is returned.
1219    
1220    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1221        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1222        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1223        and may be more useful for multi-segment matching.
1224    
1225    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1226        used to be given if ever the end of the subject was reached; now it is
1227        given only if matching could not proceed because another character was
1228        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1229        string "Z", which now yields "no match" instead of "partial match". In the
1230        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1231        final character ended with (*FAIL).
1232    
1233    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1234        if the pattern had a "must contain" character that was already found in the
1235        earlier partial match, unless partial matching was again requested. For
1236        example, with the pattern /dog.(body)?/, the "must contain" character is
1237        "g". If the first part-match was for the string "dog", restarting with
1238        "sbody" failed. This bug has been fixed.
1239    
1240    13. The string returned by pcre_dfa_exec() after a partial match has been
1241        changed so that it starts at the first inspected character rather than the
1242        first character of the match. This makes a difference only if the pattern
1243        starts with a lookbehind assertion or \b or \B (\K is not supported by
1244        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1245        matching functions compatible, and I think it's the right thing to do.
1246    
1247    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1248        so that the demonstration program is easily available in environments where
1249        PCRE has not been installed from source.
1250    
1251    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1252        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1253        library.
1254    
1255    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1256        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1257        is not the first non-POSIX option to be added. Clearly some people find
1258        these options useful.
1259    
1260    17. If a caller to the POSIX matching function regexec() passes a non-zero
1261        value for nmatch with a NULL value for pmatch, the value of
1262        nmatch is forced to zero.
1263    
1264    18. RunGrepTest did not have a test for the availability of the -u option of
1265        the diff command, as RunTest does. It now checks in the same way as
1266        RunTest, and also checks for the -b option.
1267    
1268    19. If an odd number of negated classes containing just a single character
1269        interposed, within parentheses, between a forward reference to a named
1270        subpattern and the definition of the subpattern, compilation crashed with
1271        an internal error, complaining that it could not find the referenced
1272        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1273        [The bug was that it was starting one character too far in when skipping
1274        over the character class, thus treating the ] as data rather than
1275        terminating the class. This meant it could skip too much.]
1276    
1277    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1278        /g option in pcretest when the pattern contains \K, which makes it possible
1279        to have an empty string match not at the start, even when the pattern is
1280        anchored. Updated pcretest and pcredemo to use this option.
1281    
1282    21. If the maximum number of capturing subpatterns in a recursion was greater
1283        than the maximum at the outer level, the higher number was returned, but
1284        with unset values at the outer level. The correct (outer level) value is
1285        now given.
1286    
1287    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1288        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1289        make it do so. The string so far is captured, making this feature
1290        compatible with Perl.
1291    
1292    23. The tests have been re-organized, adding tests 11 and 12, to make it
1293        possible to check the Perl 5.10 features against Perl 5.10.
1294    
1295    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1296        pattern matches a fixed length string. PCRE did not allow this; now it
1297        does. Neither allows recursion.
1298    
1299    25. I finally figured out how to implement a request to provide the minimum
1300        length of subject string that was needed in order to match a given pattern.
1301        (It was back references and recursion that I had previously got hung up
1302        on.) This code has now been added to pcre_study(); it finds a lower bound
1303        to the length of subject needed. It is not necessarily the greatest lower
1304        bound, but using it to avoid searching strings that are too short does give
1305        some useful speed-ups. The value is available to calling programs via
1306        pcre_fullinfo().
1307    
1308    26. While implementing 25, I discovered to my embarrassment that pcretest had
1309        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1310        study optimizations had never been tested with that matching function.
1311        Oops. What is worse, even when it was passed study data, there was a bug in
1312        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1313        were also very few tests of studied patterns with pcre_dfa_exec().
1314    
1315    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1316        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1317        on the other side of the coin, they are no longer allowed to have different
1318        names, because these cannot be distinguished in PCRE, and this has caused
1319        confusion. (This is a difference from Perl.)
1320    
1321    28. When duplicate subpattern names are present (necessarily with different
1322        numbers, as required by 27 above), and a test is made by name in a
1323        conditional pattern, either for a subpattern having been matched, or for
1324        recursion in such a pattern, all the associated numbered subpatterns are
1325        tested, and the overall condition is true if the condition is true for any
1326        one of them. This is the way Perl works, and is also more like the way
1327        testing by number works.
1328    
1329    
1330    Version 7.9 11-Apr-09
1331    ---------------------
1332    
1333    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1334        (pcretest), all targets were linked against these libraries. This included
1335        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1336        libraries. This caused unwanted dependencies to be created. This problem
1337        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1338        pcretest is linked with readline.
1339    
1340    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1341        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1342        moved outside it again, because FALSE and TRUE are already defined in AIX,
1343        but BOOL is not.
1344    
1345    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1346        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1347    
1348    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1349        hyphens) following file names and line numbers when outputting matching
1350        lines. This is not true; no spaces are inserted. I have also clarified the
1351        wording for the --colour (or --color) option.
1352    
1353    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1354        was not coloured; this is different to GNU grep, so I have changed it to be
1355        the same.
1356    
1357    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1358        each matching line was coloured. Now it goes on to look for further matches
1359        of any of the test patterns, which is the same behaviour as GNU grep.
1360    
1361    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1362        doesn't make sense to accept an empty string match in pcregrep, so I have
1363        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1364        seems to be how GNU grep behaves.
1365    
1366    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1367        start or after a newline", because the conditional assertion was not being
1368        correctly handled. The rule now is that both the assertion and what follows
1369        in the first alternative must satisfy the test.
1370    
1371    9.  If auto-callout was enabled in a pattern with a conditional group whose
1372        condition was an assertion, PCRE could crash during matching, both with
1373        pcre_exec() and pcre_dfa_exec().
1374    
1375    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1376        used for matching.
1377    
1378    11. Unicode property support in character classes was not working for
1379        characters (bytes) greater than 127 when not in UTF-8 mode.
1380    
1381    12. Added the -M command line option to pcretest.
1382    
1383    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1384    
1385    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1386    
1387    16. Added comments and documentation about mis-use of no_arg in the C++
1388        wrapper.
1389    
1390    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1391        from Martin Jerabek that uses macro names for all relevant character and
1392        string constants.
1393    
1394    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1395        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1396        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1397        these, but not everybody uses configure.
1398    
1399    19. A conditional group that had only one branch was not being correctly
1400        recognized as an item that could match an empty string. This meant that an
1401        enclosing group might also not be so recognized, causing infinite looping
1402        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1403        with the subject "ab", where knowledge that the repeated group can match
1404        nothing is needed in order to break the loop.
1405    
1406    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1407        exec(), but without supplying a callout function, matching went wrong.
1408    
1409    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1410        leak if the size of the offset vector was greater than 30. When the vector
1411        is smaller, the saved offsets during recursion go onto a local stack
1412        vector, but for larger vectors malloc() is used. It was failing to free
1413        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1414        error, in fact).
1415    
1416    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1417        heapframe that is used only when UTF-8 support is enabled. This caused no
1418        problem, but was untidy.
1419    
1420    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1421        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1422        included within another project.
1423    
1424    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1425        slightly modified by me:
1426    
1427          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1428              not building pcregrep.
1429    
1430          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1431              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1432    
1433    25. Forward references, both numeric and by name, in patterns that made use of
1434        duplicate group numbers, could behave incorrectly or give incorrect errors,
1435        because when scanning forward to find the reference group, PCRE was not
1436        taking into account the duplicate group numbers. A pattern such as
1437        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1438    
1439    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1440        the feature of a custom pointer more persuasive (as requested by a user).
1441    
1442    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1443        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1444        pre-defined.
1445    
1446    28. Added support for (*UTF8) at the start of a pattern.
1447    
1448    29. Arrange for flags added by the "release type" setting in CMake to be shown
1449        in the configuration summary.
1450    
1451    
1452    Version 7.8 05-Sep-08
1453    ---------------------
1454    
1455    1.  Replaced UCP searching code with optimized version as implemented for Ad
1456        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1457        stage table and inline lookup instead of a function, giving speed ups of 2
1458        to 5 times on some simple patterns that I tested. Permission was given to
1459        distribute the MultiStage2.py script that generates the tables (it's not in
1460        the tarball, but is in the Subversion repository).
1461    
1462    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1463        scripts.
1464    
1465    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1466        a group with a zero qualifier. The result of the study could be incorrect,
1467        or the function might crash, depending on the pattern.
1468    
1469    4.  Caseless matching was not working for non-ASCII characters in back
1470        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1471        It now works when Unicode Property Support is available.
1472    
1473    5.  In pcretest, an escape such as \x{de} in the data was always generating
1474        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1475        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1476        truncation.
1477    
1478    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1479    
1480    7.  Added two (int) casts to pcregrep when printing the difference of two
1481        pointers, in case they are 64-bit values.
1482    
1483    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1484        test 2 if it fails.
1485    
1486    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1487        and a #define of that name to empty if it is not externally set. This is to
1488        allow users of MSVC to set it if necessary.
1489    
1490    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1491        the convenience functions in the pcre_get.c source file.
1492    
1493    11. An option change at the start of a pattern that had top-level alternatives
1494        could cause overwriting and/or a crash. This command provoked a crash in
1495        some environments:
1496    
1497          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1498    
1499        This potential security problem was recorded as CVE-2008-2371.
1500    
1501    12. For a pattern where the match had to start at the beginning or immediately
1502        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1503        pcre_dfa_exec() could read past the end of the passed subject if there was
1504        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1505        pcretest so that it places the subject at the end of its malloc-ed buffer.
1506    
1507    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1508        exec() might read past the end of the data buffer in UTF-8 mode.
1509    
1510    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1511        the data contained the byte 0x85 as part of a UTF-8 character within its
1512        first line. This applied both to normal and DFA matching.
1513    
1514    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1515        /^[^d]*?$/8 failed to match "abc".
1516    
1517    16. Added a missing copyright notice to pcrecpp_internal.h.
1518    
1519    17. Make it more clear in the documentation that values returned from
1520        pcre_exec() in ovector are byte offsets, not character counts.
1521    
1522    18. Tidied a few places to stop certain compilers from issuing warnings.
1523    
1524    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1525        supplied by Stefan Weber. I made a further small update for 7.8 because
1526        there is a change of source arrangements: the pcre_searchfuncs.c module is
1527        replaced by pcre_ucd.c.
1528    
1529    
1530    Version 7.7 07-May-08
1531    ---------------------
1532    
1533    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1534        a string to a long long, pretend we don't even have a long long." This is
1535        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1536    
1537    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1538        pre-7.6 versions, which defined a global no_arg variable instead of putting
1539        it in the RE class. (See also #8 below.)
1540    
1541    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1542        Lopes.
1543    
1544    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1545    
1546        (1) The include/exclude patterns were being applied to the whole pathnames
1547            of files, instead of just to the final components.
1548    
1549        (2) If there was more than one level of directory, the subdirectories were
1550            skipped unless they satisfied the include/exclude conditions. This is
1551            inconsistent with GNU grep (and could even be seen as contrary to the
1552            pcregrep specification - which I improved to make it absolutely clear).
1553            The action now is always to scan all levels of directory, and just
1554            apply the include/exclude patterns to regular files.
1555    
1556    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1557        --exclude_dir in the tests to avoid scanning .svn directories.
1558    
1559    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1560        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1561        doesn't support NULs in patterns.
1562    
1563    7.  Added some missing "const"s to declarations of static tables in
1564        pcre_compile.c and pcre_dfa_exec.c.
1565    
1566    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1567        caused by fix #2  above. (Subsequently also a second patch to fix the
1568        first patch. And a third patch - this was a messy problem.)
1569    
1570    9.  Applied Craig's patch to remove the use of push_back().
1571    
1572    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1573        matching function regexec().
1574    
1575    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1576        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1577        references. PCRE supports relative numbers with this syntax (I don't think
1578        Oniguruma does).
1579    
1580    12. Previously, a group with a zero repeat such as (...){0} was completely
1581        omitted from the compiled regex. However, this means that if the group
1582        was called as a subroutine from elsewhere in the pattern, things went wrong
1583        (an internal error was given). Such groups are now left in the compiled
1584        pattern, with a new opcode that causes them to be skipped at execution
1585        time.
1586    
1587    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1588        to the way PCRE behaves:
1589    
1590        (a) A lone ] character is dis-allowed (Perl treats it as data).
1591    
1592        (b) A back reference to an unmatched subpattern matches an empty string
1593            (Perl fails the current match path).
1594    
1595        (c) A data ] in a character class must be notated as \] because if the
1596            first data character in a class is ], it defines an empty class. (In
1597            Perl it is not possible to have an empty class.) The empty class []
1598            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1599            The negative empty class [^] matches any one character, independently
1600            of the DOTALL setting.
1601    
1602    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1603        non-existent subpattern following a character class starting with ']' and
1604        containing () gave an internal compiling error instead of "reference to
1605        non-existent subpattern". Fortunately, when the pattern did exist, the
1606        compiled code was correct. (When scanning forwards to check for the
1607        existencd of the subpattern, it was treating the data ']' as terminating
1608        the class, so got the count wrong. When actually compiling, the reference
1609        was subsequently set up correctly.)
1610    
1611    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1612        it was being rejected as not supported by pcre_dfa_exec(), even though
1613        other assertions are supported. I have made pcre_dfa_exec() support
1614        (*FAIL).
1615    
1616    16. The implementation of 13c above involved the invention of a new opcode,
1617        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1618        cannot be changed at match time, I realized I could make a small
1619        improvement to matching performance by compiling OP_ALLANY instead of
1620        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1621        on the OP_ANY path.
1622    
1623    17. Compiling pcretest on Windows with readline support failed without the
1624        following two fixes: (1) Make the unistd.h include conditional on
1625        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1626    
1627    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1628        ncurses library to be included for pcretest when ReadLine support is
1629        requested, but also to allow for it to be overridden. This patch came from
1630        Daniel Bergström.
1631    
1632    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1633        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1634        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1635        spotting this.
1636    
1637    
1638    Version 7.6 28-Jan-08
1639    ---------------------
1640    
1641    1.  A character class containing a very large number of characters with
1642        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1643        overflow.
1644    
1645    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1646        HAVE_LONG_LONG is not defined.
1647    
1648    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1649        bring them up to date and include new features. This patch includes:
1650    
1651        - Fixed PH's badly added libz and libbz2 support.
1652        - Fixed a problem with static linking.
1653        - Added pcredemo. [But later removed - see 7 below.]
1654        - Fixed dftables problem and added an option.
1655        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1656            HAVE_LONG_LONG.
1657        - Added readline support for pcretest.
1658        - Added an listing of the option settings after cmake has run.
1659    
1660    4.  A user submitted a patch to Makefile that makes it easy to create
1661        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1662        Makefile.am that cause it to include this special target, without
1663        affecting anything else. Note that the same mingw target plus all
1664        the other distribution libraries and programs are now supported
1665        when configuring with CMake (see 6 below) instead of with
1666        Configure/Make.
1667    
1668    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1669        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1670        exported in the Windows port". It has not yet been confirmed that the patch
1671        solves the problem, but it does no harm.
1672    
1673    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1674        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1675        with CMake, and also correct the comment about stack recursion.
1676    
1677    7.  Remove the automatic building of pcredemo from the ./configure system and
1678        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1679        of a program that users should build themselves after PCRE is installed, so
1680        building it automatically is not really right. What is more, it gave
1681        trouble in some build environments.
1682    
1683    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1684    
1685    
1686    Version 7.5 10-Jan-08
1687    ---------------------
1688    
1689    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1690        values in parens when parsing an RE using the C++ wrapper."
1691    
1692    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1693        Characters greater than 255 were excluded from the class instead of being
1694        included.
1695    
1696    3.  The same bug as (2) above applied to negated POSIX classes such as
1697        [:^space:].
1698    
1699    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1700        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1701        I have changed it.
1702    
1703    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1704        first named subpattern) and a construct such as (?&a) would reference the
1705        first named subpattern whose name started with "a" (in other words, the
1706        length check was missing). Both these problems are fixed. "Subpattern name
1707        expected" is now given for (?&) (a zero-length name), and this patch also
1708        makes it give the same error for \k'' (previously it complained that that
1709        was a reference to a non-existent subpattern).
1710    
1711    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1712        this is right because (?- can be followed by option settings as well as by
1713        digits. I have, however, made the messages clearer.
1714    
1715    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1716        than the number used in the conditional) now cause a compile-time error.
1717        This is actually not compatible with Perl, which accepts such patterns, but
1718        treats the conditional as always being FALSE (as PCRE used to), but it
1719        seems to me that giving a diagnostic is better.
1720    
1721    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1722        and messages.
1723    
1724    9.  Fix two occurrences of "backslash" in comments that should have been
1725        "backspace".
1726    
1727    10. Remove two redundant lines of code that can never be obeyed (their function
1728        was moved elsewhere).
1729    
1730    11. The program that makes PCRE's Unicode character property table had a bug
1731        which caused it to generate incorrect table entries for sequences of
1732        characters that have the same character type, but are in different scripts.
1733        It amalgamated them into a single range, with the script of the first of
1734        them. In other words, some characters were in the wrong script. There were
1735        thirteen such cases, affecting characters in the following ranges:
1736    
1737          U+002b0 - U+002c1
1738          U+0060c - U+0060d
1739          U+0061e - U+00612
1740          U+0064b - U+0065e
1741          U+0074d - U+0076d
1742          U+01800 - U+01805
1743          U+01d00 - U+01d77
1744          U+01d9b - U+01dbf
1745          U+0200b - U+0200f
1746          U+030fc - U+030fe
1747          U+03260 - U+0327f
1748          U+0fb46 - U+0fbb1
1749          U+10450 - U+1049d
1750    
1751    12. The -o option (show only the matching part of a line) for pcregrep was not
1752        compatible with GNU grep in that, if there was more than one match in a
1753        line, it showed only the first of them. It now behaves in the same way as
1754        GNU grep.
1755    
1756    13. If the -o and -v options were combined for pcregrep, it printed a blank
1757        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1758        does the same. The return code can be used to tell if there were any
1759        non-matching lines.
1760    
1761    14. Added --file-offsets and --line-offsets to pcregrep.
1762    
1763    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1764        infinitely looping recursion. The bug was that positive lookaheads were not
1765        being skipped when checking for a possible empty match (negative lookaheads
1766        and both kinds of lookbehind were skipped).
1767    
1768    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1769        inclusion of <windows.h> to before rather than after the definition of
1770        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1771    
1772    17. Specifying a possessive quantifier with a specific limit for a Unicode
1773        character property caused pcre_compile() to compile bad code, which led at
1774        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1775        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1776        caused the error; without that there was no problem.
1777    
1778    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1779    
1780    19. Added --enable-pcretest-libreadline.
1781    
1782    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1783        RE::GlobalReplace(). As a result, the number of replacements returned was
1784        double what it should be. I removed one of the increments, but Craig sent a
1785        later patch that removed the other one (the right fix) and added unit tests
1786        that check the return values (which was not done before).
1787    
1788    21. Several CMake things:
1789    
1790        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1791            the names libpcre and libpcreposix, not just pcre and pcreposix.
1792    
1793        (2) The above change means that pcretest and pcregrep are now correctly
1794            linked with the newly-built libraries, not previously installed ones.
1795    
1796        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1797    
1798    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1799        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1800        UTF-8 newline character). The key issue is that the pattern starts .*;
1801        this means that the match must be either at the beginning, or after a
1802        newline. The bug was in the code for advancing after a failed match and
1803        checking that the new position followed a newline. It was not taking
1804        account of UTF-8 characters correctly.
1805    
1806    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1807        character classes. PCRE was not treating the sequence [:...:] as a
1808        character class unless the ... were all letters. Perl, however, seems to
1809        allow any characters between [: and :], though of course it rejects as
1810        unknown any "names" that contain non-letters, because all the known class
1811        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1812        for example, whereas PCRE did not - it did not recognize a POSIX character
1813        class. This seemed a bit dangerous, so the code has been changed to be
1814        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1815        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1816        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1817        Perl does, and where it didn't before.
1818    
1819    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1820        Windows environments %n is disabled by default.
1821    
1822    
1823    Version 7.4 21-Sep-07
1824    ---------------------
1825    
1826    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1827        means that a class such as [\s] counted as "explicit reference to CR or
1828        LF". That isn't really right - the whole point of the change was to try to
1829        help when there was an actual mention of one of the two characters. So now
1830        the change happens only if \r or \n (or a literal CR or LF) character is
1831        encountered.
1832    
1833    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1834        of both had grown to the point where there were only 3 bits left.
1835        Fortunately, there was spare space in the data structure, and so I have
1836        moved the internal flags into a new 16-bit field to free up more option
1837        bits.
1838    
1839    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1840        but did not set the internal JCHANGED flag - either of these is enough to
1841        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1842        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1843        start sets both bits.
1844    
1845    4.  Added options (at build time, compile time, exec time) to change \R from
1846        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1847    
1848    5.  doc/pcresyntax.html was missing from the distribution.
1849    
1850    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1851        compatibility, even though it is no longer used.
1852    
1853    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1854        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1855        windows.h file is present (where different names are used). [This was
1856        reversed later after testing - see 16 below.]
1857    
1858    8.  Changed all #include <config.h> to #include "config.h". There were also
1859        some further <pcre.h> cases that I changed to "pcre.h".
1860    
1861    9.  When pcregrep was used with the --colour option, it missed the line ending
1862        sequence off the lines that it output.
1863    
1864    10. It was pointed out to me that arrays of string pointers cause lots of
1865        relocations when a shared library is dynamically loaded. A technique of
1866        using a single long string with a table of offsets can drastically reduce
1867        these. I have refactored PCRE in four places to do this. The result is
1868        dramatic:
1869    
1870          Originally:                          290
1871          After changing UCP table:            187
1872          After changing error message table:   43
1873          After changing table of "verbs"       36
1874          After changing table of Posix names   22
1875    
1876        Thanks to the folks working on Gregex for glib for this insight.
1877    
1878    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1879        unicode-properties was also set.
1880    
1881    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1882    
1883    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1884        checked only for CRLF.
1885    
1886    14. Added casts to pcretest.c to avoid compiler warnings.
1887    
1888    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1889    
1890    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1891        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1892        entirely. This removes changes made in 7 above.
1893    
1894    17. The CMake files have been updated, and there is now more information about
1895        building with CMake in the NON-UNIX-USE document.
1896    
1897    
1898    Version 7.3 28-Aug-07
1899    ---------------------
1900    
1901     1. In the rejigging of the build system that eventually resulted in 7.1, the
1902        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1903        brackets there is not right, since it causes compilers to look for an
1904        installed pcre.h, not the version that is in the source that is being
1905        compiled (which of course may be different). I have changed it back to:
1906    
1907          #include "pcre.h"
1908    
1909        I have a vague recollection that the change was concerned with compiling in
1910        different directories, but in the new build system, that is taken care of
1911        by the VPATH setting the Makefile.
1912    
1913     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1914        when the subject happened to end in the byte 0x85 (e.g. if the last
1915        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1916        characters but of course it shouldn't be taken as a newline when it is part
1917        of another character. The bug was that, for an unlimited repeat of . in
1918        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1919        characters when looking for a newline.
1920    
1921     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1922    
1923     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1924        in debug output.
1925    
1926     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1927        long printing in the pcrecpp unittest when running under MinGW.
1928    
1929     6. ESC_K was left out of the EBCDIC table.
1930    
1931     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1932        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1933        limit also applies to "virtual nesting" when a pattern is recursive, and in
1934        this case 1000 isn't so big. I have been able to remove this limit at the
1935        expense of backing off one optimization in certain circumstances. Normally,
1936        when pcre_exec() would call its internal match() function recursively and
1937        immediately return the result unconditionally, it uses a "tail recursion"
1938        feature to save stack. However, when a subpattern that can match an empty
1939        string has an unlimited repetition quantifier, it no longer makes this
1940        optimization. That gives it a stack frame in which to save the data for
1941        checking that an empty string has been matched. Previously this was taken
1942        from the 1000-entry workspace that had been reserved. So now there is no
1943        explicit limit, but more stack is used.
1944    
1945     8. Applied Daniel's patches to solve problems with the import/export magic
1946        syntax that is required for Windows, and which was going wrong for the
1947        pcreposix and pcrecpp parts of the library. These were overlooked when this
1948        problem was solved for the main library.
1949    
1950     9. There were some crude static tests to avoid integer overflow when computing
1951        the size of patterns that contain repeated groups with explicit upper
1952        limits. As the maximum quantifier is 65535, the maximum group length was
1953        set at 30,000 so that the product of these two numbers did not overflow a
1954        32-bit integer. However, it turns out that people want to use groups that
1955        are longer than 30,000 bytes (though not repeat them that many times).
1956        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1957        made it possible to implement the integer overflow checks in a much more
1958        dynamic way, which I have now done. The artificial limitation on group
1959        length has been removed - we now have only the limit on the total length of
1960        the compiled pattern, which depends on the LINK_SIZE setting.
1961    
1962    10. Fixed a bug in the documentation for get/copy named substring when
1963        duplicate names are permitted. If none of the named substrings are set, the
1964        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1965        empty string.
1966    
1967    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1968        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1969        because the ] is interpreted as the first data character and the
1970        terminating ] is not found. PCRE has been made compatible with Perl in this
1971        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1972        cause memory overwriting.
1973    
1974    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1975        string has been matched (to stop an infinite loop). It was not recognizing
1976        a conditional subpattern that could match an empty string if that
1977        subpattern was within another subpattern. For example, it looped when
1978        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1979        condition was not nested. This bug has been fixed.
1980    
1981    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1982        past the start of the subject in the presence of bytes with the top bit
1983        set, for example "\x8aBCD".
1984    
1985    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1986        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1987    
1988    14. Optimized (?!) to (*FAIL).
1989    
1990    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1991        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1992        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1993        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1994        does: it's just the validity check that is more restrictive.
1995    
1996    16. Inserted checks for integer overflows during escape sequence (backslash)
1997        processing, and also fixed erroneous offset values for syntax errors during
1998        backslash processing.
1999    
2000    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2001        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2002    
2003    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2004        caused an overrun.
2005    
2006    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2007        something other than just ASCII characters) inside a group that had an
2008        unlimited repeat caused a loop at compile time (while checking to see
2009        whether the group could match an empty string).
2010    
2011    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2012        [\P{Any}] did so. (Error in the code for printing property names.)
2013    
2014    21. An orphan \E inside a character class could cause a crash.
2015    
2016    22. A repeated capturing bracket such as (A)? could cause a wild memory
2017        reference during compilation.
2018    
2019    23. There are several functions in pcre_compile() that scan along a compiled
2020        expression for various reasons (e.g. to see if it's fixed length for look
2021        behind). There were bugs in these functions when a repeated \p or \P was
2022        present in the pattern. These operators have additional parameters compared
2023        with \d, etc, and these were not being taken into account when moving along
2024        the compiled data. Specifically:
2025    
2026        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2027            length.
2028    
2029        (b) An item such as \pL+ within a repeated group could cause crashes or
2030            loops.
2031    
2032        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2033            "reference to non-existent subpattern" error.
2034    
2035        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2036    
2037    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2038        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2039    
2040    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2041    
2042    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2043        character were causing crashes (broken optimization).
2044    
2045    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2046        \p or \P) caused a compile-time loop.
2047    
2048    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2049        break. For example, the unstudied pattern [\r\n]A does not match the string
2050        "\r\nA" because change 7.0/46 below moves the current point on by two
2051        characters after failing to match at the start. However, the pattern \nA
2052        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2053        the same is true. There doesn't seem any very clean way out of this, but
2054        what I have chosen to do makes the common cases work: PCRE now takes note
2055        of whether there can be an explicit match for \r or \n anywhere in the
2056        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2057        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2058        pattern has explicit CR or LF references.
2059    
2060    29. Added (*CR) etc for changing newline setting at start of pattern.
2061    
2062    
2063    Version 7.2 19-Jun-07
2064    ---------------------
2065    
2066     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2067        which is apparently normally available under Windows.
2068    
2069     2. Re-jig the pcregrep tests with different newline settings in an attempt
2070        to make them independent of the local environment's newline setting.
2071    
2072     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2073    
2074     4. Some of the "internals" tests were previously cut out when the link size
2075        was not 2, because the output contained actual offsets. The recent new
2076        "Z" feature of pcretest means that these can be cut out, making the tests
2077        usable with all link sizes.
2078    
2079     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2080        stack recursion. This gives a massive performance boost under BSD, but just
2081        a small improvement under Linux. However, it saves one field in the frame
2082        in all cases.
2083    
2084     6. Added more features from the forthcoming Perl 5.10:
2085    
2086        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2087            recursion call. It refers to the nth most recently opened parentheses.
2088    
2089        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2090            to be opened parentheses.
2091    
2092        (c) Conditions that refer to capturing parentheses can be specified
2093            relatively, for example, (?(-2)... or (?(+3)...
2094    
2095        (d) \K resets the start of the current match so that everything before
2096            is not part of it.
2097    
2098        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2099    
2100        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2101            reference syntax.
2102    
2103        (g) (?| introduces a group in which the numbering of parentheses in each
2104            alternative starts with the same number.
2105    
2106        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2107    
2108     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2109        PCRE_INFO_JCHANGED.
2110    
2111     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2112        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2113        for detecting groups that can match an empty string.
2114    
2115     9. A pattern with a very large number of alternatives (more than several
2116        hundred) was running out of internal workspace during the pre-compile
2117        phase, where pcre_compile() figures out how much memory will be needed. A
2118        bit of new cunning has reduced the workspace needed for groups with
2119        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2120        workspace instead of running out of the 4096 that are available.
2121    
2122    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2123    
2124    11. Applied patch from Google to remove an optimization that didn't quite work.
2125        The report of the bug said:
2126    
2127          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2128          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2129          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2130    
2131    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2132        it matched the wrong number of bytes.
2133    
2134    
2135    Version 7.1 24-Apr-07
2136  ---------------------  ---------------------
2137    
2138   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one   1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2139      that is more "standard", making use of automake and other autotools.      that is more "standard", making use of automake and other Autotools. There
2140        is some re-arrangement of the files and adjustment of comments consequent
2141        on this.
2142    
2143     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2144        for recursive directory scanning broke on some systems because the files
2145        are not scanned in any specific order and on different systems the order
2146        was different. A call to "sort" has been inserted into RunGrepTest for the
2147        approprate test as a short-term fix. In the longer term there may be an
2148        alternative.
2149    
2150     3. I had an email from Eric Raymond about problems translating some of PCRE's
2151        man pages to HTML (despite the fact that I distribute HTML pages, some
2152        people do their own conversions for various reasons). The problems
2153        concerned the use of low-level troff macros .br and .in. I have therefore
2154        removed all such uses from the man pages (some were redundant, some could
2155        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2156        HTML has been updated to handle .nf/.fi and to complain if it encounters
2157        .br or .in.
2158    
2159     4. Updated comments in configure.ac that get placed in config.h.in and also
2160        arranged for config.h to be included in the distribution, with the name
2161        config.h.generic, for the benefit of those who have to compile without
2162        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2163    
2164     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2165        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2166        makevp.bat for the current PCRE, using the additional files
2167        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2168    
2169     6. A Windows user reported a minor discrepancy with test 2, which turned out
2170        to be caused by a trailing space on an input line that had got lost in his
2171        copy. The trailing space was an accident, so I've just removed it.
2172    
2173     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2174        that is needed.
2175    
2176     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2177        as "const" (a) because they are and (b) because it helps the PHP
2178        maintainers who have recently made a script to detect big data structures
2179        in the php code that should be moved to the .rodata section. I remembered
2180        to update Builducptable as well, so it won't revert if ucptable.h is ever
2181        re-created.
2182    
2183     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2184        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2185        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2186        support is not required. This saves 1.5-2K of code, which is important in
2187        some applications.
2188    
2189        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2190        so as not to refer to the tables, even though these functions will never be
2191        called when UTF-8 support is disabled. Otherwise there are problems with a
2192        shared library.
2193    
2194    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2195    
2196        (a) It was defining its arguments as char * instead of void *.
2197    
2198        (b) It was assuming that all moves were upwards in memory; this was true
2199            a long time ago when I wrote it, but is no longer the case.
2200    
2201        The emulated memove() is provided for those environments that have neither
2202        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2203        is clearly not the case, as these two bugs were recently reported.
2204    
2205    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2206        and Detrail to create the HTML documentation, the .txt form of the man
2207        pages, and it removes trailing spaces from listed files. It also creates
2208        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2209        case, it wraps all the #defines with #ifndefs. This script should be run
2210        before "make dist".
2211    
2212    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2213        with Unicode property support.
2214    
2215        (a) For a maximizing quantifier, if the two different cases of the
2216            character were of different lengths in their UTF-8 codings (there are
2217            some cases like this - I found 11), and the matching function had to
2218            back up over a mixture of the two cases, it incorrectly assumed they
2219            were both the same length.
2220    
2221        (b) When PCRE was configured to use the heap rather than the stack for
2222            recursion during matching, it was not correctly preserving the data for
2223            the other case of a UTF-8 character when checking ahead for a match
2224            while processing a minimizing repeat. If the check also involved
2225            matching a wide character, but failed, corruption could cause an
2226            erroneous result when trying to check for a repeat of the original
2227            character.
2228    
2229    13. Some tidying changes to the testing mechanism:
2230    
2231        (a) The RunTest script now detects the internal link size and whether there
2232            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2233            values substituted by "configure". (The RunGrepTest script already did
2234            this for UTF-8.) The configure.ac script no longer substitutes the
2235            relevant variables.
2236    
2237        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2238            with length and offset values. This means that the output is different
2239            for different internal link sizes. Test 2 is skipped for link sizes
2240            other than 2 because of this, bypassing the problem. Unfortunately,
2241            there was also a test in test 3 (the locale tests) that used /B and
2242            failed for link sizes other than 2. Rather than cut the whole test out,
2243            I have added a new /Z option to pcretest that replaces the length and
2244            offset values with spaces. This is now used to make test 3 independent
2245            of link size. (Test 2 will be tidied up later.)
2246    
2247    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2248        segmentation fault instead of returning the appropriate error message.
2249    
2250    15. In multiline mode when the newline sequence was set to "any", the pattern
2251        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2252        This doesn't seem right; it now treats the CRLF combination as the line
2253        ending, and so does not match in that case. It's only a pattern such as ^$
2254        that would hit this one: something like ^ABC$ would have failed after \r
2255        and then tried again after \r\n.
2256    
2257    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2258        in an attempt to make files that differ only in their line terminators
2259        compare equal. This works on Linux.
2260    
2261    17. Under certain error circumstances pcregrep might try to free random memory
2262        as it exited. This is now fixed, thanks to valgrind.
2263    
2264    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2265        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2266        was because its rules for how to advance for /g after matching an empty
2267        string at the end of a line did not allow for this case. They now check for
2268        it specially.
2269    
2270    20. pcretest is supposed to handle patterns and data of any length, by
2271        extending its buffers when necessary. It was getting this wrong when the
2272        buffer for a data line had to be extended.
2273    
2274    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2275        CRLF as a newline sequence.
2276    
2277    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2278        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2279        I have nevertheless tidied it up.
2280    
2281    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2282    
2283    24. Added a man page for pcre-config.
2284    
2285    
2286  Version 7.0 19-Dec-06  Version 7.0 19-Dec-06

Legend:
Removed from v.98  
changed lines
  Added in v.991

  ViewVC Help
Powered by ViewVC 1.1.5