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

Diff of /code/trunk/ChangeLog

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

revision 3 by nigel, Sat Feb 24 21:38:01 2007 UTC revision 967 by zherczeg, Mon May 14 07:36:39 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    9.  (*MARK) control verb is now supported by the JIT compiler.
46    
47    10. The command "./RunTest list" lists the available tests without actually
48        running any of them. (Because I keep forgetting what they all are.)
49    
50    11. Add PCRE_INFO_MAXLOOKBEHIND.
51    
52    12. Applied a (slightly modified) user-supplied patch that improves performance
53        when the heap is used for recursion (compiled with --disable-stack-for-
54        recursion). Instead of malloc and free for each heap frame each time a
55        logical recursion happens, frames are retained on a chain and re-used where
56        possible. This sometimes gives as much as 30% improvement.
57    
58    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59        call.
60    
61    14. As documented, (*COMMIT) is now confined to within a positive assertion.
62    
63    15. It is now possible to link pcretest with libedit as an alternative to
64        libreadline.
65    
66    16. (*COMMIT) control verb is now supported by the JIT compiler.
67    
68    17. The Unicode data tables have been updated to Unicode 6.1.0.
69    
70    18. Added --file-list option to pcregrep.
71    
72    19. Added binary file support to pcregrep, including the -a, --binary-files,
73        -I, and --text options.
74    
75    20. The madvise function is renamed for posix_madvise for QNX compatibility
76        reasons. Fixed by Giuseppe D'Angelo.
77    
78    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79    
80    22. Changed the option for creating symbolic links for 16-bit man pages from
81        -s to -sf so that re-installing does not cause issues.
82    
83    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84    
85    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86        matches in certain environments (the workspace was not being correctly
87        retained). Also added to pcre_dfa_exec() a simple plausibility check on
88        some of the workspace data at the beginning of a restart.
89    
90    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91        was not doing so when it should - probably a typo introduced by SVN 528
92        (change 8.10/14).
93    
94    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
95        \w+ when the character tables indicated that \x{c4} was a word character.
96        There were several related cases, all because the tests for doing a table
97        lookup were testing for characters less than 127 instead of 255.
98    
99    27. If a pattern contains capturing parentheses that are not used in a match,
100        their slots in the ovector are set to -1. For those that are higher than
101        any matched groups, this happens at the end of processing. In the case when
102        there were back references that the ovector was too small to contain
103        (causing temporary malloc'd memory to be used during matching), and the
104        highest capturing number was not used, memory off the end of the ovector
105        was incorrectly being set to -1. (It was using the size of the temporary
106        memory instead of the true size.)
107    
108    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
109        ovector size, it uses memory at the end of the block that it has got.
110    
111    29. Check for an overlong MARK name and give an error at compile time. The
112        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
113    
114    30. JIT compiler update.
115    
116    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
117        Rill for the patch.
118    
119    
120    Version 8.30 04-February-2012
121    -----------------------------
122    
123    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
124        name is defined in ctype.h.
125    
126    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
127        only in quite long subpatterns.
128    
129    3.  Removed the function pcre_info(), which has been obsolete and deprecated
130        since it was replaced by pcre_fullinfo() in February 2000.
131    
132    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
133        match a (*MARK), and the match failed at the start of the subject, a
134        reference to memory before the start of the subject could occur. This bug
135        was introduced by fix 17 of release 8.21.
136    
137    5.  A reference to an unset group with zero minimum repetition was giving
138        totally wrong answers (in non-JavaScript-compatibility mode). For example,
139        /(another)?(\1?)test/ matched against "hello world test". This bug was
140        introduced in release 8.13.
141    
142    6.  Add support for 16-bit character strings (a large amount of work involving
143        many changes and refactorings).
144    
145    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
146        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
147        from a file.
148    
149    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
150        rounding is not applied in this particular case).
151    
152    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
153        if they appear, or are escaped, in patterns.
154    
155    10. Get rid of a number of -Wunused-but-set-variable warnings.
156    
157    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
158        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
159        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
160        also returns the mark "x". This bug applied to capturing parentheses,
161        non-capturing parentheses, and atomic parentheses. It also applied to some
162        assertions.
163    
164    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
165        information out of configure.ac instead of relying on pcre.h.generic, which
166        is not stored in the repository.
167    
168    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
169        -lreadline.
170    
171    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
172    
173    15. Applied Graycode's patch to put the top-level frame on the stack rather
174        than the heap when not using the stack for recursion. This gives a
175        performance improvement in many cases when recursion is not deep.
176    
177    16. Experimental code added to "pcretest -C" to output the stack frame size.
178    
179    
180    Version 8.21 12-Dec-2011
181    ------------------------
182    
183    1.  Updating the JIT compiler.
184    
185    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
186        are added as well.
187    
188    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
189        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
190        calling _pcre_jit_exec. Some extra comments are added.
191    
192    4.  (*MARK) settings inside atomic groups that do not contain any capturing
193        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
194        was introduced by change 18 for 8.20.
195    
196    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
197        ECMA-262 standard.
198    
199    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
200        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
201        This bug was probably introduced by change 9 of 8.13.
202    
203    7.  While fixing 6 above, I noticed that a number of other items were being
204        incorrectly rejected as "not fixed length". This arose partly because newer
205        opcodes had not been added to the fixed-length checking code. I have (a)
206        corrected the bug and added tests for these items, and (b) arranged for an
207        error to occur if an unknown opcode is encountered while checking for fixed
208        length instead of just assuming "not fixed length". The items that were
209        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
210        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
211        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
212    
213    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
214        being incorrectly compiled and would have given unpredicatble results.
215    
216    9.  A possessively repeated subpattern with minimum repeat count greater than
217        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
218        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
219        the first (A) could occur when it should not.
220    
221    10. Add a cast and remove a redundant test from the code.
222    
223    11. JIT should use pcre_malloc/pcre_free for allocation.
224    
225    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
226        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
227        is anything other than /usr, -L is still shown).
228    
229    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
230    
231    14. Perl does not support \N without a following name in a [] class; PCRE now
232        also gives an error.
233    
234    15. If a forward reference was repeated with an upper limit of around 2000,
235        it caused the error "internal error: overran compiling workspace". The
236        maximum number of forward references (including repeats) was limited by the
237        internal workspace, and dependent on the LINK_SIZE. The code has been
238        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
239        the default depends on LINK_SIZE. There is a new upper limit (for safety)
240        of around 200,000 forward references. While doing this, I also speeded up
241        the filling in of repeated forward references.
242    
243    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
244        incorrectly expecting the subject to contain another "a" after the start.
245    
246    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
247        in the match, the SKIP should be ignored. This was not happening; instead
248        the SKIP was being treated as NOMATCH. For patterns such as
249        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
250        tested.
251    
252    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
253        now much more compatible with Perl, in particular in cases where the result
254        is a non-match for a non-anchored pattern. For example, if
255        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
256        "m", where previously it did not return a name. A side effect of this
257        change is that for partial matches, the last encountered mark name is
258        returned, as for non matches. A number of tests that were previously not
259        Perl-compatible have been moved into the Perl-compatible test files. The
260        refactoring has had the pleasing side effect of removing one argument from
261        the match() function, thus reducing its stack requirements.
262    
263    19. If the /S+ option was used in pcretest to study a pattern using JIT,
264        subsequent uses of /S (without +) incorrectly behaved like /S+.
265    
266    21. Retrieve executable code size support for the JIT compiler and fixing
267        some warnings.
268    
269    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
270        not work when the shorter character appeared right at the end of the
271        subject string.
272    
273    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
274        systems.
275    
276    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
277        output it when the /M option is used in pcretest.
278    
279    25. The CheckMan script was not being included in the distribution. Also, added
280        an explicit "perl" to run Perl scripts from the PrepareRelease script
281        because this is reportedly needed in Windows.
282    
283    26. If study data was being save in a file and studying had not found a set of
284        "starts with" bytes for the pattern, the data written to the file (though
285        never used) was taken from uninitialized memory and so caused valgrind to
286        complain.
287    
288    27. Updated RunTest.bat as provided by Sheri Pierce.
289    
290    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
291    
292    29. Computation of memory usage for the table of capturing group names was
293        giving an unnecessarily large value.
294    
295    
296    Version 8.20 21-Oct-2011
297    ------------------------
298    
299    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
300        a POSIX class. After further experiments with Perl, which convinced me that
301        Perl has bugs and confusions, a closing square bracket is no longer allowed
302        in a POSIX name. This bug also affected patterns with classes that started
303        with full stops.
304    
305    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
306        captured substring, but while checking the failing first alternative,
307        substring 1 is temporarily captured. If the output vector supplied to
308        pcre_exec() was not big enough for this capture, the yield of the function
309        was still zero ("insufficient space for captured substrings"). This cannot
310        be totally fixed without adding another stack variable, which seems a lot
311        of expense for a edge case. However, I have improved the situation in cases
312        such as /(a)(b)x|abc/ matched against "abc", where the return code
313        indicates that fewer than the maximum number of slots in the ovector have
314        been set.
315    
316    3.  Related to (2) above: when there are more back references in a pattern than
317        slots in the output vector, pcre_exec() uses temporary memory during
318        matching, and copies in the captures as far as possible afterwards. It was
319        using the entire output vector, but this conflicts with the specification
320        that only 2/3 is used for passing back captured substrings. Now it uses
321        only the first 2/3, for compatibility. This is, of course, another edge
322        case.
323    
324    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
325        main code base, and can be used by building with --enable-jit. When this is
326        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
327        runtime --no-jit option is given.
328    
329    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
330        ovector, the return from the function was zero, implying that there were
331        other matches that did not fit. The correct "exactly full" value is now
332        returned.
333    
334    6.  If a subpattern that was called recursively or as a subroutine contained
335        (*PRUNE) or any other control that caused it to give a non-standard return,
336        invalid errors such as "Error -26 (nested recursion at the same subject
337        position)" or even infinite loops could occur.
338    
339    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
340        computing the minimum length on reaching *ACCEPT, and so ended up with the
341        wrong value of 1 rather than 0. Further investigation indicates that
342        computing a minimum subject length in the presence of *ACCEPT is difficult
343        (think back references, subroutine calls), and so I have changed the code
344        so that no minimum is registered for a pattern that contains *ACCEPT.
345    
346    8.  If (*THEN) was present in the first (true) branch of a conditional group,
347        it was not handled as intended. [But see 16 below.]
348    
349    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
350        Sheri Pierce.
351    
352    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
353        the first byte in a match must be "a".
354    
355    11. Change 17 for 8.13 increased the recursion depth for patterns like
356        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
357        pattern contains any instances of (*THEN). If it does not, the old
358        optimizations are restored. It would be nice to do this on a per-group
359        basis, but at the moment that is not feasible.
360    
361    12. In some environments, the output of pcretest -C is CRLF terminated. This
362        broke RunTest's code that checks for the link size. A single white space
363        character after the value is now allowed for.
364    
365    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
366        For "fr", it uses the Windows-specific input and output files.
367    
368    14. If (*THEN) appeared in a group that was called recursively or as a
369        subroutine, it did not work as intended. [But see next item.]
370    
371    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
372        pattern fragments (but not containing any | characters). If A and B are
373        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
374        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
375        D. In other words, Perl considers parentheses that do not contain any |
376        characters to be part of a surrounding alternative, whereas PCRE was
377        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
378        differently. PCRE now behaves in the same way as Perl, except in the case
379        of subroutine/recursion calls such as (?1) which have in any case always
380        been different (but PCRE had them first :-).
381    
382    16. Related to 15 above: Perl does not treat the | in a conditional group as
383        creating alternatives. Such a group is treated in the same way as an
384        ordinary group without any | characters when processing (*THEN). PCRE has
385        been changed to match Perl's behaviour.
386    
387    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
388        RunGrepTest script failed.
389    
390    18. Change 22 for version 13 caused atomic groups to use more stack. This is
391        inevitable for groups that contain captures, but it can lead to a lot of
392        stack use in large patterns. The old behaviour has been restored for atomic
393        groups that do not contain any capturing parentheses.
394    
395    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
396        suppress the check for a minimum subject length at run time. (If it was
397        given to pcre_exec() or pcre_dfa_exec() it did work.)
398    
399    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
400        fail to work when decoding hex characters in data strings in EBCDIC
401        environments.
402    
403    21. It appears that in at least one Mac OS environment, the isxdigit() function
404        is implemented as a macro that evaluates to its argument more than once,
405        contravening the C 90 Standard (I haven't checked a later standard). There
406        was an instance in pcretest which caused it to go wrong when processing
407        \x{...} escapes in subject strings. The has been rewritten to avoid using
408        things like p++ in the argument of isxdigit().
409    
410    
411    Version 8.13 16-Aug-2011
412    ------------------------
413    
414    1.  The Unicode data tables have been updated to Unicode 6.0.0.
415    
416    2.  Two minor typos in pcre_internal.h have been fixed.
417    
418    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
419        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
420        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
421    
422    4.  There were a number of related bugs in the code for matching backrefences
423        caselessly in UTF-8 mode when codes for the characters concerned were
424        different numbers of bytes. For example, U+023A and U+2C65 are an upper
425        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
426        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
427        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
428        2-byte code at the end of the subject (it thought there wasn't enough data
429        left).
430    
431    5.  Comprehensive information about what went wrong is now returned by
432        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
433        as the output vector has at least 2 elements. The offset of the start of
434        the failing character and a reason code are placed in the vector.
435    
436    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
437        now returned is for the first byte of the failing character, instead of the
438        last byte inspected. This is an incompatible change, but I hope it is small
439        enough not to be a problem. It makes the returned offset consistent with
440        pcre_exec() and pcre_dfa_exec().
441    
442    7.  pcretest now gives a text phrase as well as the error number when
443        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
444        failure, the offset and reason code are output.
445    
446    8.  When \R was used with a maximizing quantifier it failed to skip backwards
447        over a \r\n pair if the subsequent match failed. Instead, it just skipped
448        back over a single character (\n). This seems wrong (because it treated the
449        two characters as a single entity when going forwards), conflicts with the
450        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
451        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
452        has been changed.
453    
454    9.  Some internal refactoring has changed the processing so that the handling
455        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
456        time (the PCRE_DOTALL option was changed this way some time ago: version
457        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
458        which was always a bit of a fudge. It also means that there is one less
459        argument for the match() function, which reduces its stack requirements
460        slightly. This change also fixes an incompatibility with Perl: the pattern
461        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
462    
463    10. More internal refactoring has drastically reduced the number of recursive
464        calls to match() for possessively repeated groups such as (abc)++ when
465        using pcre_exec().
466    
467    11. While implementing 10, a number of bugs in the handling of groups were
468        discovered and fixed:
469    
470        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
471        (a|)*(?1) gave a compile-time internal error.
472        ((a|)+)+  did not notice that the outer group could match an empty string.
473        (^a|^)+   was not marked as anchored.
474        (.*a|.*)+ was not marked as matching at start or after a newline.
475    
476    12. Yet more internal refactoring has removed another argument from the match()
477        function. Special calls to this function are now indicated by setting a
478        value in a variable in the "match data" data block.
479    
480    13. Be more explicit in pcre_study() instead of relying on "default" for
481        opcodes that mean there is no starting character; this means that when new
482        ones are added and accidentally left out of pcre_study(), testing should
483        pick them up.
484    
485    14. The -s option of pcretest has been documented for ages as being an old
486        synonym of -m (show memory usage). I have changed it to mean "force study
487        for every regex", that is, assume /S for every regex. This is similar to -i
488        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
489        using it. It makes it easier to run collections of tests with and without
490        study enabled, and thereby test pcre_study() more easily. All the standard
491        tests are now run with and without -s (but some patterns can be marked as
492        "never study" - see 20 below).
493    
494    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
495        restoration of the capturing data to the outer values was not happening
496        correctly.
497    
498    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
499        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
500        pattern had matched an empty string, and so incorrectly returned a no
501        match.
502    
503    17. There was optimizing code for the last branch of non-capturing parentheses,
504        and also for the obeyed branch of a conditional subexpression, which used
505        tail recursion to cut down on stack usage. Unfortunately, now that there is
506        the possibility of (*THEN) occurring in these branches, tail recursion is
507        no longer possible because the return has to be checked for (*THEN). These
508        two optimizations have therefore been removed. [But see 8.20/11 above.]
509    
510    18. If a pattern containing \R was studied, it was assumed that \R always
511        matched two bytes, thus causing the minimum subject length to be
512        incorrectly computed because \R can also match just one byte.
513    
514    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
515        was incorrectly computed.
516    
517    20. If /S is present twice on a test pattern in pcretest input, it now
518        *disables* studying, thereby overriding the use of -s on the command line
519        (see 14 above). This is necessary for one or two tests to keep the output
520        identical in both cases.
521    
522    21. When (*ACCEPT) was used in an assertion that matched an empty string and
523        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
524    
525    22. When an atomic group that contained a capturing parenthesis was
526        successfully matched, but the branch in which it appeared failed, the
527        capturing was not being forgotten if a higher numbered group was later
528        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
529        group 1 to "a", when in fact it should be unset. This applied to multi-
530        branched capturing and non-capturing groups, repeated or not, and also to
531        positive assertions (capturing in negative assertions does not happen
532        in PCRE) and also to nested atomic groups.
533    
534    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
535        subject after a captured substring, to make it easier to tell which of a
536        number of identical substrings has been captured.
537    
538    24. The way atomic groups are processed by pcre_exec() has been changed so that
539        if they are repeated, backtracking one repetition now resets captured
540        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
541        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
542        "aaa". Previously, it would have been "a". As part of this code
543        refactoring, the way recursive calls are handled has also been changed.
544    
545    25. If an assertion condition captured any substrings, they were not passed
546        back unless some other capturing happened later. For example, if
547        (?(?=(a))a) was matched against "a", no capturing was returned.
548    
549    26. When studying a pattern that contained subroutine calls or assertions,
550        the code for finding the minimum length of a possible match was handling
551        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
552        group 1 called group 2 while simultaneously a separate group 2 called group
553        1). A stack overflow occurred in this case. I have fixed this by limiting
554        the recursion depth to 10.
555    
556    27. Updated RunTest.bat in the distribution to the version supplied by Tom
557        Fortmann. This supports explicit test numbers on the command line, and has
558        argument validation and error reporting.
559    
560    28. An instance of \X with an unlimited repeat could fail if at any point the
561        first character it looked at was a mark character.
562    
563    29. Some minor code refactoring concerning Unicode properties and scripts
564        should reduce the stack requirement of match() slightly.
565    
566    30. Added the '=' option to pcretest to check the setting of unused capturing
567        slots at the end of the pattern, which are documented as being -1, but are
568        not included in the return count.
569    
570    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
571        compiled something random. Now it gives a compile-time error (as does
572        Perl).
573    
574    32. A *MARK encountered during the processing of a positive assertion is now
575        recorded and passed back (compatible with Perl).
576    
577    33. If --only-matching or --colour was set on a pcregrep call whose pattern
578        had alternative anchored branches, the search for a second match in a line
579        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
580        matched the line "0102" twice. The same bug affected patterns that started
581        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
582        twice.
583    
584    34. Previously, PCRE did not allow quantification of assertions. However, Perl
585        does, and because of capturing effects, quantifying parenthesized
586        assertions may at times be useful. Quantifiers are now allowed for
587        parenthesized assertions.
588    
589    35. A minor code tidy in pcre_compile() when checking options for \R usage.
590    
591    36. \g was being checked for fancy things in a character class, when it should
592        just be a literal "g".
593    
594    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
595        appearance of a nested POSIX class supersedes an apparent external class.
596        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
597        unescaped square brackets may also appear as part of class names. For
598        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
599        more like Perl. (But see 8.20/1 above.)
600    
601    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
602        was because it thought it was \N{name}, which is not supported).
603    
604    39. Add minix to OS list not supporting the -S option in pcretest.
605    
606    40. PCRE tries to detect cases of infinite recursion at compile time, but it
607        cannot analyze patterns in sufficient detail to catch mutual recursions
608        such as ((?1))((?2)). There is now a runtime test that gives an error if a
609        subgroup is called recursively as a subpattern for a second time at the
610        same position in the subject string. In previous releases this might have
611        been caught by the recursion limit, or it might have run out of stack.
612    
613    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
614        happen only once. PCRE was, however incorrectly giving a compile time error
615        "recursive call could loop indefinitely" because it cannot analyze the
616        pattern in sufficient detail. The compile time test no longer happens when
617        PCRE is compiling a conditional subpattern, but actual runaway loops are
618        now caught at runtime (see 40 above).
619    
620    42. It seems that Perl allows any characters other than a closing parenthesis
621        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
622        has been changed to be the same.
623    
624    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
625        as not to get warnings when autogen.sh is called. Also changed
626        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
627    
628    44. To help people who use pcregrep to scan files containing exceedingly long
629        lines, the following changes have been made:
630    
631        (a) The default value of the buffer size parameter has been increased from
632            8K to 20K. (The actual buffer used is three times this size.)
633    
634        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
635            PCRE is built.
636    
637        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
638            to be set at run time.
639    
640        (d) Numerical values in pcregrep options can be followed by K or M, for
641            example --buffer-size=50K.
642    
643        (e) If a line being scanned overflows pcregrep's buffer, an error is now
644            given and the return code is set to 2.
645    
646    45. Add a pointer to the latest mark to the callout data block.
647    
648    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
649        partial match of an empty string instead of no match. This was specific to
650        the use of ".".
651    
652    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
653        complete match instead of a partial match. This bug was dependent on both
654        the PCRE_UTF8 and PCRE_DOTALL options being set.
655    
656    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
657        starting byte set, because \b was not being ignored.
658    
659    
660    Version 8.12 15-Jan-2011
661    ------------------------
662    
663    1.  Fixed some typos in the markup of the man pages, and wrote a script that
664        checks for such things as part of the documentation building process.
665    
666    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
667        --match-limit and --recursion-limit options (added for 8.11). In
668        particular, this made one of the standard tests fail. (The integer value
669        went into the wrong half of a long int.)
670    
671    3.  If the --colour option was given to pcregrep with -v (invert match), it
672        did strange things, either producing crazy output, or crashing. It should,
673        of course, ignore a request for colour when reporting lines that do not
674        match.
675    
676    4.  Another pcregrep bug caused similar problems if --colour was specified with
677        -M (multiline) and the pattern match finished with a line ending.
678    
679    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
680        matched in multiline mode, the following line was shown as part of the
681        match. This seems wrong, so I have changed it.
682    
683    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
684        the check for further matches in the same line (so they could be coloured)
685        to overrun the end of the current line. If another match was found, it was
686        incorrectly shown (and then shown again when found in the next line).
687    
688    7.  If pcregrep was compiled under Windows, there was a reference to the
689        function pcregrep_exit() before it was defined. I am assuming this was
690        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
691        reported by a user. I've moved the definition above the reference.
692    
693    
694    Version 8.11 10-Dec-2010
695    ------------------------
696    
697    1.  (*THEN) was not working properly if there were untried alternatives prior
698        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
699        backtracked to try for "b" instead of moving to the next alternative branch
700        at the same level (in this case, to look for "c"). The Perl documentation
701        is clear that when (*THEN) is backtracked onto, it goes to the "next
702        alternative in the innermost enclosing group".
703    
704    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
705        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
706        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
707        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
708        (*THEN).
709    
710    3.  If \s appeared in a character class, it removed the VT character from
711        the class, even if it had been included by some previous item, for example
712        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
713        of \s, but is part of the POSIX "space" class.)
714    
715    4.  A partial match never returns an empty string (because you can always
716        match an empty string at the end of the subject); however the checking for
717        an empty string was starting at the "start of match" point. This has been
718        changed to the "earliest inspected character" point, because the returned
719        data for a partial match starts at this character. This means that, for
720        example, /(?<=abc)def/ gives a partial match for the subject "abc"
721        (previously it gave "no match").
722    
723    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
724        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
725        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
726        has an implication that the given string is incomplete (because a partial
727        match is preferred over a full match). For this reason, these items now
728        give a partial match in this situation. [Aside: previously, the one case
729        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
730        match rather than a full match, which was wrong by the old rules, but is
731        now correct.]
732    
733    6.  There was a bug in the handling of #-introduced comments, recognized when
734        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
735        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
736        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
737        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
738        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
739        places in pcre_compile().
740    
741    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
742        comments when looking ahead for named forward references to subpatterns,
743        the only newline sequence it recognized was NL. It now handles newlines
744        according to the set newline convention.
745    
746    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
747        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
748        cater for a lack of strerror(). These oversights have been fixed.
749    
750    9.  Added --match-limit and --recursion-limit to pcregrep.
751    
752    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
753    
754    11. When the -o option was used, pcregrep was setting a return code of 1, even
755        when matches were found, and --line-buffered was not being honoured.
756    
757    12. Added an optional parentheses number to the -o and --only-matching options
758        of pcregrep.
759    
760    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
761        can match an empty string. The code to do it in pcretest and pcredemo
762        needed fixing:
763    
764        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
765            only one byte after an empty string match just before CRLF (this case
766            just got forgotten; "any" and "anycrlf" were OK).
767    
768        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
769            mode when an empty string match preceded an ASCII character followed by
770            a non-ASCII character. (The code for advancing by one character rather
771            than one byte was nonsense.)
772    
773        (c) The pcredemo.c sample program did not have any code at all to handle
774            the cases when CRLF is a valid newline sequence.
775    
776    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
777        as a starting offset was within the subject string. There is now a new
778        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
779        negative or greater than the length of the string. In order to test this,
780        pcretest is extended to allow the setting of negative starting offsets.
781    
782    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
783        starting offset points to the beginning of a UTF-8 character was
784        unnecessarily clumsy. I tidied it up.
785    
786    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
787        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
788    
789    17. Nobody had reported that the --include_dir option, which was added in
790        release 7.7 should have been called --include-dir (hyphen, not underscore)
791        for compatibility with GNU grep. I have changed it to --include-dir, but
792        left --include_dir as an undocumented synonym, and the same for
793        --exclude-dir, though that is not available in GNU grep, at least as of
794        release 2.5.4.
795    
796    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
797        characters from a string of bytes) have been redefined so as not to use
798        loops, in order to improve performance in some environments. At the same
799        time, I abstracted some of the common code into auxiliary macros to save
800        repetition (this should not affect the compiled code).
801    
802    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
803        compile-time error is now given if \c is not followed by an ASCII
804        character, that is, a byte less than 128. (In EBCDIC mode, the code is
805        different, and any byte value is allowed.)
806    
807    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
808        START_OPTIMIZE option, which is now allowed at compile time - but just
809        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
810        to pcregrep and other applications that have no direct access to PCRE
811        options. The new /Y option in pcretest sets this option when calling
812        pcre_compile().
813    
814    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
815        back references. Groups containing recursive back references were forced to
816        be atomic by that change, but in the case of named groups, the amount of
817        memory required was incorrectly computed, leading to "Failed: internal
818        error: code overflow". This has been fixed.
819    
820    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
821        pcretest.c, to avoid build problems in some Borland environments.
822    
823    
824    Version 8.10 25-Jun-2010
825    ------------------------
826    
827    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
828        THEN.
829    
830    2.  (*ACCEPT) was not working when inside an atomic group.
831    
832    3.  Inside a character class, \B is treated as a literal by default, but
833        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
834        causes the error). The code is unchanged, but I tidied the documentation.
835    
836    4.  Inside a character class, PCRE always treated \R and \X as literals,
837        whereas Perl faults them if its -w option is set. I have changed PCRE so
838        that it faults them when PCRE_EXTRA is set.
839    
840    5.  Added support for \N, which always matches any character other than
841        newline. (It is the same as "." when PCRE_DOTALL is not set.)
842    
843    6.  When compiling pcregrep with newer versions of gcc which may have
844        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
845        declared with attribute warn_unused_result" were given. Just casting the
846        result to (void) does not stop the warnings; a more elaborate fudge is
847        needed. I've used a macro to implement this.
848    
849    7.  Minor change to pcretest.c to avoid a compiler warning.
850    
851    8.  Added four artifical Unicode properties to help with an option to make
852        \s etc use properties (see next item). The new properties are: Xan
853        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
854    
855    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
856        use Unicode properties. (*UCP) at the start of a pattern can be used to set
857        this option. Modified pcretest to add /W to test this facility. Added
858        REG_UCP to make it available via the POSIX interface.
859    
860    10. Added --line-buffered to pcregrep.
861    
862    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
863        studied, and the match started with a letter with a code point greater than
864        127 whose first byte was different to the first byte of the other case of
865        the letter, the other case of this starting letter was not recognized
866        (#976).
867    
868    12. If a pattern that was studied started with a repeated Unicode property
869        test, for example, \p{Nd}+, there was the theoretical possibility of
870        setting up an incorrect bitmap of starting bytes, but fortunately it could
871        not have actually happened in practice until change 8 above was made (it
872        added property types that matched character-matching opcodes).
873    
874    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
875        possible starting bytes for non-anchored patterns.
876    
877    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
878        \R, and also a number of cases that involve Unicode properties, both
879        explicit and implicit when PCRE_UCP is set.
880    
881    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
882        input, it could crash or give wrong results if characters with values
883        greater than 0xc0 were present in the subject string. (Detail: it assumed
884        UTF-8 input when processing these items.)
885    
886    16. Added a lot of (int) casts to avoid compiler warnings in systems where
887        size_t is 64-bit (#991).
888    
889    17. Added a check for running out of memory when PCRE is compiled with
890        --disable-stack-for-recursion (#990).
891    
892    18. If the last data line in a file for pcretest does not have a newline on
893        the end, a newline was missing in the output.
894    
895    19. The default pcre_chartables.c file recognizes only ASCII characters (values
896        less than 128) in its various bitmaps. However, there is a facility for
897        generating tables according to the current locale when PCRE is compiled. It
898        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
899        characters, are recognized by isspace() and therefore were getting set in
900        these tables, and indeed these tables seem to approximate to ISO 8859. This
901        caused a problem in UTF-8 mode when pcre_study() was used to create a list
902        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
903        which of course cannot start UTF-8 characters. I have changed the code so
904        that only real ASCII characters (less than 128) and the correct starting
905        bytes for UTF-8 encodings are set for characters greater than 127 when in
906        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
907        altogether.)
908    
909    20. Added the /T option to pcretest so as to be able to run tests with non-
910        standard character tables, thus making it possible to include the tests
911        used for 19 above in the standard set of tests.
912    
913    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
914        reference to a subpattern the other side of a comment that contains an
915        opening parenthesis caused either an internal compiling error, or a
916        reference to the wrong subpattern.
917    
918    
919    Version 8.02 19-Mar-2010
920    ------------------------
921    
922    1.  The Unicode data tables have been updated to Unicode 5.2.0.
923    
924    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
925        configured.
926    
927    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
928        original author of that file, following a query about its status.
929    
930    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
931        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
932    
933    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
934        quantifier applied to a forward-referencing subroutine call, could compile
935        incorrect code or give the error "internal error: previously-checked
936        referenced subpattern not found".
937    
938    6.  Both MS Visual Studio and Symbian OS have problems with initializing
939        variables to point to external functions. For these systems, therefore,
940        pcre_malloc etc. are now initialized to local functions that call the
941        relevant global functions.
942    
943    7.  There were two entries missing in the vectors called coptable and poptable
944        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
945        I've fixed the data, and added a kludgy way of testing at compile time that
946        the lengths are correct (equal to the number of opcodes).
947    
948    8.  Following on from 7, I added a similar kludge to check the length of the
949        eint vector in pcreposix.c.
950    
951    9.  Error texts for pcre_compile() are held as one long string to avoid too
952        much relocation at load time. To find a text, the string is searched,
953        counting zeros. There was no check for running off the end of the string,
954        which could happen if a new error number was added without updating the
955        string.
956    
957    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
958    
959    11. \K was not working if it appeared in an atomic group or in a group that
960        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
961        \K is "not well defined" if used in an assertion. PCRE now accepts it if
962        the assertion is positive, but not if it is negative.
963    
964    12. Change 11 fortuitously reduced the size of the stack frame used in the
965        "match()" function of pcre_exec.c by one pointer. Forthcoming
966        implementation of support for (*MARK) will need an extra pointer on the
967        stack; I have reserved it now, so that the stack frame size does not
968        decrease.
969    
970    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
971        item in branch that calls a recursion is a subroutine call - as in the
972        second branch in the above example - was incorrectly given the compile-
973        time error "recursive call could loop indefinitely" because pcre_compile()
974        was not correctly checking the subroutine for matching a non-empty string.
975    
976    14. The checks for overrunning compiling workspace could trigger after an
977        overrun had occurred. This is a "should never occur" error, but it can be
978        triggered by pathological patterns such as hundreds of nested parentheses.
979        The checks now trigger 100 bytes before the end of the workspace.
980    
981    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
982    
983    
984    Version 8.01 19-Jan-2010
985    ------------------------
986    
987    1.  If a pattern contained a conditional subpattern with only one branch (in
988        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
989        computed the wrong minimum data length (which is of course zero for such
990        subpatterns). This could cause incorrect "no match" results.
991    
992    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
993        the pattern is reset in the first branch, pcre_compile() failed with
994        "internal error: code overflow at offset...". This happened only when
995        the reset was to the original external option setting. (An optimization
996        abstracts leading options settings into an external setting, which was the
997        cause of this.)
998    
999    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1000        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1001        assertion pattern did not match (meaning that the assertion was true), it
1002        was incorrectly treated as false if the SKIP had been reached during the
1003        matching. This also applied to assertions used as conditions.
1004    
1005    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1006        assertion subpattern, including such a pattern used as a condition,
1007        unpredictable results occurred, instead of the error return
1008        PCRE_ERROR_DFA_UITEM.
1009    
1010    5.  The C++ GlobalReplace function was not working like Perl for the special
1011        situation when an empty string is matched. It now does the fancy magic
1012        stuff that is necessary.
1013    
1014    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1015        removed. (These were left over from very, very early versions of PCRE.)
1016    
1017    7.  Some cosmetic changes to the code to make life easier when compiling it
1018        as part of something else:
1019    
1020        (a) Change DEBUG to PCRE_DEBUG.
1021    
1022        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1023            called "current" as "current_branch", to prevent a collision with the
1024            Linux macro when compiled as a kernel module.
1025    
1026        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1027            prevent a collision with the Linux macro when compiled as a kernel
1028            module.
1029    
1030    8.  In pcre_compile() there are some checks for integer overflows that used to
1031        cast potentially large values to (double). This has been changed to that
1032        when building, a check for int64_t is made, and if it is found, it is used
1033        instead, thus avoiding the use of floating point arithmetic. (There is no
1034        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1035        double.
1036    
1037    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1038        2005 (difference between two addresses compared to an unsigned value).
1039    
1040    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1041        custom one, because of the following reported problem in Windows:
1042    
1043          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1044              under Win32.
1045          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1046              therefore missing the function definition.
1047          - The compiler thus generates a "C" signature for the test function.
1048          - The linker fails to find the "C" function.
1049          - PCRE fails to configure if asked to do so against libbz2.
1050    
1051    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1052        messages were output:
1053    
1054          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1055          rerunning libtoolize, to keep the correct libtool macros in-tree.
1056          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1057    
1058        I have done both of these things.
1059    
1060    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1061        most of the time, it *can* run out if it is given a pattern that contains a
1062        runaway infinite recursion. I updated the discussion in the pcrestack man
1063        page.
1064    
1065    13. Now that we have gone to the x.xx style of version numbers, the minor
1066        version may start with zero. Using 08 or 09 is a bad idea because users
1067        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1068        interpreted as invalid octal numbers. I've updated the previous comment in
1069        configure.ac, and also added a check that gives an error if 08 or 09 are
1070        used.
1071    
1072    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1073        causing partial matching to fail when the end of the subject matched \W
1074        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1075    
1076    15. There were some discrepancies between the declarations in pcre_internal.h
1077        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1078        their definitions. The declarations used "const uschar *" and the
1079        definitions used USPTR. Even though USPTR is normally defined as "const
1080        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1081        reported that: "This difference in casting confuses some C++ compilers, for
1082        example, SunCC recognizes above declarations as different functions and
1083        generates broken code for hbpcre." I have changed the declarations to use
1084        USPTR.
1085    
1086    16. GNU libtool is named differently on some systems. The autogen.sh script now
1087        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1088        (FreeBSD).
1089    
1090    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1091        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1092        comment: "Figure out how to create a longlong from a string: strtoll and
1093        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1094        instance, but it only takes 2 args instead of 3!"
1095    
1096    18. A subtle bug concerned with back references has been fixed by a change of
1097        specification, with a corresponding code fix. A pattern such as
1098        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1099        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1100        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1101        same bug. Such groups have to be quantified to be useful, or contained
1102        inside another quantified group. (If there's no repetition, the reference
1103        can never match.) The problem arises because, having left the group and
1104        moved on to the rest of the pattern, a later failure that backtracks into
1105        the group uses the captured value from the final iteration of the group
1106        rather than the correct earlier one. I have fixed this in PCRE by forcing
1107        any group that contains a reference to itself to be an atomic group; that
1108        is, there cannot be any backtracking into it once it has completed. This is
1109        similar to recursive and subroutine calls.
1110    
1111    
1112    Version 8.00 19-Oct-09
1113    ----------------------
1114    
1115    1.  The table for translating pcre_compile() error codes into POSIX error codes
1116        was out-of-date, and there was no check on the pcre_compile() error code
1117        being within the table. This could lead to an OK return being given in
1118        error.
1119    
1120    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1121        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1122        in a Windows environment.
1123    
1124    3.  The pcregrep --count option prints the count for each file even when it is
1125        zero, as does GNU grep. However, pcregrep was also printing all files when
1126        --files-with-matches was added. Now, when both options are given, it prints
1127        counts only for those files that have at least one match. (GNU grep just
1128        prints the file name in this circumstance, but including the count seems
1129        more useful - otherwise, why use --count?) Also ensured that the
1130        combination -clh just lists non-zero counts, with no names.
1131    
1132    4.  The long form of the pcregrep -F option was incorrectly implemented as
1133        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1134        but it seems right to fix it, and I didn't think it was worth preserving
1135        the old behaviour.
1136    
1137    5.  The command line items --regex=pattern and --regexp=pattern were not
1138        recognized by pcregrep, which required --regex pattern or --regexp pattern
1139        (with a space rather than an '='). The man page documented the '=' forms,
1140        which are compatible with GNU grep; these now work.
1141    
1142    6.  No libpcreposix.pc file was created for pkg-config; there was just
1143        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1144    
1145    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1146        when UCP support is not needed, by modifying the Python script that
1147        generates it from Unicode data files. This should not matter if the module
1148        is correctly used as a library, but I received one complaint about 50K of
1149        unwanted data. My guess is that the person linked everything into his
1150        program rather than using a library. Anyway, it does no harm.
1151    
1152    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1153        was a minimum greater than 1 for a wide character in a possessive
1154        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1155        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1156        character. Chaos in the form of incorrect output or a compiling loop could
1157        result.
1158    
1159    9.  The restrictions on what a pattern can contain when partial matching is
1160        requested for pcre_exec() have been removed. All patterns can now be
1161        partially matched by this function. In addition, if there are at least two
1162        slots in the offset vector, the offset of the earliest inspected character
1163        for the match and the offset of the end of the subject are set in them when
1164        PCRE_ERROR_PARTIAL is returned.
1165    
1166    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1167        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1168        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1169        and may be more useful for multi-segment matching.
1170    
1171    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1172        used to be given if ever the end of the subject was reached; now it is
1173        given only if matching could not proceed because another character was
1174        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1175        string "Z", which now yields "no match" instead of "partial match". In the
1176        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1177        final character ended with (*FAIL).
1178    
1179    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1180        if the pattern had a "must contain" character that was already found in the
1181        earlier partial match, unless partial matching was again requested. For
1182        example, with the pattern /dog.(body)?/, the "must contain" character is
1183        "g". If the first part-match was for the string "dog", restarting with
1184        "sbody" failed. This bug has been fixed.
1185    
1186    13. The string returned by pcre_dfa_exec() after a partial match has been
1187        changed so that it starts at the first inspected character rather than the
1188        first character of the match. This makes a difference only if the pattern
1189        starts with a lookbehind assertion or \b or \B (\K is not supported by
1190        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1191        matching functions compatible, and I think it's the right thing to do.
1192    
1193    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1194        so that the demonstration program is easily available in environments where
1195        PCRE has not been installed from source.
1196    
1197    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1198        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1199        library.
1200    
1201    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1202        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1203        is not the first non-POSIX option to be added. Clearly some people find
1204        these options useful.
1205    
1206    17. If a caller to the POSIX matching function regexec() passes a non-zero
1207        value for nmatch with a NULL value for pmatch, the value of
1208        nmatch is forced to zero.
1209    
1210    18. RunGrepTest did not have a test for the availability of the -u option of
1211        the diff command, as RunTest does. It now checks in the same way as
1212        RunTest, and also checks for the -b option.
1213    
1214    19. If an odd number of negated classes containing just a single character
1215        interposed, within parentheses, between a forward reference to a named
1216        subpattern and the definition of the subpattern, compilation crashed with
1217        an internal error, complaining that it could not find the referenced
1218        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1219        [The bug was that it was starting one character too far in when skipping
1220        over the character class, thus treating the ] as data rather than
1221        terminating the class. This meant it could skip too much.]
1222    
1223    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1224        /g option in pcretest when the pattern contains \K, which makes it possible
1225        to have an empty string match not at the start, even when the pattern is
1226        anchored. Updated pcretest and pcredemo to use this option.
1227    
1228    21. If the maximum number of capturing subpatterns in a recursion was greater
1229        than the maximum at the outer level, the higher number was returned, but
1230        with unset values at the outer level. The correct (outer level) value is
1231        now given.
1232    
1233    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1234        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1235        make it do so. The string so far is captured, making this feature
1236        compatible with Perl.
1237    
1238    23. The tests have been re-organized, adding tests 11 and 12, to make it
1239        possible to check the Perl 5.10 features against Perl 5.10.
1240    
1241    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1242        pattern matches a fixed length string. PCRE did not allow this; now it
1243        does. Neither allows recursion.
1244    
1245    25. I finally figured out how to implement a request to provide the minimum
1246        length of subject string that was needed in order to match a given pattern.
1247        (It was back references and recursion that I had previously got hung up
1248        on.) This code has now been added to pcre_study(); it finds a lower bound
1249        to the length of subject needed. It is not necessarily the greatest lower
1250        bound, but using it to avoid searching strings that are too short does give
1251        some useful speed-ups. The value is available to calling programs via
1252        pcre_fullinfo().
1253    
1254    26. While implementing 25, I discovered to my embarrassment that pcretest had
1255        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1256        study optimizations had never been tested with that matching function.
1257        Oops. What is worse, even when it was passed study data, there was a bug in
1258        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1259        were also very few tests of studied patterns with pcre_dfa_exec().
1260    
1261    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1262        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1263        on the other side of the coin, they are no longer allowed to have different
1264        names, because these cannot be distinguished in PCRE, and this has caused
1265        confusion. (This is a difference from Perl.)
1266    
1267    28. When duplicate subpattern names are present (necessarily with different
1268        numbers, as required by 27 above), and a test is made by name in a
1269        conditional pattern, either for a subpattern having been matched, or for
1270        recursion in such a pattern, all the associated numbered subpatterns are
1271        tested, and the overall condition is true if the condition is true for any
1272        one of them. This is the way Perl works, and is also more like the way
1273        testing by number works.
1274    
1275    
1276    Version 7.9 11-Apr-09
1277    ---------------------
1278    
1279    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1280        (pcretest), all targets were linked against these libraries. This included
1281        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1282        libraries. This caused unwanted dependencies to be created. This problem
1283        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1284        pcretest is linked with readline.
1285    
1286    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1287        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1288        moved outside it again, because FALSE and TRUE are already defined in AIX,
1289        but BOOL is not.
1290    
1291    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1292        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1293    
1294    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1295        hyphens) following file names and line numbers when outputting matching
1296        lines. This is not true; no spaces are inserted. I have also clarified the
1297        wording for the --colour (or --color) option.
1298    
1299    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1300        was not coloured; this is different to GNU grep, so I have changed it to be
1301        the same.
1302    
1303    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1304        each matching line was coloured. Now it goes on to look for further matches
1305        of any of the test patterns, which is the same behaviour as GNU grep.
1306    
1307    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1308        doesn't make sense to accept an empty string match in pcregrep, so I have
1309        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1310        seems to be how GNU grep behaves.
1311    
1312    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1313        start or after a newline", because the conditional assertion was not being
1314        correctly handled. The rule now is that both the assertion and what follows
1315        in the first alternative must satisfy the test.
1316    
1317    9.  If auto-callout was enabled in a pattern with a conditional group whose
1318        condition was an assertion, PCRE could crash during matching, both with
1319        pcre_exec() and pcre_dfa_exec().
1320    
1321    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1322        used for matching.
1323    
1324    11. Unicode property support in character classes was not working for
1325        characters (bytes) greater than 127 when not in UTF-8 mode.
1326    
1327    12. Added the -M command line option to pcretest.
1328    
1329    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1330    
1331    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1332    
1333    16. Added comments and documentation about mis-use of no_arg in the C++
1334        wrapper.
1335    
1336    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1337        from Martin Jerabek that uses macro names for all relevant character and
1338        string constants.
1339    
1340    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1341        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1342        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1343        these, but not everybody uses configure.
1344    
1345    19. A conditional group that had only one branch was not being correctly
1346        recognized as an item that could match an empty string. This meant that an
1347        enclosing group might also not be so recognized, causing infinite looping
1348        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1349        with the subject "ab", where knowledge that the repeated group can match
1350        nothing is needed in order to break the loop.
1351    
1352    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1353        exec(), but without supplying a callout function, matching went wrong.
1354    
1355    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1356        leak if the size of the offset vector was greater than 30. When the vector
1357        is smaller, the saved offsets during recursion go onto a local stack
1358        vector, but for larger vectors malloc() is used. It was failing to free
1359        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1360        error, in fact).
1361    
1362    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1363        heapframe that is used only when UTF-8 support is enabled. This caused no
1364        problem, but was untidy.
1365    
1366    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1367        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1368        included within another project.
1369    
1370    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1371        slightly modified by me:
1372    
1373          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1374              not building pcregrep.
1375    
1376          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1377              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1378    
1379    25. Forward references, both numeric and by name, in patterns that made use of
1380        duplicate group numbers, could behave incorrectly or give incorrect errors,
1381        because when scanning forward to find the reference group, PCRE was not
1382        taking into account the duplicate group numbers. A pattern such as
1383        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1384    
1385    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1386        the feature of a custom pointer more persuasive (as requested by a user).
1387    
1388    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1389        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1390        pre-defined.
1391    
1392    28. Added support for (*UTF8) at the start of a pattern.
1393    
1394    29. Arrange for flags added by the "release type" setting in CMake to be shown
1395        in the configuration summary.
1396    
1397    
1398    Version 7.8 05-Sep-08
1399    ---------------------
1400    
1401    1.  Replaced UCP searching code with optimized version as implemented for Ad
1402        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1403        stage table and inline lookup instead of a function, giving speed ups of 2
1404        to 5 times on some simple patterns that I tested. Permission was given to
1405        distribute the MultiStage2.py script that generates the tables (it's not in
1406        the tarball, but is in the Subversion repository).
1407    
1408    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1409        scripts.
1410    
1411    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1412        a group with a zero qualifier. The result of the study could be incorrect,
1413        or the function might crash, depending on the pattern.
1414    
1415    4.  Caseless matching was not working for non-ASCII characters in back
1416        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1417        It now works when Unicode Property Support is available.
1418    
1419    5.  In pcretest, an escape such as \x{de} in the data was always generating
1420        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1421        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1422        truncation.
1423    
1424    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1425    
1426    7.  Added two (int) casts to pcregrep when printing the difference of two
1427        pointers, in case they are 64-bit values.
1428    
1429    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1430        test 2 if it fails.
1431    
1432    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1433        and a #define of that name to empty if it is not externally set. This is to
1434        allow users of MSVC to set it if necessary.
1435    
1436    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1437        the convenience functions in the pcre_get.c source file.
1438    
1439    11. An option change at the start of a pattern that had top-level alternatives
1440        could cause overwriting and/or a crash. This command provoked a crash in
1441        some environments:
1442    
1443          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1444    
1445        This potential security problem was recorded as CVE-2008-2371.
1446    
1447    12. For a pattern where the match had to start at the beginning or immediately
1448        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1449        pcre_dfa_exec() could read past the end of the passed subject if there was
1450        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1451        pcretest so that it places the subject at the end of its malloc-ed buffer.
1452    
1453    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1454        exec() might read past the end of the data buffer in UTF-8 mode.
1455    
1456    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1457        the data contained the byte 0x85 as part of a UTF-8 character within its
1458        first line. This applied both to normal and DFA matching.
1459    
1460    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1461        /^[^d]*?$/8 failed to match "abc".
1462    
1463    16. Added a missing copyright notice to pcrecpp_internal.h.
1464    
1465    17. Make it more clear in the documentation that values returned from
1466        pcre_exec() in ovector are byte offsets, not character counts.
1467    
1468    18. Tidied a few places to stop certain compilers from issuing warnings.
1469    
1470    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1471        supplied by Stefan Weber. I made a further small update for 7.8 because
1472        there is a change of source arrangements: the pcre_searchfuncs.c module is
1473        replaced by pcre_ucd.c.
1474    
1475    
1476    Version 7.7 07-May-08
1477    ---------------------
1478    
1479    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1480        a string to a long long, pretend we don't even have a long long." This is
1481        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1482    
1483    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1484        pre-7.6 versions, which defined a global no_arg variable instead of putting
1485        it in the RE class. (See also #8 below.)
1486    
1487    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1488        Lopes.
1489    
1490    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1491    
1492        (1) The include/exclude patterns were being applied to the whole pathnames
1493            of files, instead of just to the final components.
1494    
1495        (2) If there was more than one level of directory, the subdirectories were
1496            skipped unless they satisfied the include/exclude conditions. This is
1497            inconsistent with GNU grep (and could even be seen as contrary to the
1498            pcregrep specification - which I improved to make it absolutely clear).
1499            The action now is always to scan all levels of directory, and just
1500            apply the include/exclude patterns to regular files.
1501    
1502    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1503        --exclude_dir in the tests to avoid scanning .svn directories.
1504    
1505    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1506        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1507        doesn't support NULs in patterns.
1508    
1509    7.  Added some missing "const"s to declarations of static tables in
1510        pcre_compile.c and pcre_dfa_exec.c.
1511    
1512    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1513        caused by fix #2  above. (Subsequently also a second patch to fix the
1514        first patch. And a third patch - this was a messy problem.)
1515    
1516    9.  Applied Craig's patch to remove the use of push_back().
1517    
1518    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1519        matching function regexec().
1520    
1521    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1522        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1523        references. PCRE supports relative numbers with this syntax (I don't think
1524        Oniguruma does).
1525    
1526    12. Previously, a group with a zero repeat such as (...){0} was completely
1527        omitted from the compiled regex. However, this means that if the group
1528        was called as a subroutine from elsewhere in the pattern, things went wrong
1529        (an internal error was given). Such groups are now left in the compiled
1530        pattern, with a new opcode that causes them to be skipped at execution
1531        time.
1532    
1533    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1534        to the way PCRE behaves:
1535    
1536        (a) A lone ] character is dis-allowed (Perl treats it as data).
1537    
1538        (b) A back reference to an unmatched subpattern matches an empty string
1539            (Perl fails the current match path).
1540    
1541        (c) A data ] in a character class must be notated as \] because if the
1542            first data character in a class is ], it defines an empty class. (In
1543            Perl it is not possible to have an empty class.) The empty class []
1544            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1545            The negative empty class [^] matches any one character, independently
1546            of the DOTALL setting.
1547    
1548    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1549        non-existent subpattern following a character class starting with ']' and
1550        containing () gave an internal compiling error instead of "reference to
1551        non-existent subpattern". Fortunately, when the pattern did exist, the
1552        compiled code was correct. (When scanning forwards to check for the
1553        existencd of the subpattern, it was treating the data ']' as terminating
1554        the class, so got the count wrong. When actually compiling, the reference
1555        was subsequently set up correctly.)
1556    
1557    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1558        it was being rejected as not supported by pcre_dfa_exec(), even though
1559        other assertions are supported. I have made pcre_dfa_exec() support
1560        (*FAIL).
1561    
1562    16. The implementation of 13c above involved the invention of a new opcode,
1563        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1564        cannot be changed at match time, I realized I could make a small
1565        improvement to matching performance by compiling OP_ALLANY instead of
1566        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1567        on the OP_ANY path.
1568    
1569    17. Compiling pcretest on Windows with readline support failed without the
1570        following two fixes: (1) Make the unistd.h include conditional on
1571        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1572    
1573    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1574        ncurses library to be included for pcretest when ReadLine support is
1575        requested, but also to allow for it to be overridden. This patch came from
1576        Daniel Bergström.
1577    
1578    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1579        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1580        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1581        spotting this.
1582    
1583    
1584    Version 7.6 28-Jan-08
1585    ---------------------
1586    
1587    1.  A character class containing a very large number of characters with
1588        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1589        overflow.
1590    
1591    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1592        HAVE_LONG_LONG is not defined.
1593    
1594    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1595        bring them up to date and include new features. This patch includes:
1596    
1597        - Fixed PH's badly added libz and libbz2 support.
1598        - Fixed a problem with static linking.
1599        - Added pcredemo. [But later removed - see 7 below.]
1600        - Fixed dftables problem and added an option.
1601        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1602            HAVE_LONG_LONG.
1603        - Added readline support for pcretest.
1604        - Added an listing of the option settings after cmake has run.
1605    
1606    4.  A user submitted a patch to Makefile that makes it easy to create
1607        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1608        Makefile.am that cause it to include this special target, without
1609        affecting anything else. Note that the same mingw target plus all
1610        the other distribution libraries and programs are now supported
1611        when configuring with CMake (see 6 below) instead of with
1612        Configure/Make.
1613    
1614    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1615        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1616        exported in the Windows port". It has not yet been confirmed that the patch
1617        solves the problem, but it does no harm.
1618    
1619    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1620        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1621        with CMake, and also correct the comment about stack recursion.
1622    
1623    7.  Remove the automatic building of pcredemo from the ./configure system and
1624        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1625        of a program that users should build themselves after PCRE is installed, so
1626        building it automatically is not really right. What is more, it gave
1627        trouble in some build environments.
1628    
1629    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1630    
1631    
1632    Version 7.5 10-Jan-08
1633    ---------------------
1634    
1635    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1636        values in parens when parsing an RE using the C++ wrapper."
1637    
1638    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1639        Characters greater than 255 were excluded from the class instead of being
1640        included.
1641    
1642    3.  The same bug as (2) above applied to negated POSIX classes such as
1643        [:^space:].
1644    
1645    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1646        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1647        I have changed it.
1648    
1649    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1650        first named subpattern) and a construct such as (?&a) would reference the
1651        first named subpattern whose name started with "a" (in other words, the
1652        length check was missing). Both these problems are fixed. "Subpattern name
1653        expected" is now given for (?&) (a zero-length name), and this patch also
1654        makes it give the same error for \k'' (previously it complained that that
1655        was a reference to a non-existent subpattern).
1656    
1657    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1658        this is right because (?- can be followed by option settings as well as by
1659        digits. I have, however, made the messages clearer.
1660    
1661    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1662        than the number used in the conditional) now cause a compile-time error.
1663        This is actually not compatible with Perl, which accepts such patterns, but
1664        treats the conditional as always being FALSE (as PCRE used to), but it
1665        seems to me that giving a diagnostic is better.
1666    
1667    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1668        and messages.
1669    
1670    9.  Fix two occurrences of "backslash" in comments that should have been
1671        "backspace".
1672    
1673    10. Remove two redundant lines of code that can never be obeyed (their function
1674        was moved elsewhere).
1675    
1676    11. The program that makes PCRE's Unicode character property table had a bug
1677        which caused it to generate incorrect table entries for sequences of
1678        characters that have the same character type, but are in different scripts.
1679        It amalgamated them into a single range, with the script of the first of
1680        them. In other words, some characters were in the wrong script. There were
1681        thirteen such cases, affecting characters in the following ranges:
1682    
1683          U+002b0 - U+002c1
1684          U+0060c - U+0060d
1685          U+0061e - U+00612
1686          U+0064b - U+0065e
1687          U+0074d - U+0076d
1688          U+01800 - U+01805
1689          U+01d00 - U+01d77
1690          U+01d9b - U+01dbf
1691          U+0200b - U+0200f
1692          U+030fc - U+030fe
1693          U+03260 - U+0327f
1694          U+0fb46 - U+0fbb1
1695          U+10450 - U+1049d
1696    
1697    12. The -o option (show only the matching part of a line) for pcregrep was not
1698        compatible with GNU grep in that, if there was more than one match in a
1699        line, it showed only the first of them. It now behaves in the same way as
1700        GNU grep.
1701    
1702    13. If the -o and -v options were combined for pcregrep, it printed a blank
1703        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1704        does the same. The return code can be used to tell if there were any
1705        non-matching lines.
1706    
1707    14. Added --file-offsets and --line-offsets to pcregrep.
1708    
1709    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1710        infinitely looping recursion. The bug was that positive lookaheads were not
1711        being skipped when checking for a possible empty match (negative lookaheads
1712        and both kinds of lookbehind were skipped).
1713    
1714    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1715        inclusion of <windows.h> to before rather than after the definition of
1716        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1717    
1718    17. Specifying a possessive quantifier with a specific limit for a Unicode
1719        character property caused pcre_compile() to compile bad code, which led at
1720        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1721        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1722        caused the error; without that there was no problem.
1723    
1724    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1725    
1726    19. Added --enable-pcretest-libreadline.
1727    
1728    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1729        RE::GlobalReplace(). As a result, the number of replacements returned was
1730        double what it should be. I removed one of the increments, but Craig sent a
1731        later patch that removed the other one (the right fix) and added unit tests
1732        that check the return values (which was not done before).
1733    
1734    21. Several CMake things:
1735    
1736        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1737            the names libpcre and libpcreposix, not just pcre and pcreposix.
1738    
1739        (2) The above change means that pcretest and pcregrep are now correctly
1740            linked with the newly-built libraries, not previously installed ones.
1741    
1742        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1743    
1744    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1745        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1746        UTF-8 newline character). The key issue is that the pattern starts .*;
1747        this means that the match must be either at the beginning, or after a
1748        newline. The bug was in the code for advancing after a failed match and
1749        checking that the new position followed a newline. It was not taking
1750        account of UTF-8 characters correctly.
1751    
1752    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1753        character classes. PCRE was not treating the sequence [:...:] as a
1754        character class unless the ... were all letters. Perl, however, seems to
1755        allow any characters between [: and :], though of course it rejects as
1756        unknown any "names" that contain non-letters, because all the known class
1757        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1758        for example, whereas PCRE did not - it did not recognize a POSIX character
1759        class. This seemed a bit dangerous, so the code has been changed to be
1760        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1761        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1762        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1763        Perl does, and where it didn't before.
1764    
1765    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1766        Windows environments %n is disabled by default.
1767    
1768    
1769    Version 7.4 21-Sep-07
1770    ---------------------
1771    
1772    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1773        means that a class such as [\s] counted as "explicit reference to CR or
1774        LF". That isn't really right - the whole point of the change was to try to
1775        help when there was an actual mention of one of the two characters. So now
1776        the change happens only if \r or \n (or a literal CR or LF) character is
1777        encountered.
1778    
1779    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1780        of both had grown to the point where there were only 3 bits left.
1781        Fortunately, there was spare space in the data structure, and so I have
1782        moved the internal flags into a new 16-bit field to free up more option
1783        bits.
1784    
1785    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1786        but did not set the internal JCHANGED flag - either of these is enough to
1787        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1788        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1789        start sets both bits.
1790    
1791    4.  Added options (at build time, compile time, exec time) to change \R from
1792        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1793    
1794    5.  doc/pcresyntax.html was missing from the distribution.
1795    
1796    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1797        compatibility, even though it is no longer used.
1798    
1799    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1800        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1801        windows.h file is present (where different names are used). [This was
1802        reversed later after testing - see 16 below.]
1803    
1804    8.  Changed all #include <config.h> to #include "config.h". There were also
1805        some further <pcre.h> cases that I changed to "pcre.h".
1806    
1807    9.  When pcregrep was used with the --colour option, it missed the line ending
1808        sequence off the lines that it output.
1809    
1810    10. It was pointed out to me that arrays of string pointers cause lots of
1811        relocations when a shared library is dynamically loaded. A technique of
1812        using a single long string with a table of offsets can drastically reduce
1813        these. I have refactored PCRE in four places to do this. The result is
1814        dramatic:
1815    
1816          Originally:                          290
1817          After changing UCP table:            187
1818          After changing error message table:   43
1819          After changing table of "verbs"       36
1820          After changing table of Posix names   22
1821    
1822        Thanks to the folks working on Gregex for glib for this insight.
1823    
1824    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1825        unicode-properties was also set.
1826    
1827    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1828    
1829    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1830        checked only for CRLF.
1831    
1832    14. Added casts to pcretest.c to avoid compiler warnings.
1833    
1834    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1835    
1836    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1837        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1838        entirely. This removes changes made in 7 above.
1839    
1840    17. The CMake files have been updated, and there is now more information about
1841        building with CMake in the NON-UNIX-USE document.
1842    
1843    
1844    Version 7.3 28-Aug-07
1845    ---------------------
1846    
1847     1. In the rejigging of the build system that eventually resulted in 7.1, the
1848        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1849        brackets there is not right, since it causes compilers to look for an
1850        installed pcre.h, not the version that is in the source that is being
1851        compiled (which of course may be different). I have changed it back to:
1852    
1853          #include "pcre.h"
1854    
1855        I have a vague recollection that the change was concerned with compiling in
1856        different directories, but in the new build system, that is taken care of
1857        by the VPATH setting the Makefile.
1858    
1859     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1860        when the subject happened to end in the byte 0x85 (e.g. if the last
1861        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1862        characters but of course it shouldn't be taken as a newline when it is part
1863        of another character. The bug was that, for an unlimited repeat of . in
1864        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1865        characters when looking for a newline.
1866    
1867     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1868    
1869     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1870        in debug output.
1871    
1872     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1873        long printing in the pcrecpp unittest when running under MinGW.
1874    
1875     6. ESC_K was left out of the EBCDIC table.
1876    
1877     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1878        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1879        limit also applies to "virtual nesting" when a pattern is recursive, and in
1880        this case 1000 isn't so big. I have been able to remove this limit at the
1881        expense of backing off one optimization in certain circumstances. Normally,
1882        when pcre_exec() would call its internal match() function recursively and
1883        immediately return the result unconditionally, it uses a "tail recursion"
1884        feature to save stack. However, when a subpattern that can match an empty
1885        string has an unlimited repetition quantifier, it no longer makes this
1886        optimization. That gives it a stack frame in which to save the data for
1887        checking that an empty string has been matched. Previously this was taken
1888        from the 1000-entry workspace that had been reserved. So now there is no
1889        explicit limit, but more stack is used.
1890    
1891     8. Applied Daniel's patches to solve problems with the import/export magic
1892        syntax that is required for Windows, and which was going wrong for the
1893        pcreposix and pcrecpp parts of the library. These were overlooked when this
1894        problem was solved for the main library.
1895    
1896     9. There were some crude static tests to avoid integer overflow when computing
1897        the size of patterns that contain repeated groups with explicit upper
1898        limits. As the maximum quantifier is 65535, the maximum group length was
1899        set at 30,000 so that the product of these two numbers did not overflow a
1900        32-bit integer. However, it turns out that people want to use groups that
1901        are longer than 30,000 bytes (though not repeat them that many times).
1902        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1903        made it possible to implement the integer overflow checks in a much more
1904        dynamic way, which I have now done. The artificial limitation on group
1905        length has been removed - we now have only the limit on the total length of
1906        the compiled pattern, which depends on the LINK_SIZE setting.
1907    
1908    10. Fixed a bug in the documentation for get/copy named substring when
1909        duplicate names are permitted. If none of the named substrings are set, the
1910        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1911        empty string.
1912    
1913    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1914        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1915        because the ] is interpreted as the first data character and the
1916        terminating ] is not found. PCRE has been made compatible with Perl in this
1917        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1918        cause memory overwriting.
1919    
1920    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1921        string has been matched (to stop an infinite loop). It was not recognizing
1922        a conditional subpattern that could match an empty string if that
1923        subpattern was within another subpattern. For example, it looped when
1924        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1925        condition was not nested. This bug has been fixed.
1926    
1927    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1928        past the start of the subject in the presence of bytes with the top bit
1929        set, for example "\x8aBCD".
1930    
1931    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1932        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1933    
1934    14. Optimized (?!) to (*FAIL).
1935    
1936    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1937        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1938        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1939        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1940        does: it's just the validity check that is more restrictive.
1941    
1942    16. Inserted checks for integer overflows during escape sequence (backslash)
1943        processing, and also fixed erroneous offset values for syntax errors during
1944        backslash processing.
1945    
1946    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1947        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1948    
1949    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1950        caused an overrun.
1951    
1952    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1953        something other than just ASCII characters) inside a group that had an
1954        unlimited repeat caused a loop at compile time (while checking to see
1955        whether the group could match an empty string).
1956    
1957    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1958        [\P{Any}] did so. (Error in the code for printing property names.)
1959    
1960    21. An orphan \E inside a character class could cause a crash.
1961    
1962    22. A repeated capturing bracket such as (A)? could cause a wild memory
1963        reference during compilation.
1964    
1965    23. There are several functions in pcre_compile() that scan along a compiled
1966        expression for various reasons (e.g. to see if it's fixed length for look
1967        behind). There were bugs in these functions when a repeated \p or \P was
1968        present in the pattern. These operators have additional parameters compared
1969        with \d, etc, and these were not being taken into account when moving along
1970        the compiled data. Specifically:
1971    
1972        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1973            length.
1974    
1975        (b) An item such as \pL+ within a repeated group could cause crashes or
1976            loops.
1977    
1978        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1979            "reference to non-existent subpattern" error.
1980    
1981        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1982    
1983    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1984        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1985    
1986    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1987    
1988    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1989        character were causing crashes (broken optimization).
1990    
1991    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1992        \p or \P) caused a compile-time loop.
1993    
1994    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1995        break. For example, the unstudied pattern [\r\n]A does not match the string
1996        "\r\nA" because change 7.0/46 below moves the current point on by two
1997        characters after failing to match at the start. However, the pattern \nA
1998        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1999        the same is true. There doesn't seem any very clean way out of this, but
2000        what I have chosen to do makes the common cases work: PCRE now takes note
2001        of whether there can be an explicit match for \r or \n anywhere in the
2002        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2003        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2004        pattern has explicit CR or LF references.
2005    
2006    29. Added (*CR) etc for changing newline setting at start of pattern.
2007    
2008    
2009    Version 7.2 19-Jun-07
2010    ---------------------
2011    
2012     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2013        which is apparently normally available under Windows.
2014    
2015     2. Re-jig the pcregrep tests with different newline settings in an attempt
2016        to make them independent of the local environment's newline setting.
2017    
2018     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2019    
2020     4. Some of the "internals" tests were previously cut out when the link size
2021        was not 2, because the output contained actual offsets. The recent new
2022        "Z" feature of pcretest means that these can be cut out, making the tests
2023        usable with all link sizes.
2024    
2025     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2026        stack recursion. This gives a massive performance boost under BSD, but just
2027        a small improvement under Linux. However, it saves one field in the frame
2028        in all cases.
2029    
2030     6. Added more features from the forthcoming Perl 5.10:
2031    
2032        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2033            recursion call. It refers to the nth most recently opened parentheses.
2034    
2035        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2036            to be opened parentheses.
2037    
2038        (c) Conditions that refer to capturing parentheses can be specified
2039            relatively, for example, (?(-2)... or (?(+3)...
2040    
2041        (d) \K resets the start of the current match so that everything before
2042            is not part of it.
2043    
2044        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2045    
2046        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2047            reference syntax.
2048    
2049        (g) (?| introduces a group in which the numbering of parentheses in each
2050            alternative starts with the same number.
2051    
2052        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2053    
2054     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2055        PCRE_INFO_JCHANGED.
2056    
2057     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2058        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2059        for detecting groups that can match an empty string.
2060    
2061     9. A pattern with a very large number of alternatives (more than several
2062        hundred) was running out of internal workspace during the pre-compile
2063        phase, where pcre_compile() figures out how much memory will be needed. A
2064        bit of new cunning has reduced the workspace needed for groups with
2065        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2066        workspace instead of running out of the 4096 that are available.
2067    
2068    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2069    
2070    11. Applied patch from Google to remove an optimization that didn't quite work.
2071        The report of the bug said:
2072    
2073          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2074          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2075          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2076    
2077    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2078        it matched the wrong number of bytes.
2079    
2080    
2081    Version 7.1 24-Apr-07
2082    ---------------------
2083    
2084     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2085        that is more "standard", making use of automake and other Autotools. There
2086        is some re-arrangement of the files and adjustment of comments consequent
2087        on this.
2088    
2089     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2090        for recursive directory scanning broke on some systems because the files
2091        are not scanned in any specific order and on different systems the order
2092        was different. A call to "sort" has been inserted into RunGrepTest for the
2093        approprate test as a short-term fix. In the longer term there may be an
2094        alternative.
2095    
2096     3. I had an email from Eric Raymond about problems translating some of PCRE's
2097        man pages to HTML (despite the fact that I distribute HTML pages, some
2098        people do their own conversions for various reasons). The problems
2099        concerned the use of low-level troff macros .br and .in. I have therefore
2100        removed all such uses from the man pages (some were redundant, some could
2101        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2102        HTML has been updated to handle .nf/.fi and to complain if it encounters
2103        .br or .in.
2104    
2105     4. Updated comments in configure.ac that get placed in config.h.in and also
2106        arranged for config.h to be included in the distribution, with the name
2107        config.h.generic, for the benefit of those who have to compile without
2108        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2109    
2110     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2111        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2112        makevp.bat for the current PCRE, using the additional files
2113        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2114    
2115     6. A Windows user reported a minor discrepancy with test 2, which turned out
2116        to be caused by a trailing space on an input line that had got lost in his
2117        copy. The trailing space was an accident, so I've just removed it.
2118    
2119     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2120        that is needed.
2121    
2122     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2123        as "const" (a) because they are and (b) because it helps the PHP
2124        maintainers who have recently made a script to detect big data structures
2125        in the php code that should be moved to the .rodata section. I remembered
2126        to update Builducptable as well, so it won't revert if ucptable.h is ever
2127        re-created.
2128    
2129     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2130        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2131        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2132        support is not required. This saves 1.5-2K of code, which is important in
2133        some applications.
2134    
2135        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2136        so as not to refer to the tables, even though these functions will never be
2137        called when UTF-8 support is disabled. Otherwise there are problems with a
2138        shared library.
2139    
2140    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2141    
2142        (a) It was defining its arguments as char * instead of void *.
2143    
2144        (b) It was assuming that all moves were upwards in memory; this was true
2145            a long time ago when I wrote it, but is no longer the case.
2146    
2147        The emulated memove() is provided for those environments that have neither
2148        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2149        is clearly not the case, as these two bugs were recently reported.
2150    
2151    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2152        and Detrail to create the HTML documentation, the .txt form of the man
2153        pages, and it removes trailing spaces from listed files. It also creates
2154        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2155        case, it wraps all the #defines with #ifndefs. This script should be run
2156        before "make dist".
2157    
2158    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2159        with Unicode property support.
2160    
2161        (a) For a maximizing quantifier, if the two different cases of the
2162            character were of different lengths in their UTF-8 codings (there are
2163            some cases like this - I found 11), and the matching function had to
2164            back up over a mixture of the two cases, it incorrectly assumed they
2165            were both the same length.
2166    
2167        (b) When PCRE was configured to use the heap rather than the stack for
2168            recursion during matching, it was not correctly preserving the data for
2169            the other case of a UTF-8 character when checking ahead for a match
2170            while processing a minimizing repeat. If the check also involved
2171            matching a wide character, but failed, corruption could cause an
2172            erroneous result when trying to check for a repeat of the original
2173            character.
2174    
2175    13. Some tidying changes to the testing mechanism:
2176    
2177        (a) The RunTest script now detects the internal link size and whether there
2178            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2179            values substituted by "configure". (The RunGrepTest script already did
2180            this for UTF-8.) The configure.ac script no longer substitutes the
2181            relevant variables.
2182    
2183        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2184            with length and offset values. This means that the output is different
2185            for different internal link sizes. Test 2 is skipped for link sizes
2186            other than 2 because of this, bypassing the problem. Unfortunately,
2187            there was also a test in test 3 (the locale tests) that used /B and
2188            failed for link sizes other than 2. Rather than cut the whole test out,
2189            I have added a new /Z option to pcretest that replaces the length and
2190            offset values with spaces. This is now used to make test 3 independent
2191            of link size. (Test 2 will be tidied up later.)
2192    
2193    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2194        segmentation fault instead of returning the appropriate error message.
2195    
2196    15. In multiline mode when the newline sequence was set to "any", the pattern
2197        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2198        This doesn't seem right; it now treats the CRLF combination as the line
2199        ending, and so does not match in that case. It's only a pattern such as ^$
2200        that would hit this one: something like ^ABC$ would have failed after \r
2201        and then tried again after \r\n.
2202    
2203    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2204        in an attempt to make files that differ only in their line terminators
2205        compare equal. This works on Linux.
2206    
2207    17. Under certain error circumstances pcregrep might try to free random memory
2208        as it exited. This is now fixed, thanks to valgrind.
2209    
2210    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2211        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2212        was because its rules for how to advance for /g after matching an empty
2213        string at the end of a line did not allow for this case. They now check for
2214        it specially.
2215    
2216    20. pcretest is supposed to handle patterns and data of any length, by
2217        extending its buffers when necessary. It was getting this wrong when the
2218        buffer for a data line had to be extended.
2219    
2220    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2221        CRLF as a newline sequence.
2222    
2223    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2224        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2225        I have nevertheless tidied it up.
2226    
2227    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2228    
2229    24. Added a man page for pcre-config.
2230    
2231    
2232    Version 7.0 19-Dec-06
2233    ---------------------
2234    
2235     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2236        moving to gcc 4.1.1.
2237    
2238     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2239        sys/time.h, which is documented as needed for this function. It doesn't
2240        seem to matter on Linux, but it showed up on some releases of OS X.
2241    
2242     3. It seems that there are systems where bytes whose values are greater than
2243        127 match isprint() in the "C" locale. The "C" locale should be the
2244        default when a C program starts up. In most systems, only ASCII printing
2245        characters match isprint(). This difference caused the output from pcretest
2246        to vary, making some of the tests fail. I have changed pcretest so that:
2247    
2248        (a) When it is outputting text in the compiled version of a pattern, bytes
2249            other than 32-126 are always shown as hex escapes.
2250    
2251        (b) When it is outputting text that is a matched part of a subject string,
2252            it does the same, unless a different locale has been set for the match
2253            (using the /L modifier). In this case, it uses isprint() to decide.
2254    
2255     4. Fixed a major bug that caused incorrect computation of the amount of memory
2256        required for a compiled pattern when options that changed within the
2257        pattern affected the logic of the preliminary scan that determines the
2258        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2259        that the computed length was too small. The symptoms of this bug were
2260        either the PCRE error "internal error: code overflow" from pcre_compile(),
2261        or a glibc crash with a message such as "pcretest: free(): invalid next
2262        size (fast)". Examples of patterns that provoked this bug (shown in
2263        pcretest format) are:
2264    
2265          /(?-x: )/x
2266          /(?x)(?-x: \s*#\s*)/
2267          /((?i)[\x{c0}])/8
2268          /(?i:[\x{c0}])/8
2269    
2270        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2271        is now done differently.
2272    
2273     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2274        wrapper classes; (b) implement a new function in the C++ scanner that is
2275        more efficient than the old way of doing things because it avoids levels of
2276        recursion in the regex matching; (c) add a paragraph to the documentation
2277        for the FullMatch() function.
2278    
2279     6. The escape sequence \n was being treated as whatever was defined as
2280        "newline". Not only was this contrary to the documentation, which states
2281        that \n is character 10 (hex 0A), but it also went horribly wrong when
2282        "newline" was defined as CRLF. This has been fixed.
2283    
2284     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2285        was being set to -1 for the "end of line" case (supposedly a value that no
2286        character can have). Though this value is never used (the check for end of
2287        line is "zero bytes in current character"), it caused compiler complaints.
2288        I've changed it to 0xffffffff.
2289    
2290     8. In pcre_version.c, the version string was being built by a sequence of
2291        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2292        string (as it is for production releases) called a macro with an empty
2293        argument. The C standard says the result of this is undefined. The gcc
2294        compiler treats it as an empty string (which was what was wanted) but it is
2295        reported that Visual C gives an error. The source has been hacked around to
2296        avoid this problem.
2297    
2298     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2299        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2300        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2301        of them did).
2302    
2303    10. Originally, pcretest opened its input and output without "b"; then I was
2304        told that "b" was needed in some environments, so it was added for release
2305        5.0 to both the input and output. (It makes no difference on Unix-like
2306        systems.) Later I was told that it is wrong for the input on Windows. I've
2307        now abstracted the modes into two macros, to make it easier to fiddle with
2308        them, and removed "b" from the input mode under Windows.
2309    
2310    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2311    
2312    12. Added -help and --help to pcretest as an official way of being reminded
2313        of the options.
2314    
2315    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2316        and pcrecpp.cc because they annoy compilers at high warning levels.
2317    
2318    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2319    
2320    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2321        scripts are not C programs :-) and which was not noticed because it works
2322        on Linux.
2323    
2324    16. pcretest is supposed to handle any length of pattern and data line (as one
2325        line or as a continued sequence of lines) by extending its input buffer if
2326        necessary. This feature was broken for very long pattern lines, leading to
2327        a string of junk being passed to pcre_compile() if the pattern was longer
2328        than about 50K.
2329    
2330    17. I have done a major re-factoring of the way pcre_compile() computes the
2331        amount of memory needed for a compiled pattern. Previously, there was code
2332        that made a preliminary scan of the pattern in order to do this. That was
2333        OK when PCRE was new, but as the facilities have expanded, it has become
2334        harder and harder to keep it in step with the real compile phase, and there
2335        have been a number of bugs (see for example, 4 above). I have now found a
2336        cunning way of running the real compile function in a "fake" mode that
2337        enables it to compute how much memory it would need, while actually only
2338        ever using a few hundred bytes of working memory and without too many
2339        tests of the mode. This should make future maintenance and development
2340        easier. A side effect of this work is that the limit of 200 on the nesting
2341        depth of parentheses has been removed (though this was never a serious
2342        limitation, I suspect). However, there is a downside: pcre_compile() now
2343        runs more slowly than before (30% or more, depending on the pattern). I
2344        hope this isn't a big issue. There is no effect on runtime performance.
2345    
2346    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2347        newline (only possible for the last line of a file) and it was a
2348        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2349    
2350    19. Added additional timing features to pcretest. (1) The -tm option now times
2351        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2352        separate command line item, by a number that specifies the number of
2353        repeats to use when timing. The default is 50000; this gives better
2354        precision, but takes uncomfortably long for very large patterns.
2355    
2356    20. Extended pcre_study() to be more clever in cases where a branch of a
2357        subpattern has no definite first character. For example, (a*|b*)[cd] would
2358        previously give no result from pcre_study(). Now it recognizes that the
2359        first character must be a, b, c, or d.
2360    
2361    21. There was an incorrect error "recursive call could loop indefinitely" if
2362        a subpattern (or the entire pattern) that was being tested for matching an
2363        empty string contained only one non-empty item after a nested subpattern.
2364        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2365        incorrectly, because the \d was being skipped in the check.
2366    
2367    22. The pcretest program now has a new pattern option /B and a command line
2368        option -b, which is equivalent to adding /B to every pattern. This causes
2369        it to show the compiled bytecode, without the additional information that
2370        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2371        is the same as /B/I).
2372    
2373    23. A new optimization is now able automatically to treat some sequences such
2374        as a*b as a*+b. More specifically, if something simple (such as a character
2375        or a simple class like \d) has an unlimited quantifier, and is followed by
2376        something that cannot possibly match the quantified thing, the quantifier
2377        is automatically "possessified".
2378    
2379    24. A recursive reference to a subpattern whose number was greater than 39
2380        went wrong under certain circumstances in UTF-8 mode. This bug could also
2381        have affected the operation of pcre_study().
2382    
2383    25. Realized that a little bit of performance could be had by replacing
2384        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2385    
2386    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2387    
2388    27. Possessive quantifiers such as a++ were previously implemented by turning
2389        them into atomic groups such as ($>a+). Now they have their own opcodes,
2390        which improves performance. This includes the automatically created ones
2391        from 23 above.
2392    
2393    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2394        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2395        the first matched character to be a colon. This applied both to named and
2396        numbered groups.
2397    
2398    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2399    
2400    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2401        building PCRE on Windows easier, so I have included it in the distribution.
2402    
2403    31. There is now a check in pcretest against a ridiculously large number being
2404        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2405        loop, the loop is abandoned.
2406    
2407    32. Forward references to subpatterns in conditions such as (?(2)...) where
2408        subpattern 2 is defined later cause pcre_compile() to search forwards in
2409        the pattern for the relevant set of parentheses. This search went wrong
2410        when there were unescaped parentheses in a character class, parentheses
2411        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2412    
2413    33. "Subroutine" calls and backreferences were previously restricted to
2414        referencing subpatterns earlier in the regex. This restriction has now
2415        been removed.
2416    
2417    34. Added a number of extra features that are going to be in Perl 5.10. On the
2418        whole, these are just syntactic alternatives for features that PCRE had
2419        previously implemented using the Python syntax or my own invention. The
2420        other formats are all retained for compatibility.
2421    
2422        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2423            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2424            also .NET compatible.
2425    
2426        (b) A recursion or subroutine call to a named group can now be defined as
2427            (?&name) as well as (?P>name).
2428    
2429        (c) A backreference to a named group can now be defined as \k<name> or
2430            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2431            5.10, are also .NET compatible.
2432    
2433        (d) A conditional reference to a named group can now use the syntax
2434            (?(<name>) or (?('name') as well as (?(name).
2435    
2436        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2437            groups (named and numbered) that are never evaluated inline, but can be
2438            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2439            is always false. There may be only one alternative in such a group.
2440    
2441        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2442            as the simple (?(R). The condition is true only if the most recent
2443            recursion is that of the given number or name. It does not search out
2444            through the entire recursion stack.
2445    
2446        (g) The escape \gN or \g{N} has been added, where N is a positive or
2447            negative number, specifying an absolute or relative reference.
2448    
2449    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2450        some "unreachable code" warnings.
2451    
2452    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2453        things, this adds five new scripts.
2454    
2455    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2456        There were also incompatibilities regarding the handling of \Q..\E inside
2457        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2458        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2459    
2460    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2461        matches an empty string, and forcibly breaks the loop. There were bugs in
2462        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2463        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2464        separate and independent bugs (that affected different cases) have been
2465        fixed.
2466    
2467    39. Refactored the code to abolish the use of different opcodes for small
2468        capturing bracket numbers. This is a tidy that I avoided doing when I
2469        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2470        The new approach is not only tidier, it makes it possible to reduce the
2471        memory needed to fix the previous bug (38).
2472    
2473    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2474        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2475        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2476        mode.
2477    
2478    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2479        report.
2480    
2481    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2482        copy construction and assignment in the C++ wrapper.
2483    
2484    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2485        couple of bugs that could have given wrong results in the "--newline=crlf"
2486        case.
2487    
2488    44. Added a number of casts and did some reorganization of signed/unsigned int
2489        variables following suggestions from Dair Grant. Also renamed the variable
2490        "this" as "item" because it is a C++ keyword.
2491    
2492    45. Arranged for dftables to add
2493    
2494          #include "pcre_internal.h"
2495    
2496        to pcre_chartables.c because without it, gcc 4.x may remove the array
2497        definition from the final binary if PCRE is built into a static library and
2498        dead code stripping is activated.
2499    
2500    46. For an unanchored pattern, if a match attempt fails at the start of a
2501        newline sequence, and the newline setting is CRLF or ANY, and the next two
2502        characters are CRLF, advance by two characters instead of one.
2503    
2504    
2505    Version 6.7 04-Jul-06
2506    ---------------------
2507    
2508     1. In order to handle tests when input lines are enormously long, pcretest has
2509        been re-factored so that it automatically extends its buffers when
2510        necessary. The code is crude, but this _is_ just a test program. The
2511        default size has been increased from 32K to 50K.
2512    
2513     2. The code in pcre_study() was using the value of the re argument before
2514        testing it for NULL. (Of course, in any sensible call of the function, it
2515        won't be NULL.)
2516    
2517     3. The memmove() emulation function in pcre_internal.h, which is used on
2518        systems that lack both memmove() and bcopy() - that is, hardly ever -
2519        was missing a "static" storage class specifier.
2520    
2521     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2522        containing an extended class (one that cannot be represented by a bitmap
2523        because it contains high-valued characters or Unicode property items, e.g.
2524        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2525        pattern, but PCRE should not loop if you do not (it no longer does).
2526        [Detail: two cases were found: (a) a repeated subpattern containing an
2527        extended class; (b) a recursive reference to a subpattern that followed a
2528        previous extended class. It wasn't skipping over the extended class
2529        correctly when UTF-8 mode was not set.]
2530    
2531     5. A negated single-character class was not being recognized as fixed-length
2532        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2533        compile error "lookbehind assertion is not fixed length".
2534    
2535     6. The RunPerlTest auxiliary script was showing an unexpected difference
2536        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2537        write a Perl script that can interpret lines of an input file either as
2538        byte characters or as UTF-8, which is what "perltest" was being required to
2539        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2540        can't do is switch easily at run time between having the "use utf8;" pragma
2541        or not. In the end, I fudged it by using the RunPerlTest script to insert
2542        "use utf8;" explicitly for the UTF-8 tests.
2543    
2544     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2545        the end of the subject string, contrary to the documentation and to what
2546        Perl does. This was true of both matching functions. Now it matches only at
2547        the start of the subject and immediately after *internal* newlines.
2548    
2549     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2550        a pointer to an int instead of a pointer to an unsigned long int. This
2551        caused problems on 64-bit systems.
2552    
2553     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2554        instance of the 'standard' template library not being so standard".
2555    
2556    10. There was no check on the number of named subpatterns nor the maximum
2557        length of a subpattern name. The product of these values is used to compute
2558        the size of the memory block for a compiled pattern. By supplying a very
2559        long subpattern name and a large number of named subpatterns, the size
2560        computation could be caused to overflow. This is now prevented by limiting
2561        the length of names to 32 characters, and the number of named subpatterns
2562        to 10,000.
2563    
2564    11. Subpatterns that are repeated with specific counts have to be replicated in
2565        the compiled pattern. The size of memory for this was computed from the
2566        length of the subpattern and the repeat count. The latter is limited to
2567        65535, but there was no limit on the former, meaning that integer overflow
2568        could in principle occur. The compiled length of a repeated subpattern is
2569        now limited to 30,000 bytes in order to prevent this.
2570    
2571    12. Added the optional facility to have named substrings with the same name.
2572    
2573    13. Added the ability to use a named substring as a condition, using the
2574        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2575        are numbers (not recommended). Forward references are permitted.
2576    
2577    14. Added forward references in named backreferences (if you see what I mean).
2578    
2579    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2580        pattern could run off the end of the subject. For example, the pattern
2581        "(?s)(.{1,5})"8 did this with the subject "ab".
2582    
2583    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2584        PCRE_CASELESS was set when matching characters that were quantified with ?
2585        or *.
2586    
2587    17. A character class other than a single negated character that had a minimum
2588        but no maximum quantifier - for example [ab]{6,} - was not handled
2589        correctly by pce_dfa_exec(). It would match only one character.
2590    
2591    18. A valid (though odd) pattern that looked like a POSIX character
2592        class but used an invalid character after [ (for example [[,abc,]]) caused
2593        pcre_compile() to give the error "Failed: internal error: code overflow" or
2594        in some cases to crash with a glibc free() error. This could even happen if
2595        the pattern terminated after [[ but there just happened to be a sequence of
2596        letters, a binary zero, and a closing ] in the memory that followed.
2597    
2598    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2599        over the years. Originally (before any Unicode support), just the bottom 8
2600        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2601        output from "man perlunicode" includes this:
2602    
2603          The regular expression compiler produces polymorphic opcodes.  That
2604          is, the pattern adapts to the data and automatically switches to
2605          the Unicode character scheme when presented with Unicode data--or
2606          instead uses a traditional byte scheme when presented with byte
2607          data.
2608    
2609        Sadly, a wide octal escape does not cause a switch, and in a string with
2610        no other multibyte characters, these octal escapes are treated as before.
2611        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2612        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2613        Unicode string.
2614    
2615        I have not perpetrated such confusion in PCRE. Up till now, it took just
2616        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2617        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2618        translate to the appropriate multibyte character.
2619    
2620    29. Applied some refactoring to reduce the number of warnings from Microsoft
2621        and Borland compilers. This has included removing the fudge introduced
2622        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2623        a warning about an unused variable.
2624    
2625    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2626        characters since release 4.0, because Perl (from release 5.004) does not.
2627        [Or at least, is documented not to: some releases seem to be in conflict
2628        with the documentation.] However, when a pattern was studied with
2629        pcre_study() and all its branches started with \s, PCRE still included VT
2630        as a possible starting character. Of course, this did no harm; it just
2631        caused an unnecessary match attempt.
2632    
2633    22. Removed a now-redundant internal flag bit that recorded the fact that case
2634        dependency changed within the pattern. This was once needed for "required
2635        byte" processing, but is no longer used. This recovers a now-scarce options
2636        bit. Also moved the least significant internal flag bit to the most-
2637        significant bit of the word, which was not previously used (hangover from
2638        the days when it was an int rather than a uint) to free up another bit for
2639        the future.
2640    
2641    23. Added support for CRLF line endings as well as CR and LF. As well as the
2642        default being selectable at build time, it can now be changed at runtime
2643        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2644        specify that it is scanning data with non-default line endings.
2645    
2646    24. Changed the definition of CXXLINK to make it agree with the definition of
2647        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2648    
2649    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2650        recursions. This makes a big different to stack usage for some patterns.
2651    
2652    26. If a subpattern containing a named recursion or subroutine reference such
2653        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2654        the space required for the compiled pattern went wrong and gave too small a
2655        value. Depending on the environment, this could lead to "Failed: internal
2656        error: code overflow at offset 49" or "glibc detected double free or
2657        corruption" errors.
2658    
2659    27. Applied patches from Google (a) to support the new newline modes and (b) to
2660        advance over multibyte UTF-8 characters in GlobalReplace.
2661    
2662    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2663        difference for some implementation of PCRE in some Windows version.
2664    
2665    29. Added some extra testing facilities to pcretest:
2666    
2667        \q<number>   in a data line sets the "match limit" value
2668        \Q<number>   in a data line sets the "match recursion limt" value
2669        -S <number>  sets the stack size, where <number> is in megabytes
2670    
2671        The -S option isn't available for Windows.
2672    
2673    
2674    Version 6.6 06-Feb-06
2675    ---------------------
2676    
2677     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2678        in pcreposix.h. I have copied the definition from pcre.h.
2679    
2680     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2681        because pcre.h is no longer a built file.
2682    
2683     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2684        not normally included in the compiled code.
2685    
2686    
2687    Version 6.5 01-Feb-06
2688    ---------------------
2689    
2690     1. When using the partial match feature with pcre_dfa_exec(), it was not
2691        anchoring the second and subsequent partial matches at the new starting
2692        point. This could lead to incorrect results. For example, with the pattern
2693        /1234/, partially matching against "123" and then "a4" gave a match.
2694    
2695     2. Changes to pcregrep:
2696    
2697        (a) All non-match returns from pcre_exec() were being treated as failures
2698            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2699            error message is output. Some extra information is given for the
2700            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2701            probably the only errors that are likely to be caused by users (by
2702            specifying a regex that has nested indefinite repeats, for instance).
2703            If there are more than 20 of these errors, pcregrep is abandoned.
2704    
2705        (b) A binary zero was treated as data while matching, but terminated the
2706            output line if it was written out. This has been fixed: binary zeroes
2707            are now no different to any other data bytes.
2708    
2709        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2710            used to set a locale for matching. The --locale=xxxx long option has
2711            been added (no short equivalent) to specify a locale explicitly on the
2712            pcregrep command, overriding the environment variables.
2713    
2714        (d) When -B was used with -n, some line numbers in the output were one less
2715            than they should have been.
2716    
2717        (e) Added the -o (--only-matching) option.
2718    
2719        (f) If -A or -C was used with -c (count only), some lines of context were
2720            accidentally printed for the final match.
2721    
2722        (g) Added the -H (--with-filename) option.
2723    
2724        (h) The combination of options -rh failed to suppress file names for files
2725            that were found from directory arguments.
2726    
2727        (i) Added the -D (--devices) and -d (--directories) options.
2728    
2729        (j) Added the -F (--fixed-strings) option.
2730    
2731        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2732    
2733        (l) Added the --colo(u)r option.
2734    
2735        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2736            is not present by default.
2737    
2738     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2739        items such as (?R) or (?1), when the recursion could match a number of
2740        alternatives. If it matched one of the alternatives, but subsequently,
2741        outside the recursion, there was a failure, the code tried to back up into
2742        the recursion. However, because of the way PCRE is implemented, this is not
2743        possible, and the result was an incorrect result from the match.
2744    
2745        In order to prevent this happening, the specification of recursion has
2746        been changed so that all such subpatterns are automatically treated as
2747        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2748    
2749     4. I had overlooked the fact that, in some locales, there are characters for
2750        which isalpha() is true but neither isupper() nor islower() are true. In
2751        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2752        and ordfeminine) are like this. This affected the treatment of \w and \W
2753        when they appeared in character classes, but not when they appeared outside
2754        a character class. The bit map for "word" characters is now created
2755        separately from the results of isalnum() instead of just taking it from the
2756        upper, lower, and digit maps. (Plus the underscore character, of course.)
2757    
2758     5. The above bug also affected the handling of POSIX character classes such as
2759        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2760        permanent tables. Instead, the bit maps for such a class were previously
2761        created as the appropriate unions of the upper, lower, and digit bitmaps.
2762        Now they are created by subtraction from the [[:word:]] class, which has
2763        its own bitmap.
2764    
2765     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2766        It is created by subtracting the vertical space characters (\x09, \x0a,
2767        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2768        subtraction was done in the overall bitmap for a character class, meaning
2769        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2770        be recognized. This bug has been fixed.
2771    
2772     7. Patches from the folks at Google:
2773    
2774          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2775          real life, but is still worth protecting against".
2776    
2777          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2778          regular expressions".
2779    
2780          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2781          have it.
2782    
2783          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2784          "configure" and the latter not, in order to fix a problem somebody had
2785          with compiling the Arg class on HP-UX.
2786    
2787          (e) Improve the error-handling of the C++ wrapper a little bit.
2788    
2789          (f) New tests for checking recursion limiting.
2790    
2791     8. The pcre_memmove() function, which is used only if the environment does not
2792        have a standard memmove() function (and is therefore rarely compiled),
2793        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2794        returning a result (though PCRE never actually uses the result).
2795    
2796     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2797        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2798        returned instead of calling malloc() with an overflowing number that would
2799        most likely cause subsequent chaos.
2800    
2801    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2802    
2803    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2804        with this option is matched, the nmatch and pmatch options of regexec() are
2805        ignored.
2806    
2807    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2808        provided in case anyone wants to the the POSIX interface with UTF-8
2809        strings.
2810    
2811    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2812        C++ linking (needed for some HP-UX environments).
2813    
2814    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2815        (unused parameter) and in the pcre_printint() function (omitted "default"
2816        switch label when the default is to do nothing).
2817    
2818    15. Added some code to make it possible, when PCRE is compiled as a C++
2819        library, to replace subject pointers for pcre_exec() with a smart pointer
2820        class, thus making it possible to process discontinuous strings.
2821    
2822    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2823        much the same function. They were added by different people who were trying
2824        to make PCRE easy to compile on non-Unix systems. It has been suggested
2825        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2826        for compiling on Windows systems. I have therefore replaced it with
2827        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2828        defaults to "extern" for C or "extern C" for C++, which works fine on
2829        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2830        SCOPE with something explicit in config.h. In addition:
2831    
2832        (a) pcreposix.h still had just "extern" instead of either of these macros;
2833            I have replaced it with PCRE_DATA_SCOPE.
2834    
2835        (b) Functions such as _pcre_xclass(), which are internal to the library,
2836            but external in the C sense, all had PCRE_EXPORT in their definitions.
2837            This is apparently wrong for the Windows case, so I have removed it.
2838            (It makes no difference on Unix-like systems.)
2839    
2840    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2841        of recursive calls to match(). This is different to MATCH_LIMIT because
2842        that limits the total number of calls to match(), not all of which increase
2843        the depth of recursion. Limiting the recursion depth limits the amount of
2844        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2845        when PCRE is compiled, and changed at run time. A patch from Google adds
2846        this functionality to the C++ interface.
2847    
2848    18. Changes to the handling of Unicode character properties:
2849    
2850        (a) Updated the table to Unicode 4.1.0.
2851    
2852        (b) Recognize characters that are not in the table as "Cn" (undefined).
2853    
2854        (c) I revised the way the table is implemented to a much improved format
2855            which includes recognition of ranges. It now supports the ranges that
2856            are defined in UnicodeData.txt, and it also amalgamates other
2857            characters into ranges. This has reduced the number of entries in the
2858            table from around 16,000 to around 3,000, thus reducing its size
2859            considerably. I realized I did not need to use a tree structure after
2860            all - a binary chop search is just as efficient. Having reduced the
2861            number of entries, I extended their size from 6 bytes to 8 bytes to
2862            allow for more data.
2863    
2864        (d) Added support for Unicode script names via properties such as \p{Han}.
2865    
2866    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2867        matching that character.
2868    
2869    20. When matching a repeated Unicode property with a minimum greater than zero,
2870        (for example \pL{2,}), PCRE could look past the end of the subject if it
2871        reached it while seeking the minimum number of characters. This could
2872        happen only if some of the characters were more than one byte long, because
2873        there is a check for at least the minimum number of bytes.
2874    
2875    21. Refactored the implementation of \p and \P so as to be more general, to
2876        allow for more different types of property in future. This has changed the
2877        compiled form incompatibly. Anybody with saved compiled patterns that use
2878        \p or \P will have to recompile them.
2879    
2880    22. Added "Any" and "L&" to the supported property types.
2881    
2882    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2883        but give a compile time error if the value is greater than 0xff.
2884    
2885    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2886        accidentally not being installed or uninstalled.
2887    
2888    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2889        made were to insert the current release number. This seemed silly, because
2890        it made things harder for people building PCRE on systems that don't run
2891        "configure". I have turned pcre.h into a distributed file, no longer built
2892        by "configure", with the version identification directly included. There is
2893        no longer a pcre.h.in file.
2894    
2895        However, this change necessitated a change to the pcre-config script as
2896        well. It is built from pcre-config.in, and one of the substitutions was the
2897        release number. I have updated configure.ac so that ./configure now finds
2898        the release number by grepping pcre.h.
2899    
2900    26. Added the ability to run the tests under valgrind.
2901    
2902    
2903    Version 6.4 05-Sep-05
2904    ---------------------
2905    
2906     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2907        "--" to be printed when multiple files were scanned, even when none of the
2908        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2909        consider it to be a bug, and have restored the previous behaviour.
2910    
2911     2. A couple of code tidies to get rid of compiler warnings.
2912    
2913     3. The pcretest program used to cheat by referring to symbols in the library
2914        whose names begin with _pcre_. These are internal symbols that are not
2915        really supposed to be visible externally, and in some environments it is
2916        possible to suppress them. The cheating is now confined to including
2917        certain files from the library's source, which is a bit cleaner.
2918    
2919     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2920        file's purpose clearer.
2921    
2922     5. Reorganized pcre_ucp_findchar().
2923    
2924    
2925    Version 6.3 15-Aug-05
2926    ---------------------
2927    
2928     1. The file libpcre.pc.in did not have general read permission in the tarball.
2929    
2930     2. There were some problems when building without C++ support:
2931    
2932        (a) If C++ support was not built, "make install" and "make test" still
2933            tried to test it.
2934    
2935        (b) There were problems when the value of CXX was explicitly set. Some
2936            changes have been made to try to fix these, and ...
2937    
2938        (c) --disable-cpp can now be used to explicitly disable C++ support.
2939    
2940        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2941            backslash in a target when C++ was disabled. This confuses some
2942            versions of "make", apparently. Using an intermediate variable solves
2943            this. (Same for CPP_LOBJ.)
2944    
2945     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2946        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2947        necessary on certain architectures.
2948    
2949     4. Added a setting of -export-symbols-regex to the link command to remove
2950        those symbols that are exported in the C sense, but actually are local
2951        within the library, and not documented. Their names all begin with
2952        "_pcre_". This is not a perfect job, because (a) we have to except some
2953        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2954        available (and never for static libraries). I have made a note to try to
2955        find a way round (a) in the future.
2956    
2957    
2958    Version 6.2 01-Aug-05
2959    ---------------------
2960    
2961     1. There was no test for integer overflow of quantifier values. A construction
2962        such as {1111111111111111} would give undefined results. What is worse, if
2963        a minimum quantifier for a parenthesized subpattern overflowed and became
2964        negative, the calculation of the memory size went wrong. This could have
2965        led to memory overwriting.
2966    
2967     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2968    
2969     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2970        operating environments where this matters.
2971    
2972     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2973        PCRE options from within the C++ wrapper.
2974    
2975     5. Named capturing subpatterns were not being correctly counted when a pattern
2976        was compiled. This caused two problems: (a) If there were more than 100
2977        such subpatterns, the calculation of the memory needed for the whole
2978        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2979        back references of the form \12, where the number was greater than 9, were
2980        not recognized as back references, even though there were sufficient
2981        previous subpatterns.
2982    
2983     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2984        versions of gcc, e.g. 2.95.4.
2985    
2986    
2987    Version 6.1 21-Jun-05
2988    ---------------------
2989    
2990     1. There was one reference to the variable "posix" in pcretest.c that was not
2991        surrounded by "#if !defined NOPOSIX".
2992    
2993     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2994        the cross-check on the old pcre_info() function, for the benefit of the
2995        cut-down version of PCRE that is currently imported into Exim.
2996    
2997     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2998        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2999        patterns that start (?i) but I don't think that matters. The use of (?i) is
3000        just an example; this all applies to the other options as well.
3001    
3002     4. Since libtool seems to echo the compile commands it is issuing, the output
3003        from "make" can be reduced a bit by putting "@" in front of each libtool
3004        compile command.
3005    
3006     5. Patch from the folks at Google for configure.in to be a bit more thorough
3007        in checking for a suitable C++ installation before trying to compile the
3008        C++ stuff. This should fix a reported problem when a compiler was present,
3009        but no suitable headers.
3010    
3011     6. The man pages all had just "PCRE" as their title. I have changed them to
3012        be the relevant file name. I have also arranged that these names are
3013        retained in the file doc/pcre.txt, which is a concatenation in text format
3014        of all the man pages except the little individual ones for each function.
3015    
3016     7. The NON-UNIX-USE file had not been updated for the different set of source
3017        files that come with release 6. I also added a few comments about the C++
3018        wrapper.
3019    
3020    
3021    Version 6.0 07-Jun-05
3022    ---------------------
3023    
3024     1. Some minor internal re-organization to help with my DFA experiments.
3025    
3026     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3027        didn't matter for the library itself when fully configured, but did matter
3028        when compiling without UCP support, or within Exim, where the ucp files are
3029        not imported.
3030    
3031     3. Refactoring of the library code to split up the various functions into
3032        different source modules. The addition of the new DFA matching code (see
3033        below) to a single monolithic source would have made it really too
3034        unwieldy, quite apart from causing all the code to be include in a
3035        statically linked application, when only some functions are used. This is
3036        relevant even without the DFA addition now that patterns can be compiled in
3037        one application and matched in another.
3038    
3039        The downside of splitting up is that there have to be some external
3040        functions and data tables that are used internally in different modules of
3041        the library but which are not part of the API. These have all had their
3042        names changed to start with "_pcre_" so that they are unlikely to clash
3043        with other external names.
3044    
3045     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3046        a different (DFA) algorithm. Although it is slower than the original
3047        function, it does have some advantages for certain types of matching
3048        problem.
3049    
3050     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3051        including restarting after a partial match.
3052    
3053     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3054        defined when compiling for Windows was sent to me. I have put it into the
3055        code, though I have no means of testing or verifying it.
3056    
3057     7. Added the pcre_refcount() auxiliary function.
3058    
3059     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3060        match before or at the first newline in the subject string. In pcretest,
3061        the /f option on a pattern can be used to set this.
3062    
3063     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3064        would behave wrongly. This has been present in PCRE since release 4.0.
3065    
3066    10. A number of changes to the pcregrep command:
3067    
3068        (a) Refactored how -x works; insert ^(...)$ instead of setting
3069            PCRE_ANCHORED and checking the length, in preparation for adding
3070            something similar for -w.
3071    
3072        (b) Added the -w (match as a word) option.
3073    
3074        (c) Refactored the way lines are read and buffered so as to have more
3075            than one at a time available.
3076    
3077        (d) Implemented a pcregrep test script.
3078    
3079        (e) Added the -M (multiline match) option. This allows patterns to match
3080            over several lines of the subject. The buffering ensures that at least
3081            8K, or the rest of the document (whichever is the shorter) is available
3082            for matching (and similarly the previous 8K for lookbehind assertions).
3083    
3084        (f) Changed the --help output so that it now says
3085    
3086              -w, --word-regex(p)
3087    
3088            instead of two lines, one with "regex" and the other with "regexp"
3089            because that confused at least one person since the short forms are the
3090            same. (This required a bit of code, as the output is generated
3091            automatically from a table. It wasn't just a text change.)
3092    
3093        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3094            option but starts with a hyphen. Could be a pattern or a path name
3095            starting with a hyphen, for instance.
3096    
3097        (h) "-" can be given as a file name to represent stdin.
3098    
3099        (i) When file names are being printed, "(standard input)" is used for
3100            the standard input, for compatibility with GNU grep. Previously
3101            "<stdin>" was used.
3102    
3103        (j) The option --label=xxx can be used to supply a name to be used for
3104            stdin when file names are being printed. There is no short form.
3105    
3106        (k) Re-factored the options decoding logic because we are going to add
3107            two more options that take data. Such options can now be given in four
3108            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3109    
3110        (l) Added the -A, -B, and -C options for requesting that lines of context
3111            around matches be printed.
3112    
3113        (m) Added the -L option to print the names of files that do not contain
3114            any matching lines, that is, the complement of -l.
3115    
3116        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3117            continue to scan other files.
3118    
3119        (o) The -s option was incorrectly implemented. For compatibility with other
3120            greps, it now suppresses the error message for a non-existent or non-
3121            accessible file (but not the return code). There is a new option called
3122            -q that suppresses the output of matching lines, which was what -s was
3123            previously doing.
3124    
3125        (p) Added --include and --exclude options to specify files for inclusion
3126            and exclusion when recursing.
3127    
3128    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3129        Hopefully, it now does.
3130    
3131    12. Missing cast in pcre_study().
3132    
3133    13. Added an "uninstall" target to the makefile.
3134    
3135    14. Replaced "extern" in the function prototypes in Makefile.in with
3136        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3137        world, but is set differently for Windows.
3138    
3139    15. Added a second compiling function called pcre_compile2(). The only
3140        difference is that it has an extra argument, which is a pointer to an
3141        integer error code. When there is a compile-time failure, this is set
3142        non-zero, in addition to the error test pointer being set to point to an
3143        error message. The new argument may be NULL if no error number is required
3144        (but then you may as well call pcre_compile(), which is now just a
3145        wrapper). This facility is provided because some applications need a
3146        numeric error indication, but it has also enabled me to tidy up the way
3147        compile-time errors are handled in the POSIX wrapper.
3148    
3149    16. Added VPATH=.libs to the makefile; this should help when building with one
3150        prefix path and installing with another. (Or so I'm told by someone who
3151        knows more about this stuff than I do.)
3152    
3153    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3154        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3155        match everything, including newlines. This is not POSIX-compatible, but
3156        somebody wanted the feature. From pcretest it can be activated by using
3157        both the P and the s flags.
3158    
3159    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3160    
3161    19. libpcre.pc was being incorrectly installed as executable.
3162    
3163    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3164        it now also looks for '\r' so that it will work unmodified on Windows.
3165    
3166    21. Added Google's contributed C++ wrapper to the distribution.
3167    
3168    22. Added some untidy missing memory free() calls in pcretest, to keep
3169        Electric Fence happy when testing.
3170    
3171    
3172    
3173    Version 5.0 13-Sep-04
3174    ---------------------
3175    
3176     1. Internal change: literal characters are no longer packed up into items
3177        containing multiple characters in a single byte-string. Each character
3178        is now matched using a separate opcode. However, there may be more than one
3179        byte in the character in UTF-8 mode.
3180    
3181     2. The pcre_callout_block structure has two new fields: pattern_position and
3182        next_item_length. These contain the offset in the pattern to the next match
3183        item, and its length, respectively.
3184    
3185     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3186        insertion of callouts before each pattern item. Added the /C option to
3187        pcretest to make use of this.
3188    
3189     4. On the advice of a Windows user, the lines
3190    
3191          #if defined(_WIN32) || defined(WIN32)
3192          _setmode( _fileno( stdout ), 0x8000 );
3193          #endif  /* defined(_WIN32) || defined(WIN32) */
3194    
3195        have been added to the source of pcretest. This apparently does useful
3196        magic in relation to line terminators.
3197    
3198     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3199        for the benefit of those environments where the "b" makes a difference.
3200    
3201     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3202        to know about it. I have put a hack into configure.in that adds in code
3203        to set GCC=yes if CC=icc. This seems to end up at a point in the
3204        generated configure script that is early enough to affect the setting of
3205        compiler options, which is what is needed, but I have no means of testing
3206        whether it really works. (The user who reported this had patched the
3207        generated configure script, which of course I cannot do.)
3208    
3209        LATER: After change 22 below (new libtool files), the configure script
3210        seems to know about icc (and also ecc). Therefore, I have commented out
3211        this hack in configure.in.
3212    
3213     7. Added support for pkg-config (2 patches were sent in).
3214    
3215     8. Negated POSIX character classes that used a combination of internal tables
3216        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3217        [[:^ascii]]. Typically, they would match almost any characters. The other
3218        POSIX classes were not broken in this way.
3219    
3220     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3221        to find the match, as PCRE was deluded into thinking that the match had to
3222        start at the start point or following a newline. The same bug applied to
3223        patterns with negative forward assertions or any backward assertions
3224        preceding ".*" at the start, unless the pattern required a fixed first
3225        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3226    
3227    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3228        starting at the last subject character, bytes beyond the end of the subject
3229        string were read.
3230    
3231    11. Renamed the variable "class" as "classbits" to make life easier for C++
3232        users. (Previously there was a macro definition, but it apparently wasn't
3233        enough.)
3234    
3235    12. Added the new field "tables" to the extra data so that tables can be passed
3236        in at exec time, or the internal tables can be re-selected. This allows
3237        a compiled regex to be saved and re-used at a later time by a different
3238        program that might have everything at different addresses.
3239    
3240    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3241        -R library as well as a -L library.
3242    
3243    14. The debugging options of pcretest (-d on the command line or D on a
3244        pattern) showed incorrect output for anything following an extended class
3245        that contained multibyte characters and which was followed by a quantifier.
3246    
3247    15. Added optional support for general category Unicode character properties
3248        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3249        support. It adds about 90K to the size of the library. The meanings of the
3250        inbuilt class escapes such as \d and \s have NOT been changed.
3251    
3252    16. Updated pcredemo.c to include calls to free() to release the memory for the
3253        compiled pattern.
3254    
3255    17. The generated file chartables.c was being created in the source directory
3256        instead of in the building directory. This caused the build to fail if the
3257        source directory was different from the building directory, and was
3258        read-only.
3259    
3260    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3261        file. No doubt somebody will tell me if they don't make sense... Also added
3262        Dan Mooney's comments about building on OpenVMS.
3263    
3264    19. Added support for partial matching via the PCRE_PARTIAL option for
3265        pcre_exec() and the \P data escape in pcretest.
3266    
3267    20. Extended pcretest with 3 new pattern features:
3268    
3269        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3270              write the compiled pattern to the file whose name is "rest-of-line".
3271              This is a straight binary dump of the data, with the saved pointer to
3272              the character tables forced to be NULL. The study data, if any, is
3273              written too. After writing, pcretest reads a new pattern.
3274    
3275        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3276              compiled pattern from the given file. There must not be any
3277              occurrences of "<" in the file name (pretty unlikely); if there are,
3278              pcretest will instead treat the initial "<" as a pattern delimiter.
3279              After reading in the pattern, pcretest goes on to read data lines as
3280              usual.
3281    
3282        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3283              and 16-bit fields in a compiled pattern, to simulate a pattern that
3284              was compiled on a host of opposite endianness.
3285    
3286    21. The pcre-exec() function can now cope with patterns that were compiled on
3287        hosts of opposite endianness, with this restriction:
3288    
3289          As for any compiled expression that is saved and used later, the tables
3290          pointer field cannot be preserved; the extra_data field in the arguments
3291          to pcre_exec() should be used to pass in a tables address if a value
3292          other than the default internal tables were used at compile time.
3293    
3294    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3295        now diagnosed as an error. Previously, most of the time, a negative number
3296        would have been treated as zero, but if in addition "ovector" was passed as
3297        NULL, a crash could occur.
3298    
3299    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3300        new versions from the libtool 1.5 distribution (the last one is a copy of
3301        a file called libtool.m4). This seems to have fixed the need to patch
3302        "configure" to support Darwin 1.3 (which I used to do). However, I still
3303        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3304        workstation).
3305    
3306    24. Changed the PCRE licence to be the more standard "BSD" licence.
3307    
3308    
3309    Version 4.5 01-Dec-03
3310    ---------------------
3311    
3312     1. There has been some re-arrangement of the code for the match() function so
3313        that it can be compiled in a version that does not call itself recursively.
3314        Instead, it keeps those local variables that need separate instances for
3315        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3316        needs to "recurse". Keeping track of where control must go is done by means
3317        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3318        hide most of the details from the main code, and operates only if
3319        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3320        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3321        operating.
3322    
3323        To make it easier for callers to provide specially tailored get/free
3324        functions for this usage, two new functions, pcre_stack_malloc, and
3325        pcre_stack_free, are used. They are always called in strict stacking order,
3326        and the size of block requested is always the same.
3327    
3328        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3329        PCRE has been compiled to use the stack or the heap for recursion. The
3330        -C option of pcretest uses this to show which version is compiled.
3331    
3332        A new data escape \S, is added to pcretest; it causes the amounts of store
3333        obtained and freed by both kinds of malloc/free at match time to be added
3334        to the output.
3335    
3336     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3337        what's available on my current Linux desktop machine.
3338    
3339     3. When matching a UTF-8 string, the test for a valid string at the start has
3340        been extended. If start_offset is not zero, PCRE now checks that it points
3341        to a byte that is the start of a UTF-8 character. If not, it returns
3342        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3343        this is necessary because there may be backward assertions in the pattern.
3344        When matching the same subject several times, it may save resources to use
3345        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3346    
3347     4. The code for checking the validity of UTF-8 strings has been tightened so
3348        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3349        containing "overlong sequences".
3350    
3351     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3352        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3353        should have been "&", but it just so happened that all the cases this let
3354        through by mistake were picked up later in the function.
3355    
3356     6. I had used a variable called "isblank" - this is a C99 function, causing
3357        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3358    
3359     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3360        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3361        so the version is shown; (c) stop "make test" echoing "./RunTest".
3362    
3363     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3364    
3365     9. The prototype for memmove() for systems that don't have it was using
3366        size_t, but the inclusion of the header that defines size_t was later. I've
3367        moved the #includes for the C headers earlier to avoid this.
3368    
3369    10. Added some adjustments to the code to make it easier to compiler on certain
3370        special systems:
3371    
3372          (a) Some "const" qualifiers were missing.
3373          (b) Added the macro EXPORT before all exported functions; by default this
3374              is defined to be empty.
3375          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3376              that it reads its output file name as an argument instead of writing
3377              to the standard output and assuming this can be redirected.
3378    
3379    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3380        class containing characters with values greater than 255, PCRE compilation
3381        went into a loop.
3382    
3383    12. A recursive reference to a subpattern that was within another subpattern
3384        that had a minimum quantifier of zero caused PCRE to crash. For example,
3385        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3386        recursion. If the recursively-called subpattern itself had a zero repeat,
3387        that was OK.
3388    
3389    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3390        buffer into which it was copied (for escape processing) was still set at
3391        1024, so long lines caused crashes.
3392    
3393    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3394        "internal error: code overflow...". This applied to any character class
3395        that was followed by a possessive quantifier.
3396    
3397    15. Modified the Makefile to add libpcre.la as a prerequisite for
3398        libpcreposix.la because I was told this is needed for a parallel build to
3399        work.
3400    
3401    16. If a pattern that contained .* following optional items at the start was
3402        studied, the wrong optimizing data was generated, leading to matching
3403        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3404        matching string must start with a or b or c. The correct conclusion for
3405        this pattern is that a match can start with any character.
3406    
3407    
3408    Version 4.4 13-Aug-03
3409    ---------------------
3410    
3411     1. In UTF-8 mode, a character class containing characters with values between
3412        127 and 255 was not handled correctly if the compiled pattern was studied.
3413        In fixing this, I have also improved the studying algorithm for such
3414        classes (slightly).
3415    
3416     2. Three internal functions had redundant arguments passed to them. Removal
3417        might give a very teeny performance improvement.
3418    
3419     3. Documentation bug: the value of the capture_top field in a callout is *one
3420        more than* the number of the hightest numbered captured substring.
3421    
3422     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3423        in incorrectly linking with a previously installed version. They now link
3424        explicitly with libpcre.la.
3425    
3426     5. configure.in no longer needs to recognize Cygwin specially.
3427    
3428     6. A problem in pcre.in for Windows platforms is fixed.
3429    
3430     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3431        pcretest, it used to include the size of the study block as part of its
3432        output. Unfortunately, the structure contains a field that has a different
3433        size on different hardware architectures. This meant that the tests that
3434        showed this size failed. As the block is currently always of a fixed size,
3435        this information isn't actually particularly useful in pcretest output, so
3436        I have just removed it.
3437    
3438     8. Three pre-processor statements accidentally did not start in column 1.
3439        Sadly, there are *still* compilers around that complain, even though
3440        standard C has not required this for well over a decade. Sigh.
3441    
3442     9. In pcretest, the code for checking callouts passed small integers in the
3443        callout_data field, which is a void * field. However, some picky compilers
3444        complained about the casts involved for this on 64-bit systems. Now
3445        pcretest passes the address of the small integer instead, which should get
3446        rid of the warnings.
3447    
3448    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3449        both compile and run time, and gives an error if an invalid UTF-8 sequence
3450        is found. There is a option for disabling this check in cases where the
3451        string is known to be correct and/or the maximum performance is wanted.
3452    
3453    11. In response to a bug report, I changed one line in Makefile.in from
3454    
3455            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3456        to
3457            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3458    
3459        to look similar to other lines, but I have no way of telling whether this
3460        is the right thing to do, as I do not use Windows. No doubt I'll get told
3461        if it's wrong...
3462    
3463    
3464    Version 4.3 21-May-03
3465    ---------------------
3466    
3467    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3468       Makefile.
3469    
3470    2. Some refactoring to improve the quality of the code:
3471    
3472       (i)   The utf8_table... variables are now declared "const".
3473    
3474       (ii)  The code for \cx, which used the "case flipping" table to upper case
3475             lower case letters, now just substracts 32. This is ASCII-specific,
3476             but the whole concept of \cx is ASCII-specific, so it seems
3477             reasonable.
3478    
3479       (iii) PCRE was using its character types table to recognize decimal and
3480             hexadecimal digits in the pattern. This is silly, because it handles
3481             only 0-9, a-f, and A-F, but the character types table is locale-
3482             specific, which means strange things might happen. A private
3483             table is now used for this - though it costs 256 bytes, a table is
3484             much faster than multiple explicit tests. Of course, the standard
3485             character types table is still used for matching digits in subject
3486             strings against \d.
3487    
3488       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3489             ending in _t are). So I've renamed it as ESC_tee.
3490    
3491    3. The first argument for regexec() in the POSIX wrapper should have been
3492       defined as "const".
3493    
3494    4. Changed pcretest to use malloc() for its buffers so that they can be
3495       Electric Fenced for debugging.
3496    
3497    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3498       to read one or more bytes before the start of the subject string. Often this
3499       had no effect on PCRE's behaviour, but in some circumstances it could
3500       provoke a segmentation fault.
3501    
3502    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3503       to try to read one or more bytes before the start of the subject string.
3504    
3505    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3506       UTF-8 support could misbehave in various ways if the subject string
3507       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3508       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3509       back over UTF-8 characters.)
3510    
3511    
3512    Version 4.2 14-Apr-03
3513    ---------------------
3514    
3515    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3516    
3517    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3518         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3519         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3520         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3521         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3522           and BUILD_EXEEXT
3523         Note: automatic setting of the BUILD variables is not yet working
3524         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3525           compile-time but not at link-time
3526         [LINK]: use for linking executables only