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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.81  
changed lines
  Added in v.958

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12