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

Diff of /code/trunk/ChangeLog

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

revision 29 by nigel, Sat Feb 24 21:38:53 2007 UTC revision 960 by ph10, Thu Apr 19 17:30:38 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    9.  (*MARK) control verb is now supported by the JIT compiler.
46    
47    10. The command "./RunTest list" lists the available tests without actually
48        running any of them. (Because I keep forgetting what they all are.)
49    
50    11. Add PCRE_INFO_MAXLOOKBEHIND.
51    
52    12. Applied a (slightly modified) user-supplied patch that improves performance
53        when the heap is used for recursion (compiled with --disable-stack-for-
54        recursion). Instead of malloc and free for each heap frame each time a
55        logical recursion happens, frames are retained on a chain and re-used where
56        possible. This sometimes gives as much as 30% improvement.
57    
58    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59        call.
60    
61    14. As documented, (*COMMIT) is now confined to within a positive assertion.
62    
63    15. It is now possible to link pcretest with libedit as an alternative to
64        libreadline.
65    
66    16. (*COMMIT) control verb is now supported by the JIT compiler.
67    
68    17. The Unicode data tables have been updated to Unicode 6.1.0.
69    
70    18. Added --file-list option to pcregrep.
71    
72    19. Added binary file support to pcregrep, including the -a, --binary-files,
73        -I, and --text options.
74    
75    20. The madvise function is renamed for posix_madvise for QNX compatibility
76        reasons. Fixed by Giuseppe D'Angelo.
77    
78    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79    
80    22. Changed the option for creating symbolic links for 16-bit man pages from
81        -s to -sf so that re-installing does not cause issues.
82    
83    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84    
85    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86        matches in certain environments (the workspace was not being correctly
87        retained). Also added to pcre_dfa_exec() a simple plausibility check on
88        some of the workspace data at the beginning of a restart.
89    
90    
91    Version 8.30 04-February-2012
92    -----------------------------
93    
94    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
95        name is defined in ctype.h.
96    
97    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
98        only in quite long subpatterns.
99    
100    3.  Removed the function pcre_info(), which has been obsolete and deprecated
101        since it was replaced by pcre_fullinfo() in February 2000.
102    
103    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
104        match a (*MARK), and the match failed at the start of the subject, a
105        reference to memory before the start of the subject could occur. This bug
106        was introduced by fix 17 of release 8.21.
107    
108    5.  A reference to an unset group with zero minimum repetition was giving
109        totally wrong answers (in non-JavaScript-compatibility mode). For example,
110        /(another)?(\1?)test/ matched against "hello world test". This bug was
111        introduced in release 8.13.
112    
113    6.  Add support for 16-bit character strings (a large amount of work involving
114        many changes and refactorings).
115    
116    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
117        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
118        from a file.
119    
120    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
121        rounding is not applied in this particular case).
122    
123    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
124        if they appear, or are escaped, in patterns.
125    
126    10. Get rid of a number of -Wunused-but-set-variable warnings.
127    
128    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
129        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
130        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
131        also returns the mark "x". This bug applied to capturing parentheses,
132        non-capturing parentheses, and atomic parentheses. It also applied to some
133        assertions.
134    
135    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
136        information out of configure.ac instead of relying on pcre.h.generic, which
137        is not stored in the repository.
138    
139    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
140        -lreadline.
141    
142    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
143    
144    15. Applied Graycode's patch to put the top-level frame on the stack rather
145        than the heap when not using the stack for recursion. This gives a
146        performance improvement in many cases when recursion is not deep.
147    
148    16. Experimental code added to "pcretest -C" to output the stack frame size.
149    
150    
151    Version 8.21 12-Dec-2011
152    ------------------------
153    
154    1.  Updating the JIT compiler.
155    
156    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
157        are added as well.
158    
159    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
160        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
161        calling _pcre_jit_exec. Some extra comments are added.
162    
163    4.  (*MARK) settings inside atomic groups that do not contain any capturing
164        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
165        was introduced by change 18 for 8.20.
166    
167    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
168        ECMA-262 standard.
169    
170    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
171        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
172        This bug was probably introduced by change 9 of 8.13.
173    
174    7.  While fixing 6 above, I noticed that a number of other items were being
175        incorrectly rejected as "not fixed length". This arose partly because newer
176        opcodes had not been added to the fixed-length checking code. I have (a)
177        corrected the bug and added tests for these items, and (b) arranged for an
178        error to occur if an unknown opcode is encountered while checking for fixed
179        length instead of just assuming "not fixed length". The items that were
180        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
181        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
182        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
183    
184    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
185        being incorrectly compiled and would have given unpredicatble results.
186    
187    9.  A possessively repeated subpattern with minimum repeat count greater than
188        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
189        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
190        the first (A) could occur when it should not.
191    
192    10. Add a cast and remove a redundant test from the code.
193    
194    11. JIT should use pcre_malloc/pcre_free for allocation.
195    
196    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
197        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
198        is anything other than /usr, -L is still shown).
199    
200    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
201    
202    14. Perl does not support \N without a following name in a [] class; PCRE now
203        also gives an error.
204    
205    15. If a forward reference was repeated with an upper limit of around 2000,
206        it caused the error "internal error: overran compiling workspace". The
207        maximum number of forward references (including repeats) was limited by the
208        internal workspace, and dependent on the LINK_SIZE. The code has been
209        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
210        the default depends on LINK_SIZE. There is a new upper limit (for safety)
211        of around 200,000 forward references. While doing this, I also speeded up
212        the filling in of repeated forward references.
213    
214    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
215        incorrectly expecting the subject to contain another "a" after the start.
216    
217    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
218        in the match, the SKIP should be ignored. This was not happening; instead
219        the SKIP was being treated as NOMATCH. For patterns such as
220        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
221        tested.
222    
223    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
224        now much more compatible with Perl, in particular in cases where the result
225        is a non-match for a non-anchored pattern. For example, if
226        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
227        "m", where previously it did not return a name. A side effect of this
228        change is that for partial matches, the last encountered mark name is
229        returned, as for non matches. A number of tests that were previously not
230        Perl-compatible have been moved into the Perl-compatible test files. The
231        refactoring has had the pleasing side effect of removing one argument from
232        the match() function, thus reducing its stack requirements.
233    
234    19. If the /S+ option was used in pcretest to study a pattern using JIT,
235        subsequent uses of /S (without +) incorrectly behaved like /S+.
236    
237    21. Retrieve executable code size support for the JIT compiler and fixing
238        some warnings.
239    
240    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
241        not work when the shorter character appeared right at the end of the
242        subject string.
243    
244    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
245        systems.
246    
247    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
248        output it when the /M option is used in pcretest.
249    
250    25. The CheckMan script was not being included in the distribution. Also, added
251        an explicit "perl" to run Perl scripts from the PrepareRelease script
252        because this is reportedly needed in Windows.
253    
254    26. If study data was being save in a file and studying had not found a set of
255        "starts with" bytes for the pattern, the data written to the file (though
256        never used) was taken from uninitialized memory and so caused valgrind to
257        complain.
258    
259    27. Updated RunTest.bat as provided by Sheri Pierce.
260    
261    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
262    
263    29. Computation of memory usage for the table of capturing group names was
264        giving an unnecessarily large value.
265    
266    
267    Version 8.20 21-Oct-2011
268    ------------------------
269    
270    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
271        a POSIX class. After further experiments with Perl, which convinced me that
272        Perl has bugs and confusions, a closing square bracket is no longer allowed
273        in a POSIX name. This bug also affected patterns with classes that started
274        with full stops.
275    
276    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
277        captured substring, but while checking the failing first alternative,
278        substring 1 is temporarily captured. If the output vector supplied to
279        pcre_exec() was not big enough for this capture, the yield of the function
280        was still zero ("insufficient space for captured substrings"). This cannot
281        be totally fixed without adding another stack variable, which seems a lot
282        of expense for a edge case. However, I have improved the situation in cases
283        such as /(a)(b)x|abc/ matched against "abc", where the return code
284        indicates that fewer than the maximum number of slots in the ovector have
285        been set.
286    
287    3.  Related to (2) above: when there are more back references in a pattern than
288        slots in the output vector, pcre_exec() uses temporary memory during
289        matching, and copies in the captures as far as possible afterwards. It was
290        using the entire output vector, but this conflicts with the specification
291        that only 2/3 is used for passing back captured substrings. Now it uses
292        only the first 2/3, for compatibility. This is, of course, another edge
293        case.
294    
295    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
296        main code base, and can be used by building with --enable-jit. When this is
297        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
298        runtime --no-jit option is given.
299    
300    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
301        ovector, the return from the function was zero, implying that there were
302        other matches that did not fit. The correct "exactly full" value is now
303        returned.
304    
305    6.  If a subpattern that was called recursively or as a subroutine contained
306        (*PRUNE) or any other control that caused it to give a non-standard return,
307        invalid errors such as "Error -26 (nested recursion at the same subject
308        position)" or even infinite loops could occur.
309    
310    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
311        computing the minimum length on reaching *ACCEPT, and so ended up with the
312        wrong value of 1 rather than 0. Further investigation indicates that
313        computing a minimum subject length in the presence of *ACCEPT is difficult
314        (think back references, subroutine calls), and so I have changed the code
315        so that no minimum is registered for a pattern that contains *ACCEPT.
316    
317    8.  If (*THEN) was present in the first (true) branch of a conditional group,
318        it was not handled as intended. [But see 16 below.]
319    
320    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
321        Sheri Pierce.
322    
323    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
324        the first byte in a match must be "a".
325    
326    11. Change 17 for 8.13 increased the recursion depth for patterns like
327        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
328        pattern contains any instances of (*THEN). If it does not, the old
329        optimizations are restored. It would be nice to do this on a per-group
330        basis, but at the moment that is not feasible.
331    
332    12. In some environments, the output of pcretest -C is CRLF terminated. This
333        broke RunTest's code that checks for the link size. A single white space
334        character after the value is now allowed for.
335    
336    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
337        For "fr", it uses the Windows-specific input and output files.
338    
339    14. If (*THEN) appeared in a group that was called recursively or as a
340        subroutine, it did not work as intended. [But see next item.]
341    
342    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
343        pattern fragments (but not containing any | characters). If A and B are
344        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
345        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
346        D. In other words, Perl considers parentheses that do not contain any |
347        characters to be part of a surrounding alternative, whereas PCRE was
348        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
349        differently. PCRE now behaves in the same way as Perl, except in the case
350        of subroutine/recursion calls such as (?1) which have in any case always
351        been different (but PCRE had them first :-).
352    
353    16. Related to 15 above: Perl does not treat the | in a conditional group as
354        creating alternatives. Such a group is treated in the same way as an
355        ordinary group without any | characters when processing (*THEN). PCRE has
356        been changed to match Perl's behaviour.
357    
358    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
359        RunGrepTest script failed.
360    
361    18. Change 22 for version 13 caused atomic groups to use more stack. This is
362        inevitable for groups that contain captures, but it can lead to a lot of
363        stack use in large patterns. The old behaviour has been restored for atomic
364        groups that do not contain any capturing parentheses.
365    
366    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
367        suppress the check for a minimum subject length at run time. (If it was
368        given to pcre_exec() or pcre_dfa_exec() it did work.)
369    
370    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
371        fail to work when decoding hex characters in data strings in EBCDIC
372        environments.
373    
374    21. It appears that in at least one Mac OS environment, the isxdigit() function
375        is implemented as a macro that evaluates to its argument more than once,
376        contravening the C 90 Standard (I haven't checked a later standard). There
377        was an instance in pcretest which caused it to go wrong when processing
378        \x{...} escapes in subject strings. The has been rewritten to avoid using
379        things like p++ in the argument of isxdigit().
380    
381    
382    Version 8.13 16-Aug-2011
383    ------------------------
384    
385    1.  The Unicode data tables have been updated to Unicode 6.0.0.
386    
387    2.  Two minor typos in pcre_internal.h have been fixed.
388    
389    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
390        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
391        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
392    
393    4.  There were a number of related bugs in the code for matching backrefences
394        caselessly in UTF-8 mode when codes for the characters concerned were
395        different numbers of bytes. For example, U+023A and U+2C65 are an upper
396        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
397        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
398        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
399        2-byte code at the end of the subject (it thought there wasn't enough data
400        left).
401    
402    5.  Comprehensive information about what went wrong is now returned by
403        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
404        as the output vector has at least 2 elements. The offset of the start of
405        the failing character and a reason code are placed in the vector.
406    
407    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
408        now returned is for the first byte of the failing character, instead of the
409        last byte inspected. This is an incompatible change, but I hope it is small
410        enough not to be a problem. It makes the returned offset consistent with
411        pcre_exec() and pcre_dfa_exec().
412    
413    7.  pcretest now gives a text phrase as well as the error number when
414        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
415        failure, the offset and reason code are output.
416    
417    8.  When \R was used with a maximizing quantifier it failed to skip backwards
418        over a \r\n pair if the subsequent match failed. Instead, it just skipped
419        back over a single character (\n). This seems wrong (because it treated the
420        two characters as a single entity when going forwards), conflicts with the
421        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
422        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
423        has been changed.
424    
425    9.  Some internal refactoring has changed the processing so that the handling
426        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
427        time (the PCRE_DOTALL option was changed this way some time ago: version
428        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
429        which was always a bit of a fudge. It also means that there is one less
430        argument for the match() function, which reduces its stack requirements
431        slightly. This change also fixes an incompatibility with Perl: the pattern
432        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
433    
434    10. More internal refactoring has drastically reduced the number of recursive
435        calls to match() for possessively repeated groups such as (abc)++ when
436        using pcre_exec().
437    
438    11. While implementing 10, a number of bugs in the handling of groups were
439        discovered and fixed:
440    
441        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
442        (a|)*(?1) gave a compile-time internal error.
443        ((a|)+)+  did not notice that the outer group could match an empty string.
444        (^a|^)+   was not marked as anchored.
445        (.*a|.*)+ was not marked as matching at start or after a newline.
446    
447    12. Yet more internal refactoring has removed another argument from the match()
448        function. Special calls to this function are now indicated by setting a
449        value in a variable in the "match data" data block.
450    
451    13. Be more explicit in pcre_study() instead of relying on "default" for
452        opcodes that mean there is no starting character; this means that when new
453        ones are added and accidentally left out of pcre_study(), testing should
454        pick them up.
455    
456    14. The -s option of pcretest has been documented for ages as being an old
457        synonym of -m (show memory usage). I have changed it to mean "force study
458        for every regex", that is, assume /S for every regex. This is similar to -i
459        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
460        using it. It makes it easier to run collections of tests with and without
461        study enabled, and thereby test pcre_study() more easily. All the standard
462        tests are now run with and without -s (but some patterns can be marked as
463        "never study" - see 20 below).
464    
465    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
466        restoration of the capturing data to the outer values was not happening
467        correctly.
468    
469    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
470        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
471        pattern had matched an empty string, and so incorrectly returned a no
472        match.
473    
474    17. There was optimizing code for the last branch of non-capturing parentheses,
475        and also for the obeyed branch of a conditional subexpression, which used
476        tail recursion to cut down on stack usage. Unfortunately, now that there is
477        the possibility of (*THEN) occurring in these branches, tail recursion is
478        no longer possible because the return has to be checked for (*THEN). These
479        two optimizations have therefore been removed. [But see 8.20/11 above.]
480    
481    18. If a pattern containing \R was studied, it was assumed that \R always
482        matched two bytes, thus causing the minimum subject length to be
483        incorrectly computed because \R can also match just one byte.
484    
485    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
486        was incorrectly computed.
487    
488    20. If /S is present twice on a test pattern in pcretest input, it now
489        *disables* studying, thereby overriding the use of -s on the command line
490        (see 14 above). This is necessary for one or two tests to keep the output
491        identical in both cases.
492    
493    21. When (*ACCEPT) was used in an assertion that matched an empty string and
494        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
495    
496    22. When an atomic group that contained a capturing parenthesis was
497        successfully matched, but the branch in which it appeared failed, the
498        capturing was not being forgotten if a higher numbered group was later
499        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
500        group 1 to "a", when in fact it should be unset. This applied to multi-
501        branched capturing and non-capturing groups, repeated or not, and also to
502        positive assertions (capturing in negative assertions does not happen
503        in PCRE) and also to nested atomic groups.
504    
505    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
506        subject after a captured substring, to make it easier to tell which of a
507        number of identical substrings has been captured.
508    
509    24. The way atomic groups are processed by pcre_exec() has been changed so that
510        if they are repeated, backtracking one repetition now resets captured
511        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
512        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
513        "aaa". Previously, it would have been "a". As part of this code
514        refactoring, the way recursive calls are handled has also been changed.
515    
516    25. If an assertion condition captured any substrings, they were not passed
517        back unless some other capturing happened later. For example, if
518        (?(?=(a))a) was matched against "a", no capturing was returned.
519    
520    26. When studying a pattern that contained subroutine calls or assertions,
521        the code for finding the minimum length of a possible match was handling
522        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
523        group 1 called group 2 while simultaneously a separate group 2 called group
524        1). A stack overflow occurred in this case. I have fixed this by limiting
525        the recursion depth to 10.
526    
527    27. Updated RunTest.bat in the distribution to the version supplied by Tom
528        Fortmann. This supports explicit test numbers on the command line, and has
529        argument validation and error reporting.
530    
531    28. An instance of \X with an unlimited repeat could fail if at any point the
532        first character it looked at was a mark character.
533    
534    29. Some minor code refactoring concerning Unicode properties and scripts
535        should reduce the stack requirement of match() slightly.
536    
537    30. Added the '=' option to pcretest to check the setting of unused capturing
538        slots at the end of the pattern, which are documented as being -1, but are
539        not included in the return count.
540    
541    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
542        compiled something random. Now it gives a compile-time error (as does
543        Perl).
544    
545    32. A *MARK encountered during the processing of a positive assertion is now
546        recorded and passed back (compatible with Perl).
547    
548    33. If --only-matching or --colour was set on a pcregrep call whose pattern
549        had alternative anchored branches, the search for a second match in a line
550        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
551        matched the line "0102" twice. The same bug affected patterns that started
552        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
553        twice.
554    
555    34. Previously, PCRE did not allow quantification of assertions. However, Perl
556        does, and because of capturing effects, quantifying parenthesized
557        assertions may at times be useful. Quantifiers are now allowed for
558        parenthesized assertions.
559    
560    35. A minor code tidy in pcre_compile() when checking options for \R usage.
561    
562    36. \g was being checked for fancy things in a character class, when it should
563        just be a literal "g".
564    
565    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
566        appearance of a nested POSIX class supersedes an apparent external class.
567        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
568        unescaped square brackets may also appear as part of class names. For
569        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
570        more like Perl. (But see 8.20/1 above.)
571    
572    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
573        was because it thought it was \N{name}, which is not supported).
574    
575    39. Add minix to OS list not supporting the -S option in pcretest.
576    
577    40. PCRE tries to detect cases of infinite recursion at compile time, but it
578        cannot analyze patterns in sufficient detail to catch mutual recursions
579        such as ((?1))((?2)). There is now a runtime test that gives an error if a
580        subgroup is called recursively as a subpattern for a second time at the
581        same position in the subject string. In previous releases this might have
582        been caught by the recursion limit, or it might have run out of stack.
583    
584    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
585        happen only once. PCRE was, however incorrectly giving a compile time error
586        "recursive call could loop indefinitely" because it cannot analyze the
587        pattern in sufficient detail. The compile time test no longer happens when
588        PCRE is compiling a conditional subpattern, but actual runaway loops are
589        now caught at runtime (see 40 above).
590    
591    42. It seems that Perl allows any characters other than a closing parenthesis
592        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
593        has been changed to be the same.
594    
595    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
596        as not to get warnings when autogen.sh is called. Also changed
597        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
598    
599    44. To help people who use pcregrep to scan files containing exceedingly long
600        lines, the following changes have been made:
601    
602        (a) The default value of the buffer size parameter has been increased from
603            8K to 20K. (The actual buffer used is three times this size.)
604    
605        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
606            PCRE is built.
607    
608        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
609            to be set at run time.
610    
611        (d) Numerical values in pcregrep options can be followed by K or M, for
612            example --buffer-size=50K.
613    
614        (e) If a line being scanned overflows pcregrep's buffer, an error is now
615            given and the return code is set to 2.
616    
617    45. Add a pointer to the latest mark to the callout data block.
618    
619    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
620        partial match of an empty string instead of no match. This was specific to
621        the use of ".".
622    
623    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
624        complete match instead of a partial match. This bug was dependent on both
625        the PCRE_UTF8 and PCRE_DOTALL options being set.
626    
627    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
628        starting byte set, because \b was not being ignored.
629    
630    
631    Version 8.12 15-Jan-2011
632    ------------------------
633    
634    1.  Fixed some typos in the markup of the man pages, and wrote a script that
635        checks for such things as part of the documentation building process.
636    
637    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
638        --match-limit and --recursion-limit options (added for 8.11). In
639        particular, this made one of the standard tests fail. (The integer value
640        went into the wrong half of a long int.)
641    
642    3.  If the --colour option was given to pcregrep with -v (invert match), it
643        did strange things, either producing crazy output, or crashing. It should,
644        of course, ignore a request for colour when reporting lines that do not
645        match.
646    
647    4.  Another pcregrep bug caused similar problems if --colour was specified with
648        -M (multiline) and the pattern match finished with a line ending.
649    
650    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
651        matched in multiline mode, the following line was shown as part of the
652        match. This seems wrong, so I have changed it.
653    
654    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
655        the check for further matches in the same line (so they could be coloured)
656        to overrun the end of the current line. If another match was found, it was
657        incorrectly shown (and then shown again when found in the next line).
658    
659    7.  If pcregrep was compiled under Windows, there was a reference to the
660        function pcregrep_exit() before it was defined. I am assuming this was
661        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
662        reported by a user. I've moved the definition above the reference.
663    
664    
665    Version 8.11 10-Dec-2010
666    ------------------------
667    
668    1.  (*THEN) was not working properly if there were untried alternatives prior
669        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
670        backtracked to try for "b" instead of moving to the next alternative branch
671        at the same level (in this case, to look for "c"). The Perl documentation
672        is clear that when (*THEN) is backtracked onto, it goes to the "next
673        alternative in the innermost enclosing group".
674    
675    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
676        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
677        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
678        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
679        (*THEN).
680    
681    3.  If \s appeared in a character class, it removed the VT character from
682        the class, even if it had been included by some previous item, for example
683        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
684        of \s, but is part of the POSIX "space" class.)
685    
686    4.  A partial match never returns an empty string (because you can always
687        match an empty string at the end of the subject); however the checking for
688        an empty string was starting at the "start of match" point. This has been
689        changed to the "earliest inspected character" point, because the returned
690        data for a partial match starts at this character. This means that, for
691        example, /(?<=abc)def/ gives a partial match for the subject "abc"
692        (previously it gave "no match").
693    
694    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
695        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
696        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
697        has an implication that the given string is incomplete (because a partial
698        match is preferred over a full match). For this reason, these items now
699        give a partial match in this situation. [Aside: previously, the one case
700        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
701        match rather than a full match, which was wrong by the old rules, but is
702        now correct.]
703    
704    6.  There was a bug in the handling of #-introduced comments, recognized when
705        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
706        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
707        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
708        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
709        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
710        places in pcre_compile().
711    
712    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
713        comments when looking ahead for named forward references to subpatterns,
714        the only newline sequence it recognized was NL. It now handles newlines
715        according to the set newline convention.
716    
717    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
718        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
719        cater for a lack of strerror(). These oversights have been fixed.
720    
721    9.  Added --match-limit and --recursion-limit to pcregrep.
722    
723    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
724    
725    11. When the -o option was used, pcregrep was setting a return code of 1, even
726        when matches were found, and --line-buffered was not being honoured.
727    
728    12. Added an optional parentheses number to the -o and --only-matching options
729        of pcregrep.
730    
731    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
732        can match an empty string. The code to do it in pcretest and pcredemo
733        needed fixing:
734    
735        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
736            only one byte after an empty string match just before CRLF (this case
737            just got forgotten; "any" and "anycrlf" were OK).
738    
739        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
740            mode when an empty string match preceded an ASCII character followed by
741            a non-ASCII character. (The code for advancing by one character rather
742            than one byte was nonsense.)
743    
744        (c) The pcredemo.c sample program did not have any code at all to handle
745            the cases when CRLF is a valid newline sequence.
746    
747    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
748        as a starting offset was within the subject string. There is now a new
749        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
750        negative or greater than the length of the string. In order to test this,
751        pcretest is extended to allow the setting of negative starting offsets.
752    
753    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
754        starting offset points to the beginning of a UTF-8 character was
755        unnecessarily clumsy. I tidied it up.
756    
757    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
758        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
759    
760    17. Nobody had reported that the --include_dir option, which was added in
761        release 7.7 should have been called --include-dir (hyphen, not underscore)
762        for compatibility with GNU grep. I have changed it to --include-dir, but
763        left --include_dir as an undocumented synonym, and the same for
764        --exclude-dir, though that is not available in GNU grep, at least as of
765        release 2.5.4.
766    
767    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
768        characters from a string of bytes) have been redefined so as not to use
769        loops, in order to improve performance in some environments. At the same
770        time, I abstracted some of the common code into auxiliary macros to save
771        repetition (this should not affect the compiled code).
772    
773    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
774        compile-time error is now given if \c is not followed by an ASCII
775        character, that is, a byte less than 128. (In EBCDIC mode, the code is
776        different, and any byte value is allowed.)
777    
778    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
779        START_OPTIMIZE option, which is now allowed at compile time - but just
780        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
781        to pcregrep and other applications that have no direct access to PCRE
782        options. The new /Y option in pcretest sets this option when calling
783        pcre_compile().
784    
785    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
786        back references. Groups containing recursive back references were forced to
787        be atomic by that change, but in the case of named groups, the amount of
788        memory required was incorrectly computed, leading to "Failed: internal
789        error: code overflow". This has been fixed.
790    
791    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
792        pcretest.c, to avoid build problems in some Borland environments.
793    
794    
795    Version 8.10 25-Jun-2010
796    ------------------------
797    
798    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
799        THEN.
800    
801    2.  (*ACCEPT) was not working when inside an atomic group.
802    
803    3.  Inside a character class, \B is treated as a literal by default, but
804        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
805        causes the error). The code is unchanged, but I tidied the documentation.
806    
807    4.  Inside a character class, PCRE always treated \R and \X as literals,
808        whereas Perl faults them if its -w option is set. I have changed PCRE so
809        that it faults them when PCRE_EXTRA is set.
810    
811    5.  Added support for \N, which always matches any character other than
812        newline. (It is the same as "." when PCRE_DOTALL is not set.)
813    
814    6.  When compiling pcregrep with newer versions of gcc which may have
815        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
816        declared with attribute warn_unused_result" were given. Just casting the
817        result to (void) does not stop the warnings; a more elaborate fudge is
818        needed. I've used a macro to implement this.
819    
820    7.  Minor change to pcretest.c to avoid a compiler warning.
821    
822    8.  Added four artifical Unicode properties to help with an option to make
823        \s etc use properties (see next item). The new properties are: Xan
824        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
825    
826    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
827        use Unicode properties. (*UCP) at the start of a pattern can be used to set
828        this option. Modified pcretest to add /W to test this facility. Added
829        REG_UCP to make it available via the POSIX interface.
830    
831    10. Added --line-buffered to pcregrep.
832    
833    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
834        studied, and the match started with a letter with a code point greater than
835        127 whose first byte was different to the first byte of the other case of
836        the letter, the other case of this starting letter was not recognized
837        (#976).
838    
839    12. If a pattern that was studied started with a repeated Unicode property
840        test, for example, \p{Nd}+, there was the theoretical possibility of
841        setting up an incorrect bitmap of starting bytes, but fortunately it could
842        not have actually happened in practice until change 8 above was made (it
843        added property types that matched character-matching opcodes).
844    
845    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
846        possible starting bytes for non-anchored patterns.
847    
848    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
849        \R, and also a number of cases that involve Unicode properties, both
850        explicit and implicit when PCRE_UCP is set.
851    
852    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
853        input, it could crash or give wrong results if characters with values
854        greater than 0xc0 were present in the subject string. (Detail: it assumed
855        UTF-8 input when processing these items.)
856    
857    16. Added a lot of (int) casts to avoid compiler warnings in systems where
858        size_t is 64-bit (#991).
859    
860    17. Added a check for running out of memory when PCRE is compiled with
861        --disable-stack-for-recursion (#990).
862    
863    18. If the last data line in a file for pcretest does not have a newline on
864        the end, a newline was missing in the output.
865    
866    19. The default pcre_chartables.c file recognizes only ASCII characters (values
867        less than 128) in its various bitmaps. However, there is a facility for
868        generating tables according to the current locale when PCRE is compiled. It
869        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
870        characters, are recognized by isspace() and therefore were getting set in
871        these tables, and indeed these tables seem to approximate to ISO 8859. This
872        caused a problem in UTF-8 mode when pcre_study() was used to create a list
873        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
874        which of course cannot start UTF-8 characters. I have changed the code so
875        that only real ASCII characters (less than 128) and the correct starting
876        bytes for UTF-8 encodings are set for characters greater than 127 when in
877        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
878        altogether.)
879    
880    20. Added the /T option to pcretest so as to be able to run tests with non-
881        standard character tables, thus making it possible to include the tests
882        used for 19 above in the standard set of tests.
883    
884    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
885        reference to a subpattern the other side of a comment that contains an
886        opening parenthesis caused either an internal compiling error, or a
887        reference to the wrong subpattern.
888    
889    
890    Version 8.02 19-Mar-2010
891    ------------------------
892    
893    1.  The Unicode data tables have been updated to Unicode 5.2.0.
894    
895    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
896        configured.
897    
898    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
899        original author of that file, following a query about its status.
900    
901    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
902        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
903    
904    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
905        quantifier applied to a forward-referencing subroutine call, could compile
906        incorrect code or give the error "internal error: previously-checked
907        referenced subpattern not found".
908    
909    6.  Both MS Visual Studio and Symbian OS have problems with initializing
910        variables to point to external functions. For these systems, therefore,
911        pcre_malloc etc. are now initialized to local functions that call the
912        relevant global functions.
913    
914    7.  There were two entries missing in the vectors called coptable and poptable
915        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
916        I've fixed the data, and added a kludgy way of testing at compile time that
917        the lengths are correct (equal to the number of opcodes).
918    
919    8.  Following on from 7, I added a similar kludge to check the length of the
920        eint vector in pcreposix.c.
921    
922    9.  Error texts for pcre_compile() are held as one long string to avoid too
923        much relocation at load time. To find a text, the string is searched,
924        counting zeros. There was no check for running off the end of the string,
925        which could happen if a new error number was added without updating the
926        string.
927    
928    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
929    
930    11. \K was not working if it appeared in an atomic group or in a group that
931        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
932        \K is "not well defined" if used in an assertion. PCRE now accepts it if
933        the assertion is positive, but not if it is negative.
934    
935    12. Change 11 fortuitously reduced the size of the stack frame used in the
936        "match()" function of pcre_exec.c by one pointer. Forthcoming
937        implementation of support for (*MARK) will need an extra pointer on the
938        stack; I have reserved it now, so that the stack frame size does not
939        decrease.
940    
941    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
942        item in branch that calls a recursion is a subroutine call - as in the
943        second branch in the above example - was incorrectly given the compile-
944        time error "recursive call could loop indefinitely" because pcre_compile()
945        was not correctly checking the subroutine for matching a non-empty string.
946    
947    14. The checks for overrunning compiling workspace could trigger after an
948        overrun had occurred. This is a "should never occur" error, but it can be
949        triggered by pathological patterns such as hundreds of nested parentheses.
950        The checks now trigger 100 bytes before the end of the workspace.
951    
952    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
953    
954    
955    Version 8.01 19-Jan-2010
956    ------------------------
957    
958    1.  If a pattern contained a conditional subpattern with only one branch (in
959        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
960        computed the wrong minimum data length (which is of course zero for such
961        subpatterns). This could cause incorrect "no match" results.
962    
963    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
964        the pattern is reset in the first branch, pcre_compile() failed with
965        "internal error: code overflow at offset...". This happened only when
966        the reset was to the original external option setting. (An optimization
967        abstracts leading options settings into an external setting, which was the
968        cause of this.)
969    
970    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
971        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
972        assertion pattern did not match (meaning that the assertion was true), it
973        was incorrectly treated as false if the SKIP had been reached during the
974        matching. This also applied to assertions used as conditions.
975    
976    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
977        assertion subpattern, including such a pattern used as a condition,
978        unpredictable results occurred, instead of the error return
979        PCRE_ERROR_DFA_UITEM.
980    
981    5.  The C++ GlobalReplace function was not working like Perl for the special
982        situation when an empty string is matched. It now does the fancy magic
983        stuff that is necessary.
984    
985    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
986        removed. (These were left over from very, very early versions of PCRE.)
987    
988    7.  Some cosmetic changes to the code to make life easier when compiling it
989        as part of something else:
990    
991        (a) Change DEBUG to PCRE_DEBUG.
992    
993        (b) In pcre_compile(), rename the member of the "branch_chain" structure
994            called "current" as "current_branch", to prevent a collision with the
995            Linux macro when compiled as a kernel module.
996    
997        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
998            prevent a collision with the Linux macro when compiled as a kernel
999            module.
1000    
1001    8.  In pcre_compile() there are some checks for integer overflows that used to
1002        cast potentially large values to (double). This has been changed to that
1003        when building, a check for int64_t is made, and if it is found, it is used
1004        instead, thus avoiding the use of floating point arithmetic. (There is no
1005        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1006        double.
1007    
1008    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1009        2005 (difference between two addresses compared to an unsigned value).
1010    
1011    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1012        custom one, because of the following reported problem in Windows:
1013    
1014          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1015              under Win32.
1016          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1017              therefore missing the function definition.
1018          - The compiler thus generates a "C" signature for the test function.
1019          - The linker fails to find the "C" function.
1020          - PCRE fails to configure if asked to do so against libbz2.
1021    
1022    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1023        messages were output:
1024    
1025          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1026          rerunning libtoolize, to keep the correct libtool macros in-tree.
1027          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1028    
1029        I have done both of these things.
1030    
1031    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1032        most of the time, it *can* run out if it is given a pattern that contains a
1033        runaway infinite recursion. I updated the discussion in the pcrestack man
1034        page.
1035    
1036    13. Now that we have gone to the x.xx style of version numbers, the minor
1037        version may start with zero. Using 08 or 09 is a bad idea because users
1038        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1039        interpreted as invalid octal numbers. I've updated the previous comment in
1040        configure.ac, and also added a check that gives an error if 08 or 09 are
1041        used.
1042    
1043    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1044        causing partial matching to fail when the end of the subject matched \W
1045        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1046    
1047    15. There were some discrepancies between the declarations in pcre_internal.h
1048        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1049        their definitions. The declarations used "const uschar *" and the
1050        definitions used USPTR. Even though USPTR is normally defined as "const
1051        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1052        reported that: "This difference in casting confuses some C++ compilers, for
1053        example, SunCC recognizes above declarations as different functions and
1054        generates broken code for hbpcre." I have changed the declarations to use
1055        USPTR.
1056    
1057    16. GNU libtool is named differently on some systems. The autogen.sh script now
1058        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1059        (FreeBSD).
1060    
1061    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1062        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1063        comment: "Figure out how to create a longlong from a string: strtoll and
1064        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1065        instance, but it only takes 2 args instead of 3!"
1066    
1067    18. A subtle bug concerned with back references has been fixed by a change of
1068        specification, with a corresponding code fix. A pattern such as
1069        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1070        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1071        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1072        same bug. Such groups have to be quantified to be useful, or contained
1073        inside another quantified group. (If there's no repetition, the reference
1074        can never match.) The problem arises because, having left the group and
1075        moved on to the rest of the pattern, a later failure that backtracks into
1076        the group uses the captured value from the final iteration of the group
1077        rather than the correct earlier one. I have fixed this in PCRE by forcing
1078        any group that contains a reference to itself to be an atomic group; that
1079        is, there cannot be any backtracking into it once it has completed. This is
1080        similar to recursive and subroutine calls.
1081    
1082    
1083    Version 8.00 19-Oct-09
1084    ----------------------
1085    
1086    1.  The table for translating pcre_compile() error codes into POSIX error codes
1087        was out-of-date, and there was no check on the pcre_compile() error code
1088        being within the table. This could lead to an OK return being given in
1089        error.
1090    
1091    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1092        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1093        in a Windows environment.
1094    
1095    3.  The pcregrep --count option prints the count for each file even when it is
1096        zero, as does GNU grep. However, pcregrep was also printing all files when
1097        --files-with-matches was added. Now, when both options are given, it prints
1098        counts only for those files that have at least one match. (GNU grep just
1099        prints the file name in this circumstance, but including the count seems
1100        more useful - otherwise, why use --count?) Also ensured that the
1101        combination -clh just lists non-zero counts, with no names.
1102    
1103    4.  The long form of the pcregrep -F option was incorrectly implemented as
1104        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1105        but it seems right to fix it, and I didn't think it was worth preserving
1106        the old behaviour.
1107    
1108    5.  The command line items --regex=pattern and --regexp=pattern were not
1109        recognized by pcregrep, which required --regex pattern or --regexp pattern
1110        (with a space rather than an '='). The man page documented the '=' forms,
1111        which are compatible with GNU grep; these now work.
1112    
1113    6.  No libpcreposix.pc file was created for pkg-config; there was just
1114        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1115    
1116    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1117        when UCP support is not needed, by modifying the Python script that
1118        generates it from Unicode data files. This should not matter if the module
1119        is correctly used as a library, but I received one complaint about 50K of
1120        unwanted data. My guess is that the person linked everything into his
1121        program rather than using a library. Anyway, it does no harm.
1122    
1123    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1124        was a minimum greater than 1 for a wide character in a possessive
1125        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1126        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1127        character. Chaos in the form of incorrect output or a compiling loop could
1128        result.
1129    
1130    9.  The restrictions on what a pattern can contain when partial matching is
1131        requested for pcre_exec() have been removed. All patterns can now be
1132        partially matched by this function. In addition, if there are at least two
1133        slots in the offset vector, the offset of the earliest inspected character
1134        for the match and the offset of the end of the subject are set in them when
1135        PCRE_ERROR_PARTIAL is returned.
1136    
1137    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1138        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1139        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1140        and may be more useful for multi-segment matching.
1141    
1142    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1143        used to be given if ever the end of the subject was reached; now it is
1144        given only if matching could not proceed because another character was
1145        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1146        string "Z", which now yields "no match" instead of "partial match". In the
1147        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1148        final character ended with (*FAIL).
1149    
1150    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1151        if the pattern had a "must contain" character that was already found in the
1152        earlier partial match, unless partial matching was again requested. For
1153        example, with the pattern /dog.(body)?/, the "must contain" character is
1154        "g". If the first part-match was for the string "dog", restarting with
1155        "sbody" failed. This bug has been fixed.
1156    
1157    13. The string returned by pcre_dfa_exec() after a partial match has been
1158        changed so that it starts at the first inspected character rather than the
1159        first character of the match. This makes a difference only if the pattern
1160        starts with a lookbehind assertion or \b or \B (\K is not supported by
1161        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1162        matching functions compatible, and I think it's the right thing to do.
1163    
1164    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1165        so that the demonstration program is easily available in environments where
1166        PCRE has not been installed from source.
1167    
1168    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1169        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1170        library.
1171    
1172    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1173        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1174        is not the first non-POSIX option to be added. Clearly some people find
1175        these options useful.
1176    
1177    17. If a caller to the POSIX matching function regexec() passes a non-zero
1178        value for nmatch with a NULL value for pmatch, the value of
1179        nmatch is forced to zero.
1180    
1181    18. RunGrepTest did not have a test for the availability of the -u option of
1182        the diff command, as RunTest does. It now checks in the same way as
1183        RunTest, and also checks for the -b option.
1184    
1185    19. If an odd number of negated classes containing just a single character
1186        interposed, within parentheses, between a forward reference to a named
1187        subpattern and the definition of the subpattern, compilation crashed with
1188        an internal error, complaining that it could not find the referenced
1189        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1190        [The bug was that it was starting one character too far in when skipping
1191        over the character class, thus treating the ] as data rather than
1192        terminating the class. This meant it could skip too much.]
1193    
1194    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1195        /g option in pcretest when the pattern contains \K, which makes it possible
1196        to have an empty string match not at the start, even when the pattern is
1197        anchored. Updated pcretest and pcredemo to use this option.
1198    
1199    21. If the maximum number of capturing subpatterns in a recursion was greater
1200        than the maximum at the outer level, the higher number was returned, but
1201        with unset values at the outer level. The correct (outer level) value is
1202        now given.
1203    
1204    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1205        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1206        make it do so. The string so far is captured, making this feature
1207        compatible with Perl.
1208    
1209    23. The tests have been re-organized, adding tests 11 and 12, to make it
1210        possible to check the Perl 5.10 features against Perl 5.10.
1211    
1212    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1213        pattern matches a fixed length string. PCRE did not allow this; now it
1214        does. Neither allows recursion.
1215    
1216    25. I finally figured out how to implement a request to provide the minimum
1217        length of subject string that was needed in order to match a given pattern.
1218        (It was back references and recursion that I had previously got hung up
1219        on.) This code has now been added to pcre_study(); it finds a lower bound
1220        to the length of subject needed. It is not necessarily the greatest lower
1221        bound, but using it to avoid searching strings that are too short does give
1222        some useful speed-ups. The value is available to calling programs via
1223        pcre_fullinfo().
1224    
1225    26. While implementing 25, I discovered to my embarrassment that pcretest had
1226        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1227        study optimizations had never been tested with that matching function.
1228        Oops. What is worse, even when it was passed study data, there was a bug in
1229        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1230        were also very few tests of studied patterns with pcre_dfa_exec().
1231    
1232    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1233        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1234        on the other side of the coin, they are no longer allowed to have different
1235        names, because these cannot be distinguished in PCRE, and this has caused
1236        confusion. (This is a difference from Perl.)
1237    
1238    28. When duplicate subpattern names are present (necessarily with different
1239        numbers, as required by 27 above), and a test is made by name in a
1240        conditional pattern, either for a subpattern having been matched, or for
1241        recursion in such a pattern, all the associated numbered subpatterns are
1242        tested, and the overall condition is true if the condition is true for any
1243        one of them. This is the way Perl works, and is also more like the way
1244        testing by number works.
1245    
1246    
1247    Version 7.9 11-Apr-09
1248    ---------------------
1249    
1250    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1251        (pcretest), all targets were linked against these libraries. This included
1252        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1253        libraries. This caused unwanted dependencies to be created. This problem
1254        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1255        pcretest is linked with readline.
1256    
1257    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1258        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1259        moved outside it again, because FALSE and TRUE are already defined in AIX,
1260        but BOOL is not.
1261    
1262    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1263        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1264    
1265    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1266        hyphens) following file names and line numbers when outputting matching
1267        lines. This is not true; no spaces are inserted. I have also clarified the
1268        wording for the --colour (or --color) option.
1269    
1270    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1271        was not coloured; this is different to GNU grep, so I have changed it to be
1272        the same.
1273    
1274    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1275        each matching line was coloured. Now it goes on to look for further matches
1276        of any of the test patterns, which is the same behaviour as GNU grep.
1277    
1278    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1279        doesn't make sense to accept an empty string match in pcregrep, so I have
1280        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1281        seems to be how GNU grep behaves.
1282    
1283    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1284        start or after a newline", because the conditional assertion was not being
1285        correctly handled. The rule now is that both the assertion and what follows
1286        in the first alternative must satisfy the test.
1287    
1288    9.  If auto-callout was enabled in a pattern with a conditional group whose
1289        condition was an assertion, PCRE could crash during matching, both with
1290        pcre_exec() and pcre_dfa_exec().
1291    
1292    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1293        used for matching.
1294    
1295    11. Unicode property support in character classes was not working for
1296        characters (bytes) greater than 127 when not in UTF-8 mode.
1297    
1298    12. Added the -M command line option to pcretest.
1299    
1300    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1301    
1302    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1303    
1304    16. Added comments and documentation about mis-use of no_arg in the C++
1305        wrapper.
1306    
1307    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1308        from Martin Jerabek that uses macro names for all relevant character and
1309        string constants.
1310    
1311    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1312        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1313        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1314        these, but not everybody uses configure.
1315    
1316    19. A conditional group that had only one branch was not being correctly
1317        recognized as an item that could match an empty string. This meant that an
1318        enclosing group might also not be so recognized, causing infinite looping
1319        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1320        with the subject "ab", where knowledge that the repeated group can match
1321        nothing is needed in order to break the loop.
1322    
1323    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1324        exec(), but without supplying a callout function, matching went wrong.
1325    
1326    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1327        leak if the size of the offset vector was greater than 30. When the vector
1328        is smaller, the saved offsets during recursion go onto a local stack
1329        vector, but for larger vectors malloc() is used. It was failing to free
1330        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1331        error, in fact).
1332    
1333    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1334        heapframe that is used only when UTF-8 support is enabled. This caused no
1335        problem, but was untidy.
1336    
1337    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1338        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1339        included within another project.
1340    
1341    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1342        slightly modified by me:
1343    
1344          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1345              not building pcregrep.
1346    
1347          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1348              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1349    
1350    25. Forward references, both numeric and by name, in patterns that made use of
1351        duplicate group numbers, could behave incorrectly or give incorrect errors,
1352        because when scanning forward to find the reference group, PCRE was not
1353        taking into account the duplicate group numbers. A pattern such as
1354        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1355    
1356    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1357        the feature of a custom pointer more persuasive (as requested by a user).
1358    
1359    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1360        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1361        pre-defined.
1362    
1363    28. Added support for (*UTF8) at the start of a pattern.
1364    
1365    29. Arrange for flags added by the "release type" setting in CMake to be shown
1366        in the configuration summary.
1367    
1368    
1369    Version 7.8 05-Sep-08
1370    ---------------------
1371    
1372    1.  Replaced UCP searching code with optimized version as implemented for Ad
1373        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1374        stage table and inline lookup instead of a function, giving speed ups of 2
1375        to 5 times on some simple patterns that I tested. Permission was given to
1376        distribute the MultiStage2.py script that generates the tables (it's not in
1377        the tarball, but is in the Subversion repository).
1378    
1379    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1380        scripts.
1381    
1382    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1383        a group with a zero qualifier. The result of the study could be incorrect,
1384        or the function might crash, depending on the pattern.
1385    
1386    4.  Caseless matching was not working for non-ASCII characters in back
1387        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1388        It now works when Unicode Property Support is available.
1389    
1390    5.  In pcretest, an escape such as \x{de} in the data was always generating
1391        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1392        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1393        truncation.
1394    
1395    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1396    
1397    7.  Added two (int) casts to pcregrep when printing the difference of two
1398        pointers, in case they are 64-bit values.
1399    
1400    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1401        test 2 if it fails.
1402    
1403    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1404        and a #define of that name to empty if it is not externally set. This is to
1405        allow users of MSVC to set it if necessary.
1406    
1407    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1408        the convenience functions in the pcre_get.c source file.
1409    
1410    11. An option change at the start of a pattern that had top-level alternatives
1411        could cause overwriting and/or a crash. This command provoked a crash in
1412        some environments:
1413    
1414          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1415    
1416        This potential security problem was recorded as CVE-2008-2371.
1417    
1418    12. For a pattern where the match had to start at the beginning or immediately
1419        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1420        pcre_dfa_exec() could read past the end of the passed subject if there was
1421        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1422        pcretest so that it places the subject at the end of its malloc-ed buffer.
1423    
1424    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1425        exec() might read past the end of the data buffer in UTF-8 mode.
1426    
1427    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1428        the data contained the byte 0x85 as part of a UTF-8 character within its
1429        first line. This applied both to normal and DFA matching.
1430    
1431    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1432        /^[^d]*?$/8 failed to match "abc".
1433    
1434    16. Added a missing copyright notice to pcrecpp_internal.h.
1435    
1436    17. Make it more clear in the documentation that values returned from
1437        pcre_exec() in ovector are byte offsets, not character counts.
1438    
1439    18. Tidied a few places to stop certain compilers from issuing warnings.
1440    
1441    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1442        supplied by Stefan Weber. I made a further small update for 7.8 because
1443        there is a change of source arrangements: the pcre_searchfuncs.c module is
1444        replaced by pcre_ucd.c.
1445    
1446    
1447    Version 7.7 07-May-08
1448    ---------------------
1449    
1450    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1451        a string to a long long, pretend we don't even have a long long." This is
1452        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1453    
1454    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1455        pre-7.6 versions, which defined a global no_arg variable instead of putting
1456        it in the RE class. (See also #8 below.)
1457    
1458    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1459        Lopes.
1460    
1461    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1462    
1463        (1) The include/exclude patterns were being applied to the whole pathnames
1464            of files, instead of just to the final components.
1465    
1466        (2) If there was more than one level of directory, the subdirectories were
1467            skipped unless they satisfied the include/exclude conditions. This is
1468            inconsistent with GNU grep (and could even be seen as contrary to the
1469            pcregrep specification - which I improved to make it absolutely clear).
1470            The action now is always to scan all levels of directory, and just
1471            apply the include/exclude patterns to regular files.
1472    
1473    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1474        --exclude_dir in the tests to avoid scanning .svn directories.
1475    
1476    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1477        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1478        doesn't support NULs in patterns.
1479    
1480    7.  Added some missing "const"s to declarations of static tables in
1481        pcre_compile.c and pcre_dfa_exec.c.
1482    
1483    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1484        caused by fix #2  above. (Subsequently also a second patch to fix the
1485        first patch. And a third patch - this was a messy problem.)
1486    
1487    9.  Applied Craig's patch to remove the use of push_back().
1488    
1489    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1490        matching function regexec().
1491    
1492    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1493        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1494        references. PCRE supports relative numbers with this syntax (I don't think
1495        Oniguruma does).
1496    
1497    12. Previously, a group with a zero repeat such as (...){0} was completely
1498        omitted from the compiled regex. However, this means that if the group
1499        was called as a subroutine from elsewhere in the pattern, things went wrong
1500        (an internal error was given). Such groups are now left in the compiled
1501        pattern, with a new opcode that causes them to be skipped at execution
1502        time.
1503    
1504    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1505        to the way PCRE behaves:
1506    
1507        (a) A lone ] character is dis-allowed (Perl treats it as data).
1508    
1509        (b) A back reference to an unmatched subpattern matches an empty string
1510            (Perl fails the current match path).
1511    
1512        (c) A data ] in a character class must be notated as \] because if the
1513            first data character in a class is ], it defines an empty class. (In
1514            Perl it is not possible to have an empty class.) The empty class []
1515            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1516            The negative empty class [^] matches any one character, independently
1517            of the DOTALL setting.
1518    
1519    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1520        non-existent subpattern following a character class starting with ']' and
1521        containing () gave an internal compiling error instead of "reference to
1522        non-existent subpattern". Fortunately, when the pattern did exist, the
1523        compiled code was correct. (When scanning forwards to check for the
1524        existencd of the subpattern, it was treating the data ']' as terminating
1525        the class, so got the count wrong. When actually compiling, the reference
1526        was subsequently set up correctly.)
1527    
1528    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1529        it was being rejected as not supported by pcre_dfa_exec(), even though
1530        other assertions are supported. I have made pcre_dfa_exec() support
1531        (*FAIL).
1532    
1533    16. The implementation of 13c above involved the invention of a new opcode,
1534        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1535        cannot be changed at match time, I realized I could make a small
1536        improvement to matching performance by compiling OP_ALLANY instead of
1537        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1538        on the OP_ANY path.
1539    
1540    17. Compiling pcretest on Windows with readline support failed without the
1541        following two fixes: (1) Make the unistd.h include conditional on
1542        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1543    
1544    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1545        ncurses library to be included for pcretest when ReadLine support is
1546        requested, but also to allow for it to be overridden. This patch came from
1547        Daniel Bergström.
1548    
1549    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1550        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1551        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1552        spotting this.
1553    
1554    
1555    Version 7.6 28-Jan-08
1556    ---------------------
1557    
1558    1.  A character class containing a very large number of characters with
1559        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1560        overflow.
1561    
1562    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1563        HAVE_LONG_LONG is not defined.
1564    
1565    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1566        bring them up to date and include new features. This patch includes:
1567    
1568        - Fixed PH's badly added libz and libbz2 support.
1569        - Fixed a problem with static linking.
1570        - Added pcredemo. [But later removed - see 7 below.]
1571        - Fixed dftables problem and added an option.
1572        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1573            HAVE_LONG_LONG.
1574        - Added readline support for pcretest.
1575        - Added an listing of the option settings after cmake has run.
1576    
1577    4.  A user submitted a patch to Makefile that makes it easy to create
1578        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1579        Makefile.am that cause it to include this special target, without
1580        affecting anything else. Note that the same mingw target plus all
1581        the other distribution libraries and programs are now supported
1582        when configuring with CMake (see 6 below) instead of with
1583        Configure/Make.
1584    
1585    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1586        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1587        exported in the Windows port". It has not yet been confirmed that the patch
1588        solves the problem, but it does no harm.
1589    
1590    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1591        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1592        with CMake, and also correct the comment about stack recursion.
1593    
1594    7.  Remove the automatic building of pcredemo from the ./configure system and
1595        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1596        of a program that users should build themselves after PCRE is installed, so
1597        building it automatically is not really right. What is more, it gave
1598        trouble in some build environments.
1599    
1600    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1601    
1602    
1603    Version 7.5 10-Jan-08
1604    ---------------------
1605    
1606    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1607        values in parens when parsing an RE using the C++ wrapper."
1608    
1609    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1610        Characters greater than 255 were excluded from the class instead of being
1611        included.
1612    
1613    3.  The same bug as (2) above applied to negated POSIX classes such as
1614        [:^space:].
1615    
1616    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1617        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1618        I have changed it.
1619    
1620    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1621        first named subpattern) and a construct such as (?&a) would reference the
1622        first named subpattern whose name started with "a" (in other words, the
1623        length check was missing). Both these problems are fixed. "Subpattern name
1624        expected" is now given for (?&) (a zero-length name), and this patch also
1625        makes it give the same error for \k'' (previously it complained that that
1626        was a reference to a non-existent subpattern).
1627    
1628    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1629        this is right because (?- can be followed by option settings as well as by
1630        digits. I have, however, made the messages clearer.
1631    
1632    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1633        than the number used in the conditional) now cause a compile-time error.
1634        This is actually not compatible with Perl, which accepts such patterns, but
1635        treats the conditional as always being FALSE (as PCRE used to), but it
1636        seems to me that giving a diagnostic is better.
1637    
1638    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1639        and messages.
1640    
1641    9.  Fix two occurrences of "backslash" in comments that should have been
1642        "backspace".
1643    
1644    10. Remove two redundant lines of code that can never be obeyed (their function
1645        was moved elsewhere).
1646    
1647    11. The program that makes PCRE's Unicode character property table had a bug
1648        which caused it to generate incorrect table entries for sequences of
1649        characters that have the same character type, but are in different scripts.
1650        It amalgamated them into a single range, with the script of the first of
1651        them. In other words, some characters were in the wrong script. There were
1652        thirteen such cases, affecting characters in the following ranges:
1653    
1654          U+002b0 - U+002c1
1655          U+0060c - U+0060d
1656          U+0061e - U+00612
1657          U+0064b - U+0065e
1658          U+0074d - U+0076d
1659          U+01800 - U+01805
1660          U+01d00 - U+01d77
1661          U+01d9b - U+01dbf
1662          U+0200b - U+0200f
1663          U+030fc - U+030fe
1664          U+03260 - U+0327f
1665          U+0fb46 - U+0fbb1
1666          U+10450 - U+1049d
1667    
1668    12. The -o option (show only the matching part of a line) for pcregrep was not
1669        compatible with GNU grep in that, if there was more than one match in a
1670        line, it showed only the first of them. It now behaves in the same way as
1671        GNU grep.
1672    
1673    13. If the -o and -v options were combined for pcregrep, it printed a blank
1674        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1675        does the same. The return code can be used to tell if there were any
1676        non-matching lines.
1677    
1678    14. Added --file-offsets and --line-offsets to pcregrep.
1679    
1680    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1681        infinitely looping recursion. The bug was that positive lookaheads were not
1682        being skipped when checking for a possible empty match (negative lookaheads
1683        and both kinds of lookbehind were skipped).
1684    
1685    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1686        inclusion of <windows.h> to before rather than after the definition of
1687        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1688    
1689    17. Specifying a possessive quantifier with a specific limit for a Unicode
1690        character property caused pcre_compile() to compile bad code, which led at
1691        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1692        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1693        caused the error; without that there was no problem.
1694    
1695    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1696    
1697    19. Added --enable-pcretest-libreadline.
1698    
1699    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1700        RE::GlobalReplace(). As a result, the number of replacements returned was
1701        double what it should be. I removed one of the increments, but Craig sent a
1702        later patch that removed the other one (the right fix) and added unit tests
1703        that check the return values (which was not done before).
1704    
1705    21. Several CMake things:
1706    
1707        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1708            the names libpcre and libpcreposix, not just pcre and pcreposix.
1709    
1710        (2) The above change means that pcretest and pcregrep are now correctly
1711            linked with the newly-built libraries, not previously installed ones.
1712    
1713        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1714    
1715    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1716        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1717        UTF-8 newline character). The key issue is that the pattern starts .*;
1718        this means that the match must be either at the beginning, or after a
1719        newline. The bug was in the code for advancing after a failed match and
1720        checking that the new position followed a newline. It was not taking
1721        account of UTF-8 characters correctly.
1722    
1723    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1724        character classes. PCRE was not treating the sequence [:...:] as a
1725        character class unless the ... were all letters. Perl, however, seems to
1726        allow any characters between [: and :], though of course it rejects as
1727        unknown any "names" that contain non-letters, because all the known class
1728        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1729        for example, whereas PCRE did not - it did not recognize a POSIX character
1730        class. This seemed a bit dangerous, so the code has been changed to be
1731        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1732        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1733        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1734        Perl does, and where it didn't before.
1735    
1736    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1737        Windows environments %n is disabled by default.
1738    
1739    
1740    Version 7.4 21-Sep-07
1741    ---------------------
1742    
1743    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1744        means that a class such as [\s] counted as "explicit reference to CR or
1745        LF". That isn't really right - the whole point of the change was to try to
1746        help when there was an actual mention of one of the two characters. So now
1747        the change happens only if \r or \n (or a literal CR or LF) character is
1748        encountered.
1749    
1750    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1751        of both had grown to the point where there were only 3 bits left.
1752        Fortunately, there was spare space in the data structure, and so I have
1753        moved the internal flags into a new 16-bit field to free up more option
1754        bits.
1755    
1756    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1757        but did not set the internal JCHANGED flag - either of these is enough to
1758        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1759        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1760        start sets both bits.
1761    
1762    4.  Added options (at build time, compile time, exec time) to change \R from
1763        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1764    
1765    5.  doc/pcresyntax.html was missing from the distribution.
1766    
1767    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1768        compatibility, even though it is no longer used.
1769    
1770    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1771        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1772        windows.h file is present (where different names are used). [This was
1773        reversed later after testing - see 16 below.]
1774    
1775    8.  Changed all #include <config.h> to #include "config.h". There were also
1776        some further <pcre.h> cases that I changed to "pcre.h".
1777    
1778    9.  When pcregrep was used with the --colour option, it missed the line ending
1779        sequence off the lines that it output.
1780    
1781    10. It was pointed out to me that arrays of string pointers cause lots of
1782        relocations when a shared library is dynamically loaded. A technique of
1783        using a single long string with a table of offsets can drastically reduce
1784        these. I have refactored PCRE in four places to do this. The result is
1785        dramatic:
1786    
1787          Originally:                          290
1788          After changing UCP table:            187
1789          After changing error message table:   43
1790          After changing table of "verbs"       36
1791          After changing table of Posix names   22
1792    
1793        Thanks to the folks working on Gregex for glib for this insight.
1794    
1795    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1796        unicode-properties was also set.
1797    
1798    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1799    
1800    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1801        checked only for CRLF.
1802    
1803    14. Added casts to pcretest.c to avoid compiler warnings.
1804    
1805    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1806    
1807    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1808        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1809        entirely. This removes changes made in 7 above.
1810    
1811    17. The CMake files have been updated, and there is now more information about
1812        building with CMake in the NON-UNIX-USE document.
1813    
1814    
1815    Version 7.3 28-Aug-07
1816    ---------------------
1817    
1818     1. In the rejigging of the build system that eventually resulted in 7.1, the
1819        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1820        brackets there is not right, since it causes compilers to look for an
1821        installed pcre.h, not the version that is in the source that is being
1822        compiled (which of course may be different). I have changed it back to:
1823    
1824          #include "pcre.h"
1825    
1826        I have a vague recollection that the change was concerned with compiling in
1827        different directories, but in the new build system, that is taken care of
1828        by the VPATH setting the Makefile.
1829    
1830     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1831        when the subject happened to end in the byte 0x85 (e.g. if the last
1832        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1833        characters but of course it shouldn't be taken as a newline when it is part
1834        of another character. The bug was that, for an unlimited repeat of . in
1835        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1836        characters when looking for a newline.
1837    
1838     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1839    
1840     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1841        in debug output.
1842    
1843     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1844        long printing in the pcrecpp unittest when running under MinGW.
1845    
1846     6. ESC_K was left out of the EBCDIC table.
1847    
1848     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1849        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1850        limit also applies to "virtual nesting" when a pattern is recursive, and in
1851        this case 1000 isn't so big. I have been able to remove this limit at the
1852        expense of backing off one optimization in certain circumstances. Normally,
1853        when pcre_exec() would call its internal match() function recursively and
1854        immediately return the result unconditionally, it uses a "tail recursion"
1855        feature to save stack. However, when a subpattern that can match an empty
1856        string has an unlimited repetition quantifier, it no longer makes this
1857        optimization. That gives it a stack frame in which to save the data for
1858        checking that an empty string has been matched. Previously this was taken
1859        from the 1000-entry workspace that had been reserved. So now there is no
1860        explicit limit, but more stack is used.
1861    
1862     8. Applied Daniel's patches to solve problems with the import/export magic
1863        syntax that is required for Windows, and which was going wrong for the
1864        pcreposix and pcrecpp parts of the library. These were overlooked when this
1865        problem was solved for the main library.
1866    
1867     9. There were some crude static tests to avoid integer overflow when computing
1868        the size of patterns that contain repeated groups with explicit upper
1869        limits. As the maximum quantifier is 65535, the maximum group length was
1870        set at 30,000 so that the product of these two numbers did not overflow a
1871        32-bit integer. However, it turns out that people want to use groups that
1872        are longer than 30,000 bytes (though not repeat them that many times).
1873        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1874        made it possible to implement the integer overflow checks in a much more
1875        dynamic way, which I have now done. The artificial limitation on group
1876        length has been removed - we now have only the limit on the total length of
1877        the compiled pattern, which depends on the LINK_SIZE setting.
1878    
1879    10. Fixed a bug in the documentation for get/copy named substring when
1880        duplicate names are permitted. If none of the named substrings are set, the
1881        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1882        empty string.
1883    
1884    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1885        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1886        because the ] is interpreted as the first data character and the
1887        terminating ] is not found. PCRE has been made compatible with Perl in this
1888        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1889        cause memory overwriting.
1890    
1891    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1892        string has been matched (to stop an infinite loop). It was not recognizing
1893        a conditional subpattern that could match an empty string if that
1894        subpattern was within another subpattern. For example, it looped when
1895        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1896        condition was not nested. This bug has been fixed.
1897    
1898    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1899        past the start of the subject in the presence of bytes with the top bit
1900        set, for example "\x8aBCD".
1901    
1902    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1903        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1904    
1905    14. Optimized (?!) to (*FAIL).
1906    
1907    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1908        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1909        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1910        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1911        does: it's just the validity check that is more restrictive.
1912    
1913    16. Inserted checks for integer overflows during escape sequence (backslash)
1914        processing, and also fixed erroneous offset values for syntax errors during
1915        backslash processing.
1916    
1917    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1918        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1919    
1920    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1921        caused an overrun.
1922    
1923    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1924        something other than just ASCII characters) inside a group that had an
1925        unlimited repeat caused a loop at compile time (while checking to see
1926        whether the group could match an empty string).
1927    
1928    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1929        [\P{Any}] did so. (Error in the code for printing property names.)
1930    
1931    21. An orphan \E inside a character class could cause a crash.
1932    
1933    22. A repeated capturing bracket such as (A)? could cause a wild memory
1934        reference during compilation.
1935    
1936    23. There are several functions in pcre_compile() that scan along a compiled
1937        expression for various reasons (e.g. to see if it's fixed length for look
1938        behind). There were bugs in these functions when a repeated \p or \P was
1939        present in the pattern. These operators have additional parameters compared
1940        with \d, etc, and these were not being taken into account when moving along
1941        the compiled data. Specifically:
1942    
1943        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1944            length.
1945    
1946        (b) An item such as \pL+ within a repeated group could cause crashes or
1947            loops.
1948    
1949        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1950            "reference to non-existent subpattern" error.
1951    
1952        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1953    
1954    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1955        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1956    
1957    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1958    
1959    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1960        character were causing crashes (broken optimization).
1961    
1962    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1963        \p or \P) caused a compile-time loop.
1964    
1965    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1966        break. For example, the unstudied pattern [\r\n]A does not match the string
1967        "\r\nA" because change 7.0/46 below moves the current point on by two
1968        characters after failing to match at the start. However, the pattern \nA
1969        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1970        the same is true. There doesn't seem any very clean way out of this, but
1971        what I have chosen to do makes the common cases work: PCRE now takes note
1972        of whether there can be an explicit match for \r or \n anywhere in the
1973        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1974        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1975        pattern has explicit CR or LF references.
1976    
1977    29. Added (*CR) etc for changing newline setting at start of pattern.
1978    
1979    
1980    Version 7.2 19-Jun-07
1981    ---------------------
1982    
1983     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1984        which is apparently normally available under Windows.
1985    
1986     2. Re-jig the pcregrep tests with different newline settings in an attempt
1987        to make them independent of the local environment's newline setting.
1988    
1989     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1990    
1991     4. Some of the "internals" tests were previously cut out when the link size
1992        was not 2, because the output contained actual offsets. The recent new
1993        "Z" feature of pcretest means that these can be cut out, making the tests
1994        usable with all link sizes.
1995    
1996     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1997        stack recursion. This gives a massive performance boost under BSD, but just
1998        a small improvement under Linux. However, it saves one field in the frame
1999        in all cases.
2000    
2001     6. Added more features from the forthcoming Perl 5.10:
2002    
2003        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2004            recursion call. It refers to the nth most recently opened parentheses.
2005    
2006        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2007            to be opened parentheses.
2008    
2009        (c) Conditions that refer to capturing parentheses can be specified
2010            relatively, for example, (?(-2)... or (?(+3)...
2011    
2012        (d) \K resets the start of the current match so that everything before
2013            is not part of it.
2014    
2015        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2016    
2017        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2018            reference syntax.
2019    
2020        (g) (?| introduces a group in which the numbering of parentheses in each
2021            alternative starts with the same number.
2022    
2023        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2024    
2025     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2026        PCRE_INFO_JCHANGED.
2027    
2028     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2029        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2030        for detecting groups that can match an empty string.
2031    
2032     9. A pattern with a very large number of alternatives (more than several
2033        hundred) was running out of internal workspace during the pre-compile
2034        phase, where pcre_compile() figures out how much memory will be needed. A
2035        bit of new cunning has reduced the workspace needed for groups with
2036        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2037        workspace instead of running out of the 4096 that are available.
2038    
2039    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2040    
2041    11. Applied patch from Google to remove an optimization that didn't quite work.
2042        The report of the bug said:
2043    
2044          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2045          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2046          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2047    
2048    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2049        it matched the wrong number of bytes.
2050    
2051    
2052    Version 7.1 24-Apr-07
2053    ---------------------
2054    
2055     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2056        that is more "standard", making use of automake and other Autotools. There
2057        is some re-arrangement of the files and adjustment of comments consequent
2058        on this.
2059    
2060     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2061        for recursive directory scanning broke on some systems because the files
2062        are not scanned in any specific order and on different systems the order
2063        was different. A call to "sort" has been inserted into RunGrepTest for the
2064        approprate test as a short-term fix. In the longer term there may be an
2065        alternative.
2066    
2067     3. I had an email from Eric Raymond about problems translating some of PCRE's
2068        man pages to HTML (despite the fact that I distribute HTML pages, some
2069        people do their own conversions for various reasons). The problems
2070        concerned the use of low-level troff macros .br and .in. I have therefore
2071        removed all such uses from the man pages (some were redundant, some could
2072        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2073        HTML has been updated to handle .nf/.fi and to complain if it encounters
2074        .br or .in.
2075    
2076     4. Updated comments in configure.ac that get placed in config.h.in and also
2077        arranged for config.h to be included in the distribution, with the name
2078        config.h.generic, for the benefit of those who have to compile without
2079        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2080    
2081     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2082        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2083        makevp.bat for the current PCRE, using the additional files
2084        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2085    
2086     6. A Windows user reported a minor discrepancy with test 2, which turned out
2087        to be caused by a trailing space on an input line that had got lost in his
2088        copy. The trailing space was an accident, so I've just removed it.
2089    
2090     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2091        that is needed.
2092    
2093     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2094        as "const" (a) because they are and (b) because it helps the PHP
2095        maintainers who have recently made a script to detect big data structures
2096        in the php code that should be moved to the .rodata section. I remembered
2097        to update Builducptable as well, so it won't revert if ucptable.h is ever
2098        re-created.
2099    
2100     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2101        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2102        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2103        support is not required. This saves 1.5-2K of code, which is important in
2104        some applications.
2105    
2106        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2107        so as not to refer to the tables, even though these functions will never be
2108        called when UTF-8 support is disabled. Otherwise there are problems with a
2109        shared library.
2110    
2111    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2112    
2113        (a) It was defining its arguments as char * instead of void *.
2114    
2115        (b) It was assuming that all moves were upwards in memory; this was true
2116            a long time ago when I wrote it, but is no longer the case.
2117    
2118        The emulated memove() is provided for those environments that have neither
2119        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2120        is clearly not the case, as these two bugs were recently reported.
2121    
2122    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2123        and Detrail to create the HTML documentation, the .txt form of the man
2124        pages, and it removes trailing spaces from listed files. It also creates
2125        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2126        case, it wraps all the #defines with #ifndefs. This script should be run
2127        before "make dist".
2128    
2129    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2130        with Unicode property support.
2131    
2132        (a) For a maximizing quantifier, if the two different cases of the
2133            character were of different lengths in their UTF-8 codings (there are
2134            some cases like this - I found 11), and the matching function had to
2135            back up over a mixture of the two cases, it incorrectly assumed they
2136            were both the same length.
2137    
2138        (b) When PCRE was configured to use the heap rather than the stack for
2139            recursion during matching, it was not correctly preserving the data for
2140            the other case of a UTF-8 character when checking ahead for a match
2141            while processing a minimizing repeat. If the check also involved
2142            matching a wide character, but failed, corruption could cause an
2143            erroneous result when trying to check for a repeat of the original
2144            character.
2145    
2146    13. Some tidying changes to the testing mechanism:
2147    
2148        (a) The RunTest script now detects the internal link size and whether there
2149            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2150            values substituted by "configure". (The RunGrepTest script already did
2151            this for UTF-8.) The configure.ac script no longer substitutes the
2152            relevant variables.
2153    
2154        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2155            with length and offset values. This means that the output is different
2156            for different internal link sizes. Test 2 is skipped for link sizes
2157            other than 2 because of this, bypassing the problem. Unfortunately,
2158            there was also a test in test 3 (the locale tests) that used /B and
2159            failed for link sizes other than 2. Rather than cut the whole test out,
2160            I have added a new /Z option to pcretest that replaces the length and
2161            offset values with spaces. This is now used to make test 3 independent
2162            of link size. (Test 2 will be tidied up later.)
2163    
2164    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2165        segmentation fault instead of returning the appropriate error message.
2166    
2167    15. In multiline mode when the newline sequence was set to "any", the pattern
2168        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2169        This doesn't seem right; it now treats the CRLF combination as the line
2170        ending, and so does not match in that case. It's only a pattern such as ^$
2171        that would hit this one: something like ^ABC$ would have failed after \r
2172        and then tried again after \r\n.
2173    
2174    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2175        in an attempt to make files that differ only in their line terminators
2176        compare equal. This works on Linux.
2177    
2178    17. Under certain error circumstances pcregrep might try to free random memory
2179        as it exited. This is now fixed, thanks to valgrind.
2180    
2181    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2182        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2183        was because its rules for how to advance for /g after matching an empty
2184        string at the end of a line did not allow for this case. They now check for
2185        it specially.
2186    
2187    20. pcretest is supposed to handle patterns and data of any length, by
2188        extending its buffers when necessary. It was getting this wrong when the
2189        buffer for a data line had to be extended.
2190    
2191    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2192        CRLF as a newline sequence.
2193    
2194    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2195        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2196        I have nevertheless tidied it up.
2197    
2198    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2199    
2200    24. Added a man page for pcre-config.
2201    
2202    
2203    Version 7.0 19-Dec-06
2204    ---------------------
2205    
2206     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2207        moving to gcc 4.1.1.
2208    
2209     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2210        sys/time.h, which is documented as needed for this function. It doesn't
2211        seem to matter on Linux, but it showed up on some releases of OS X.
2212    
2213     3. It seems that there are systems where bytes whose values are greater than
2214        127 match isprint() in the "C" locale. The "C" locale should be the
2215        default when a C program starts up. In most systems, only ASCII printing
2216        characters match isprint(). This difference caused the output from pcretest
2217        to vary, making some of the tests fail. I have changed pcretest so that:
2218    
2219        (a) When it is outputting text in the compiled version of a pattern, bytes
2220            other than 32-126 are always shown as hex escapes.
2221    
2222        (b) When it is outputting text that is a matched part of a subject string,
2223            it does the same, unless a different locale has been set for the match
2224            (using the /L modifier). In this case, it uses isprint() to decide.
2225    
2226     4. Fixed a major bug that caused incorrect computation of the amount of memory
2227        required for a compiled pattern when options that changed within the
2228        pattern affected the logic of the preliminary scan that determines the
2229        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2230        that the computed length was too small. The symptoms of this bug were
2231        either the PCRE error "internal error: code overflow" from pcre_compile(),
2232        or a glibc crash with a message such as "pcretest: free(): invalid next
2233        size (fast)". Examples of patterns that provoked this bug (shown in
2234        pcretest format) are:
2235    
2236          /(?-x: )/x
2237          /(?x)(?-x: \s*#\s*)/
2238          /((?i)[\x{c0}])/8
2239          /(?i:[\x{c0}])/8
2240    
2241        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2242        is now done differently.
2243    
2244     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2245        wrapper classes; (b) implement a new function in the C++ scanner that is
2246        more efficient than the old way of doing things because it avoids levels of
2247        recursion in the regex matching; (c) add a paragraph to the documentation
2248        for the FullMatch() function.
2249    
2250     6. The escape sequence \n was being treated as whatever was defined as
2251        "newline". Not only was this contrary to the documentation, which states
2252        that \n is character 10 (hex 0A), but it also went horribly wrong when
2253        "newline" was defined as CRLF. This has been fixed.
2254    
2255     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2256        was being set to -1 for the "end of line" case (supposedly a value that no
2257        character can have). Though this value is never used (the check for end of
2258        line is "zero bytes in current character"), it caused compiler complaints.
2259        I've changed it to 0xffffffff.
2260    
2261     8. In pcre_version.c, the version string was being built by a sequence of
2262        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2263        string (as it is for production releases) called a macro with an empty
2264        argument. The C standard says the result of this is undefined. The gcc
2265        compiler treats it as an empty string (which was what was wanted) but it is
2266        reported that Visual C gives an error. The source has been hacked around to
2267        avoid this problem.
2268    
2269     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2270        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2271        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2272        of them did).
2273    
2274    10. Originally, pcretest opened its input and output without "b"; then I was
2275        told that "b" was needed in some environments, so it was added for release
2276        5.0 to both the input and output. (It makes no difference on Unix-like
2277        systems.) Later I was told that it is wrong for the input on Windows. I've
2278        now abstracted the modes into two macros, to make it easier to fiddle with
2279        them, and removed "b" from the input mode under Windows.
2280    
2281    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2282    
2283    12. Added -help and --help to pcretest as an official way of being reminded
2284        of the options.
2285    
2286    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2287        and pcrecpp.cc because they annoy compilers at high warning levels.
2288    
2289    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2290    
2291    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2292        scripts are not C programs :-) and which was not noticed because it works
2293        on Linux.
2294    
2295    16. pcretest is supposed to handle any length of pattern and data line (as one
2296        line or as a continued sequence of lines) by extending its input buffer if
2297        necessary. This feature was broken for very long pattern lines, leading to
2298        a string of junk being passed to pcre_compile() if the pattern was longer
2299        than about 50K.
2300    
2301    17. I have done a major re-factoring of the way pcre_compile() computes the
2302        amount of memory needed for a compiled pattern. Previously, there was code
2303        that made a preliminary scan of the pattern in order to do this. That was
2304        OK when PCRE was new, but as the facilities have expanded, it has become
2305        harder and harder to keep it in step with the real compile phase, and there
2306        have been a number of bugs (see for example, 4 above). I have now found a
2307        cunning way of running the real compile function in a "fake" mode that
2308        enables it to compute how much memory it would need, while actually only
2309        ever using a few hundred bytes of working memory and without too many
2310        tests of the mode. This should make future maintenance and development
2311        easier. A side effect of this work is that the limit of 200 on the nesting
2312        depth of parentheses has been removed (though this was never a serious
2313        limitation, I suspect). However, there is a downside: pcre_compile() now
2314        runs more slowly than before (30% or more, depending on the pattern). I
2315        hope this isn't a big issue. There is no effect on runtime performance.
2316    
2317    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2318        newline (only possible for the last line of a file) and it was a
2319        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2320    
2321    19. Added additional timing features to pcretest. (1) The -tm option now times
2322        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2323        separate command line item, by a number that specifies the number of
2324        repeats to use when timing. The default is 50000; this gives better
2325        precision, but takes uncomfortably long for very large patterns.
2326    
2327    20. Extended pcre_study() to be more clever in cases where a branch of a
2328        subpattern has no definite first character. For example, (a*|b*)[cd] would
2329        previously give no result from pcre_study(). Now it recognizes that the
2330        first character must be a, b, c, or d.
2331    
2332    21. There was an incorrect error "recursive call could loop indefinitely" if
2333        a subpattern (or the entire pattern) that was being tested for matching an
2334        empty string contained only one non-empty item after a nested subpattern.
2335        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2336        incorrectly, because the \d was being skipped in the check.
2337    
2338    22. The pcretest program now has a new pattern option /B and a command line
2339        option -b, which is equivalent to adding /B to every pattern. This causes
2340        it to show the compiled bytecode, without the additional information that
2341        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2342        is the same as /B/I).
2343    
2344    23. A new optimization is now able automatically to treat some sequences such
2345        as a*b as a*+b. More specifically, if something simple (such as a character
2346        or a simple class like \d) has an unlimited quantifier, and is followed by
2347        something that cannot possibly match the quantified thing, the quantifier
2348        is automatically "possessified".
2349    
2350    24. A recursive reference to a subpattern whose number was greater than 39
2351        went wrong under certain circumstances in UTF-8 mode. This bug could also
2352        have affected the operation of pcre_study().
2353    
2354    25. Realized that a little bit of performance could be had by replacing
2355        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2356    
2357    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2358    
2359    27. Possessive quantifiers such as a++ were previously implemented by turning
2360        them into atomic groups such as ($>a+). Now they have their own opcodes,
2361        which improves performance. This includes the automatically created ones
2362        from 23 above.
2363    
2364    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2365        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2366        the first matched character to be a colon. This applied both to named and
2367        numbered groups.
2368    
2369    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2370    
2371    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2372        building PCRE on Windows easier, so I have included it in the distribution.
2373    
2374    31. There is now a check in pcretest against a ridiculously large number being
2375        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2376        loop, the loop is abandoned.
2377    
2378    32. Forward references to subpatterns in conditions such as (?(2)...) where
2379        subpattern 2 is defined later cause pcre_compile() to search forwards in
2380        the pattern for the relevant set of parentheses. This search went wrong
2381        when there were unescaped parentheses in a character class, parentheses
2382        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2383    
2384    33. "Subroutine" calls and backreferences were previously restricted to
2385        referencing subpatterns earlier in the regex. This restriction has now
2386        been removed.
2387    
2388    34. Added a number of extra features that are going to be in Perl 5.10. On the
2389        whole, these are just syntactic alternatives for features that PCRE had
2390        previously implemented using the Python syntax or my own invention. The
2391        other formats are all retained for compatibility.
2392    
2393        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2394            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2395            also .NET compatible.
2396    
2397        (b) A recursion or subroutine call to a named group can now be defined as
2398            (?&name) as well as (?P>name).
2399    
2400        (c) A backreference to a named group can now be defined as \k<name> or
2401            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2402            5.10, are also .NET compatible.
2403    
2404        (d) A conditional reference to a named group can now use the syntax
2405            (?(<name>) or (?('name') as well as (?(name).
2406    
2407        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2408            groups (named and numbered) that are never evaluated inline, but can be
2409            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2410            is always false. There may be only one alternative in such a group.
2411    
2412        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2413            as the simple (?(R). The condition is true only if the most recent
2414            recursion is that of the given number or name. It does not search out
2415            through the entire recursion stack.
2416    
2417        (g) The escape \gN or \g{N} has been added, where N is a positive or
2418            negative number, specifying an absolute or relative reference.
2419    
2420    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2421        some "unreachable code" warnings.
2422    
2423    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2424        things, this adds five new scripts.
2425    
2426    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2427        There were also incompatibilities regarding the handling of \Q..\E inside
2428        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2429        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2430    
2431    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2432        matches an empty string, and forcibly breaks the loop. There were bugs in
2433        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2434        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2435        separate and independent bugs (that affected different cases) have been
2436        fixed.
2437    
2438    39. Refactored the code to abolish the use of different opcodes for small
2439        capturing bracket numbers. This is a tidy that I avoided doing when I
2440        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2441        The new approach is not only tidier, it makes it possible to reduce the
2442        memory needed to fix the previous bug (38).
2443    
2444    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2445        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2446        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2447        mode.
2448    
2449    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2450        report.
2451    
2452    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2453        copy construction and assignment in the C++ wrapper.
2454    
2455    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2456        couple of bugs that could have given wrong results in the "--newline=crlf"
2457        case.
2458    
2459    44. Added a number of casts and did some reorganization of signed/unsigned int
2460        variables following suggestions from Dair Grant. Also renamed the variable
2461        "this" as "item" because it is a C++ keyword.
2462    
2463    45. Arranged for dftables to add
2464    
2465          #include "pcre_internal.h"
2466    
2467        to pcre_chartables.c because without it, gcc 4.x may remove the array
2468        definition from the final binary if PCRE is built into a static library and
2469        dead code stripping is activated.
2470    
2471    46. For an unanchored pattern, if a match attempt fails at the start of a
2472        newline sequence, and the newline setting is CRLF or ANY, and the next two
2473        characters are CRLF, advance by two characters instead of one.
2474    
2475    
2476    Version 6.7 04-Jul-06
2477    ---------------------
2478    
2479     1. In order to handle tests when input lines are enormously long, pcretest has
2480        been re-factored so that it automatically extends its buffers when
2481        necessary. The code is crude, but this _is_ just a test program. The
2482        default size has been increased from 32K to 50K.
2483    
2484     2. The code in pcre_study() was using the value of the re argument before
2485        testing it for NULL. (Of course, in any sensible call of the function, it
2486        won't be NULL.)
2487    
2488     3. The memmove() emulation function in pcre_internal.h, which is used on
2489        systems that lack both memmove() and bcopy() - that is, hardly ever -
2490        was missing a "static" storage class specifier.
2491    
2492     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2493        containing an extended class (one that cannot be represented by a bitmap
2494        because it contains high-valued characters or Unicode property items, e.g.
2495        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2496        pattern, but PCRE should not loop if you do not (it no longer does).
2497        [Detail: two cases were found: (a) a repeated subpattern containing an
2498        extended class; (b) a recursive reference to a subpattern that followed a
2499        previous extended class. It wasn't skipping over the extended class
2500        correctly when UTF-8 mode was not set.]
2501    
2502     5. A negated single-character class was not being recognized as fixed-length
2503        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2504        compile error "lookbehind assertion is not fixed length".
2505    
2506     6. The RunPerlTest auxiliary script was showing an unexpected difference
2507        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2508        write a Perl script that can interpret lines of an input file either as
2509        byte characters or as UTF-8, which is what "perltest" was being required to
2510        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2511        can't do is switch easily at run time between having the "use utf8;" pragma
2512        or not. In the end, I fudged it by using the RunPerlTest script to insert
2513        "use utf8;" explicitly for the UTF-8 tests.
2514    
2515     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2516        the end of the subject string, contrary to the documentation and to what
2517        Perl does. This was true of both matching functions. Now it matches only at
2518        the start of the subject and immediately after *internal* newlines.
2519    
2520     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2521        a pointer to an int instead of a pointer to an unsigned long int. This
2522        caused problems on 64-bit systems.
2523    
2524     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2525        instance of the 'standard' template library not being so standard".
2526    
2527    10. There was no check on the number of named subpatterns nor the maximum
2528        length of a subpattern name. The product of these values is used to compute
2529        the size of the memory block for a compiled pattern. By supplying a very
2530        long subpattern name and a large number of named subpatterns, the size
2531        computation could be caused to overflow. This is now prevented by limiting
2532        the length of names to 32 characters, and the number of named subpatterns
2533        to 10,000.
2534    
2535    11. Subpatterns that are repeated with specific counts have to be replicated in
2536        the compiled pattern. The size of memory for this was computed from the
2537        length of the subpattern and the repeat count. The latter is limited to
2538        65535, but there was no limit on the former, meaning that integer overflow
2539        could in principle occur. The compiled length of a repeated subpattern is
2540        now limited to 30,000 bytes in order to prevent this.
2541    
2542    12. Added the optional facility to have named substrings with the same name.
2543    
2544    13. Added the ability to use a named substring as a condition, using the
2545        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2546        are numbers (not recommended). Forward references are permitted.
2547    
2548    14. Added forward references in named backreferences (if you see what I mean).
2549    
2550    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2551        pattern could run off the end of the subject. For example, the pattern
2552        "(?s)(.{1,5})"8 did this with the subject "ab".
2553    
2554    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2555        PCRE_CASELESS was set when matching characters that were quantified with ?
2556        or *.
2557    
2558    17. A character class other than a single negated character that had a minimum
2559        but no maximum quantifier - for example [ab]{6,} - was not handled
2560        correctly by pce_dfa_exec(). It would match only one character.
2561    
2562    18. A valid (though odd) pattern that looked like a POSIX character
2563        class but used an invalid character after [ (for example [[,abc,]]) caused
2564        pcre_compile() to give the error "Failed: internal error: code overflow" or
2565        in some cases to crash with a glibc free() error. This could even happen if
2566        the pattern terminated after [[ but there just happened to be a sequence of
2567        letters, a binary zero, and a closing ] in the memory that followed.
2568    
2569    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2570        over the years. Originally (before any Unicode support), just the bottom 8
2571        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2572        output from "man perlunicode" includes this:
2573    
2574          The regular expression compiler produces polymorphic opcodes.  That
2575          is, the pattern adapts to the data and automatically switches to
2576          the Unicode character scheme when presented with Unicode data--or
2577          instead uses a traditional byte scheme when presented with byte
2578          data.
2579    
2580        Sadly, a wide octal escape does not cause a switch, and in a string with
2581        no other multibyte characters, these octal escapes are treated as before.
2582        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2583        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2584        Unicode string.
2585    
2586        I have not perpetrated such confusion in PCRE. Up till now, it took just
2587        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2588        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2589        translate to the appropriate multibyte character.
2590    
2591    29. Applied some refactoring to reduce the number of warnings from Microsoft
2592        and Borland compilers. This has included removing the fudge introduced
2593        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2594        a warning about an unused variable.
2595    
2596    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2597        characters since release 4.0, because Perl (from release 5.004) does not.
2598        [Or at least, is documented not to: some releases seem to be in conflict
2599        with the documentation.] However, when a pattern was studied with
2600        pcre_study() and all its branches started with \s, PCRE still included VT
2601        as a possible starting character. Of course, this did no harm; it just
2602        caused an unnecessary match attempt.
2603    
2604    22. Removed a now-redundant internal flag bit that recorded the fact that case
2605        dependency changed within the pattern. This was once needed for "required
2606        byte" processing, but is no longer used. This recovers a now-scarce options
2607        bit. Also moved the least significant internal flag bit to the most-
2608        significant bit of the word, which was not previously used (hangover from
2609        the days when it was an int rather than a uint) to free up another bit for
2610        the future.
2611    
2612    23. Added support for CRLF line endings as well as CR and LF. As well as the
2613        default being selectable at build time, it can now be changed at runtime
2614        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2615        specify that it is scanning data with non-default line endings.
2616    
2617    24. Changed the definition of CXXLINK to make it agree with the definition of
2618        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2619    
2620    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2621        recursions. This makes a big different to stack usage for some patterns.
2622    
2623    26. If a subpattern containing a named recursion or subroutine reference such
2624        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2625        the space required for the compiled pattern went wrong and gave too small a
2626        value. Depending on the environment, this could lead to "Failed: internal
2627        error: code overflow at offset 49" or "glibc detected double free or
2628        corruption" errors.
2629    
2630    27. Applied patches from Google (a) to support the new newline modes and (b) to
2631        advance over multibyte UTF-8 characters in GlobalReplace.
2632    
2633    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2634        difference for some implementation of PCRE in some Windows version.
2635    
2636    29. Added some extra testing facilities to pcretest:
2637    
2638        \q<number>   in a data line sets the "match limit" value
2639        \Q<number>   in a data line sets the "match recursion limt" value
2640        -S <number>  sets the stack size, where <number> is in megabytes
2641    
2642        The -S option isn't available for Windows.
2643    
2644    
2645    Version 6.6 06-Feb-06
2646    ---------------------
2647    
2648     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2649        in pcreposix.h. I have copied the definition from pcre.h.
2650    
2651     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2652        because pcre.h is no longer a built file.
2653    
2654     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2655        not normally included in the compiled code.
2656    
2657    
2658    Version 6.5 01-Feb-06
2659    ---------------------
2660    
2661     1. When using the partial match feature with pcre_dfa_exec(), it was not
2662        anchoring the second and subsequent partial matches at the new starting
2663        point. This could lead to incorrect results. For example, with the pattern
2664        /1234/, partially matching against "123" and then "a4" gave a match.
2665    
2666     2. Changes to pcregrep:
2667    
2668        (a) All non-match returns from pcre_exec() were being treated as failures
2669            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2670            error message is output. Some extra information is given for the
2671            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2672            probably the only errors that are likely to be caused by users (by
2673            specifying a regex that has nested indefinite repeats, for instance).
2674            If there are more than 20 of these errors, pcregrep is abandoned.
2675    
2676        (b) A binary zero was treated as data while matching, but terminated the
2677            output line if it was written out. This has been fixed: binary zeroes
2678            are now no different to any other data bytes.
2679    
2680        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2681            used to set a locale for matching. The --locale=xxxx long option has
2682            been added (no short equivalent) to specify a locale explicitly on the
2683            pcregrep command, overriding the environment variables.
2684    
2685        (d) When -B was used with -n, some line numbers in the output were one less
2686            than they should have been.
2687    
2688        (e) Added the -o (--only-matching) option.
2689    
2690        (f) If -A or -C was used with -c (count only), some lines of context were
2691            accidentally printed for the final match.
2692    
2693        (g) Added the -H (--with-filename) option.
2694    
2695        (h) The combination of options -rh failed to suppress file names for files
2696            that were found from directory arguments.
2697    
2698        (i) Added the -D (--devices) and -d (--directories) options.
2699    
2700        (j) Added the -F (--fixed-strings) option.
2701    
2702        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2703    
2704        (l) Added the --colo(u)r option.
2705    
2706        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2707            is not present by default.
2708    
2709     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2710        items such as (?R) or (?1), when the recursion could match a number of
2711        alternatives. If it matched one of the alternatives, but subsequently,
2712        outside the recursion, there was a failure, the code tried to back up into
2713        the recursion. However, because of the way PCRE is implemented, this is not
2714        possible, and the result was an incorrect result from the match.
2715    
2716        In order to prevent this happening, the specification of recursion has
2717        been changed so that all such subpatterns are automatically treated as
2718        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2719    
2720     4. I had overlooked the fact that, in some locales, there are characters for
2721        which isalpha() is true but neither isupper() nor islower() are true. In
2722        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2723        and ordfeminine) are like this. This affected the treatment of \w and \W
2724        when they appeared in character classes, but not when they appeared outside
2725        a character class. The bit map for "word" characters is now created
2726        separately from the results of isalnum() instead of just taking it from the
2727        upper, lower, and digit maps. (Plus the underscore character, of course.)
2728    
2729     5. The above bug also affected the handling of POSIX character classes such as
2730        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2731        permanent tables. Instead, the bit maps for such a class were previously
2732        created as the appropriate unions of the upper, lower, and digit bitmaps.
2733        Now they are created by subtraction from the [[:word:]] class, which has
2734        its own bitmap.
2735    
2736     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2737        It is created by subtracting the vertical space characters (\x09, \x0a,
2738        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2739        subtraction was done in the overall bitmap for a character class, meaning
2740        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2741        be recognized. This bug has been fixed.
2742    
2743     7. Patches from the folks at Google:
2744    
2745          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2746          real life, but is still worth protecting against".
2747    
2748          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2749          regular expressions".
2750    
2751          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2752          have it.
2753    
2754          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2755          "configure" and the latter not, in order to fix a problem somebody had
2756          with compiling the Arg class on HP-UX.
2757    
2758          (e) Improve the error-handling of the C++ wrapper a little bit.
2759    
2760          (f) New tests for checking recursion limiting.
2761    
2762     8. The pcre_memmove() function, which is used only if the environment does not
2763        have a standard memmove() function (and is therefore rarely compiled),
2764        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2765        returning a result (though PCRE never actually uses the result).
2766    
2767     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2768        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2769        returned instead of calling malloc() with an overflowing number that would
2770        most likely cause subsequent chaos.
2771    
2772    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2773    
2774    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2775        with this option is matched, the nmatch and pmatch options of regexec() are
2776        ignored.
2777    
2778    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2779        provided in case anyone wants to the the POSIX interface with UTF-8
2780        strings.
2781    
2782    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2783        C++ linking (needed for some HP-UX environments).
2784    
2785    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2786        (unused parameter) and in the pcre_printint() function (omitted "default"
2787        switch label when the default is to do nothing).
2788    
2789    15. Added some code to make it possible, when PCRE is compiled as a C++
2790        library, to replace subject pointers for pcre_exec() with a smart pointer
2791        class, thus making it possible to process discontinuous strings.
2792    
2793    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2794        much the same function. They were added by different people who were trying
2795        to make PCRE easy to compile on non-Unix systems. It has been suggested
2796        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2797        for compiling on Windows systems. I have therefore replaced it with
2798        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2799        defaults to "extern" for C or "extern C" for C++, which works fine on
2800        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2801        SCOPE with something explicit in config.h. In addition:
2802    
2803        (a) pcreposix.h still had just "extern" instead of either of these macros;
2804            I have replaced it with PCRE_DATA_SCOPE.
2805    
2806        (b) Functions such as _pcre_xclass(), which are internal to the library,
2807            but external in the C sense, all had PCRE_EXPORT in their definitions.
2808            This is apparently wrong for the Windows case, so I have removed it.
2809            (It makes no difference on Unix-like systems.)
2810    
2811    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2812        of recursive calls to match(). This is different to MATCH_LIMIT because
2813        that limits the total number of calls to match(), not all of which increase
2814        the depth of recursion. Limiting the recursion depth limits the amount of
2815        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2816        when PCRE is compiled, and changed at run time. A patch from Google adds
2817        this functionality to the C++ interface.
2818    
2819    18. Changes to the handling of Unicode character properties:
2820    
2821        (a) Updated the table to Unicode 4.1.0.
2822    
2823        (b) Recognize characters that are not in the table as "Cn" (undefined).
2824    
2825        (c) I revised the way the table is implemented to a much improved format
2826            which includes recognition of ranges. It now supports the ranges that
2827            are defined in UnicodeData.txt, and it also amalgamates other
2828            characters into ranges. This has reduced the number of entries in the
2829            table from around 16,000 to around 3,000, thus reducing its size
2830            considerably. I realized I did not need to use a tree structure after
2831            all - a binary chop search is just as efficient. Having reduced the
2832            number of entries, I extended their size from 6 bytes to 8 bytes to
2833            allow for more data.
2834    
2835        (d) Added support for Unicode script names via properties such as \p{Han}.
2836    
2837    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2838        matching that character.
2839    
2840    20. When matching a repeated Unicode property with a minimum greater than zero,
2841        (for example \pL{2,}), PCRE could look past the end of the subject if it
2842        reached it while seeking the minimum number of characters. This could
2843        happen only if some of the characters were more than one byte long, because
2844        there is a check for at least the minimum number of bytes.
2845    
2846    21. Refactored the implementation of \p and \P so as to be more general, to
2847        allow for more different types of property in future. This has changed the
2848        compiled form incompatibly. Anybody with saved compiled patterns that use
2849        \p or \P will have to recompile them.
2850    
2851    22. Added "Any" and "L&" to the supported property types.
2852    
2853    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2854        but give a compile time error if the value is greater than 0xff.
2855    
2856    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2857        accidentally not being installed or uninstalled.
2858    
2859    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2860        made were to insert the current release number. This seemed silly, because
2861        it made things harder for people building PCRE on systems that don't run
2862        "configure". I have turned pcre.h into a distributed file, no longer built
2863        by "configure", with the version identification directly included. There is
2864        no longer a pcre.h.in file.
2865    
2866        However, this change necessitated a change to the pcre-config script as
2867        well. It is built from pcre-config.in, and one of the substitutions was the
2868        release number. I have updated configure.ac so that ./configure now finds
2869        the release number by grepping pcre.h.
2870    
2871    26. Added the ability to run the tests under valgrind.
2872    
2873    
2874    Version 6.4 05-Sep-05
2875    ---------------------
2876    
2877     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2878        "--" to be printed when multiple files were scanned, even when none of the
2879        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2880        consider it to be a bug, and have restored the previous behaviour.
2881    
2882     2. A couple of code tidies to get rid of compiler warnings.
2883    
2884     3. The pcretest program used to cheat by referring to symbols in the library
2885        whose names begin with _pcre_. These are internal symbols that are not
2886        really supposed to be visible externally, and in some environments it is
2887        possible to suppress them. The cheating is now confined to including
2888        certain files from the library's source, which is a bit cleaner.
2889    
2890     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2891        file's purpose clearer.
2892    
2893     5. Reorganized pcre_ucp_findchar().
2894    
2895    
2896    Version 6.3 15-Aug-05
2897    ---------------------
2898    
2899     1. The file libpcre.pc.in did not have general read permission in the tarball.
2900    
2901     2. There were some problems when building without C++ support:
2902    
2903        (a) If C++ support was not built, "make install" and "make test" still
2904            tried to test it.
2905    
2906        (b) There were problems when the value of CXX was explicitly set. Some
2907            changes have been made to try to fix these, and ...
2908    
2909        (c) --disable-cpp can now be used to explicitly disable C++ support.
2910    
2911        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2912            backslash in a target when C++ was disabled. This confuses some
2913            versions of "make", apparently. Using an intermediate variable solves
2914            this. (Same for CPP_LOBJ.)
2915    
2916     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2917        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2918        necessary on certain architectures.
2919    
2920     4. Added a setting of -export-symbols-regex to the link command to remove
2921        those symbols that are exported in the C sense, but actually are local
2922        within the library, and not documented. Their names all begin with
2923        "_pcre_". This is not a perfect job, because (a) we have to except some
2924        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2925        available (and never for static libraries). I have made a note to try to
2926        find a way round (a) in the future.
2927    
2928    
2929    Version 6.2 01-Aug-05
2930    ---------------------
2931    
2932     1. There was no test for integer overflow of quantifier values. A construction
2933        such as {1111111111111111} would give undefined results. What is worse, if
2934        a minimum quantifier for a parenthesized subpattern overflowed and became
2935        negative, the calculation of the memory size went wrong. This could have
2936        led to memory overwriting.
2937    
2938     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2939    
2940     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2941        operating environments where this matters.
2942    
2943     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2944        PCRE options from within the C++ wrapper.
2945    
2946     5. Named capturing subpatterns were not being correctly counted when a pattern
2947        was compiled. This caused two problems: (a) If there were more than 100
2948        such subpatterns, the calculation of the memory needed for the whole
2949        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2950        back references of the form \12, where the number was greater than 9, were
2951        not recognized as back references, even though there were sufficient
2952        previous subpatterns.
2953    
2954     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2955        versions of gcc, e.g. 2.95.4.
2956    
2957    
2958    Version 6.1 21-Jun-05
2959    ---------------------
2960    
2961     1. There was one reference to the variable "posix" in pcretest.c that was not
2962        surrounded by "#if !defined NOPOSIX".
2963    
2964     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2965        the cross-check on the old pcre_info() function, for the benefit of the
2966        cut-down version of PCRE that is currently imported into Exim.
2967    
2968     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2969        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2970        patterns that start (?i) but I don't think that matters. The use of (?i) is
2971        just an example; this all applies to the other options as well.
2972    
2973     4. Since libtool seems to echo the compile commands it is issuing, the output
2974        from "make" can be reduced a bit by putting "@" in front of each libtool
2975        compile command.
2976    
2977     5. Patch from the folks at Google for configure.in to be a bit more thorough
2978        in checking for a suitable C++ installation before trying to compile the
2979        C++ stuff. This should fix a reported problem when a compiler was present,
2980        but no suitable headers.
2981    
2982     6. The man pages all had just "PCRE" as their title. I have changed them to
2983        be the relevant file name. I have also arranged that these names are
2984        retained in the file doc/pcre.txt, which is a concatenation in text format
2985        of all the man pages except the little individual ones for each function.
2986    
2987     7. The NON-UNIX-USE file had not been updated for the different set of source
2988        files that come with release 6. I also added a few comments about the C++
2989        wrapper.
2990    
2991    
2992    Version 6.0 07-Jun-05
2993    ---------------------
2994    
2995     1. Some minor internal re-organization to help with my DFA experiments.
2996    
2997     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2998        didn't matter for the library itself when fully configured, but did matter
2999        when compiling without UCP support, or within Exim, where the ucp files are
3000        not imported.
3001    
3002     3. Refactoring of the library code to split up the various functions into
3003        different source modules. The addition of the new DFA matching code (see
3004        below) to a single monolithic source would have made it really too
3005        unwieldy, quite apart from causing all the code to be include in a
3006        statically linked application, when only some functions are used. This is
3007        relevant even without the DFA addition now that patterns can be compiled in
3008        one application and matched in another.
3009    
3010        The downside of splitting up is that there have to be some external
3011        functions and data tables that are used internally in different modules of
3012        the library but which are not part of the API. These have all had their
3013        names changed to start with "_pcre_" so that they are unlikely to clash
3014        with other external names.
3015    
3016     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3017        a different (DFA) algorithm. Although it is slower than the original
3018        function, it does have some advantages for certain types of matching
3019        problem.
3020    
3021     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3022        including restarting after a partial match.
3023    
3024     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3025        defined when compiling for Windows was sent to me. I have put it into the
3026        code, though I have no means of testing or verifying it.
3027    
3028     7. Added the pcre_refcount() auxiliary function.
3029    
3030     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3031        match before or at the first newline in the subject string. In pcretest,
3032        the /f option on a pattern can be used to set this.
3033    
3034     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3035        would behave wrongly. This has been present in PCRE since release 4.0.
3036    
3037    10. A number of changes to the pcregrep command:
3038    
3039        (a) Refactored how -x works; insert ^(...)$ instead of setting
3040            PCRE_ANCHORED and checking the length, in preparation for adding
3041            something similar for -w.
3042    
3043        (b) Added the -w (match as a word) option.
3044    
3045        (c) Refactored the way lines are read and buffered so as to have more
3046            than one at a time available.
3047    
3048        (d) Implemented a pcregrep test script.
3049    
3050        (e) Added the -M (multiline match) option. This allows patterns to match
3051            over several lines of the subject. The buffering ensures that at least
3052            8K, or the rest of the document (whichever is the shorter) is available
3053            for matching (and similarly the previous 8K for lookbehind assertions).
3054    
3055        (f) Changed the --help output so that it now says
3056    
3057              -w, --word-regex(p)
3058    
3059            instead of two lines, one with "regex" and the other with "regexp"
3060            because that confused at least one person since the short forms are the
3061            same. (This required a bit of code, as the output is generated
3062            automatically from a table. It wasn't just a text change.)
3063    
3064        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3065            option but starts with a hyphen. Could be a pattern or a path name
3066            starting with a hyphen, for instance.
3067    
3068        (h) "-" can be given as a file name to represent stdin.
3069    
3070        (i) When file names are being printed, "(standard input)" is used for
3071            the standard input, for compatibility with GNU grep. Previously
3072            "<stdin>" was used.
3073    
3074        (j) The option --label=xxx can be used to supply a name to be used for
3075            stdin when file names are being printed. There is no short form.
3076    
3077        (k) Re-factored the options decoding logic because we are going to add
3078            two more options that take data. Such options can now be given in four
3079            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3080    
3081        (l) Added the -A, -B, and -C options for requesting that lines of context
3082            around matches be printed.
3083    
3084        (m) Added the -L option to print the names of files that do not contain
3085            any matching lines, that is, the complement of -l.
3086    
3087        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3088            continue to scan other files.
3089    
3090        (o) The -s option was incorrectly implemented. For compatibility with other
3091            greps, it now suppresses the error message for a non-existent or non-
3092            accessible file (but not the return code). There is a new option called
3093            -q that suppresses the output of matching lines, which was what -s was
3094            previously doing.
3095    
3096        (p) Added --include and --exclude options to specify files for inclusion
3097            and exclusion when recursing.
3098    
3099    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3100        Hopefully, it now does.
3101    
3102    12. Missing cast in pcre_study().
3103    
3104    13. Added an "uninstall" target to the makefile.
3105    
3106    14. Replaced "extern" in the function prototypes in Makefile.in with
3107        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3108        world, but is set differently for Windows.
3109    
3110    15. Added a second compiling function called pcre_compile2(). The only
3111        difference is that it has an extra argument, which is a pointer to an
3112        integer error code. When there is a compile-time failure, this is set
3113        non-zero, in addition to the error test pointer being set to point to an
3114        error message. The new argument may be NULL if no error number is required
3115        (but then you may as well call pcre_compile(), which is now just a
3116        wrapper). This facility is provided because some applications need a
3117        numeric error indication, but it has also enabled me to tidy up the way
3118        compile-time errors are handled in the POSIX wrapper.
3119    
3120    16. Added VPATH=.libs to the makefile; this should help when building with one
3121        prefix path and installing with another. (Or so I'm told by someone who
3122        knows more about this stuff than I do.)
3123    
3124    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3125        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3126        match everything, including newlines. This is not POSIX-compatible, but
3127        somebody wanted the feature. From pcretest it can be activated by using
3128        both the P and the s flags.
3129    
3130    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3131    
3132    19. libpcre.pc was being incorrectly installed as executable.
3133    
3134    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3135        it now also looks for '\r' so that it will work unmodified on Windows.
3136    
3137    21. Added Google's contributed C++ wrapper to the distribution.
3138    
3139    22. Added some untidy missing memory free() calls in pcretest, to keep
3140        Electric Fence happy when testing.
3141    
3142    
3143    
3144    Version 5.0 13-Sep-04
3145    ---------------------
3146    
3147     1. Internal change: literal characters are no longer packed up into items
3148        containing multiple characters in a single byte-string. Each character
3149        is now matched using a separate opcode. However, there may be more than one
3150        byte in the character in UTF-8 mode.
3151    
3152     2. The pcre_callout_block structure has two new fields: pattern_position and
3153        next_item_length. These contain the offset in the pattern to the next match
3154        item, and its length, respectively.
3155    
3156     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3157        insertion of callouts before each pattern item. Added the /C option to
3158        pcretest to make use of this.
3159    
3160     4. On the advice of a Windows user, the lines
3161    
3162          #if defined(_WIN32) || defined(WIN32)
3163          _setmode( _fileno( stdout ), 0x8000 );
3164          #endif  /* defined(_WIN32) || defined(WIN32) */
3165    
3166        have been added to the source of pcretest. This apparently does useful
3167        magic in relation to line terminators.
3168    
3169     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3170        for the benefit of those environments where the "b" makes a difference.
3171    
3172     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3173        to know about it. I have put a hack into configure.in that adds in code
3174        to set GCC=yes if CC=icc. This seems to end up at a point in the
3175        generated configure script that is early enough to affect the setting of
3176        compiler options, which is what is needed, but I have no means of testing
3177        whether it really works. (The user who reported this had patched the
3178        generated configure script, which of course I cannot do.)
3179    
3180        LATER: After change 22 below (new libtool files), the configure script
3181        seems to know about icc (and also ecc). Therefore, I have commented out
3182        this hack in configure.in.
3183    
3184     7. Added support for pkg-config (2 patches were sent in).
3185    
3186     8. Negated POSIX character classes that used a combination of internal tables
3187        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3188        [[:^ascii]]. Typically, they would match almost any characters. The other
3189        POSIX classes were not broken in this way.
3190    
3191     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3192        to find the match, as PCRE was deluded into thinking that the match had to
3193        start at the start point or following a newline. The same bug applied to
3194        patterns with negative forward assertions or any backward assertions
3195        preceding ".*" at the start, unless the pattern required a fixed first
3196        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3197    
3198    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3199        starting at the last subject character, bytes beyond the end of the subject
3200        string were read.
3201    
3202    11. Renamed the variable "class" as "classbits" to make life easier for C++
3203        users. (Previously there was a macro definition, but it apparently wasn't
3204        enough.)
3205    
3206    12. Added the new field "tables" to the extra data so that tables can be passed
3207        in at exec time, or the internal tables can be re-selected. This allows
3208        a compiled regex to be saved and re-used at a later time by a different
3209        program that might have everything at different addresses.
3210    
3211    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3212        -R library as well as a -L library.
3213    
3214    14. The debugging options of pcretest (-d on the command line or D on a
3215        pattern) showed incorrect output for anything following an extended class
3216        that contained multibyte characters and which was followed by a quantifier.
3217    
3218    15. Added optional support for general category Unicode character properties
3219        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3220        support. It adds about 90K to the size of the library. The meanings of the
3221        inbuilt class escapes such as \d and \s have NOT been changed.
3222    
3223    16. Updated pcredemo.c to include calls to free() to release the memory for the
3224        compiled pattern.
3225    
3226    17. The generated file chartables.c was being created in the source directory
3227        instead of in the building directory. This caused the build to fail if the
3228        source directory was different from the building directory, and was
3229        read-only.
3230    
3231    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3232        file. No doubt somebody will tell me if they don't make sense... Also added
3233        Dan Mooney's comments about building on OpenVMS.
3234    
3235    19. Added support for partial matching via the PCRE_PARTIAL option for
3236        pcre_exec() and the \P data escape in pcretest.
3237    
3238    20. Extended pcretest with 3 new pattern features:
3239    
3240        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3241              write the compiled pattern to the file whose name is "rest-of-line".
3242              This is a straight binary dump of the data, with the saved pointer to
3243              the character tables forced to be NULL. The study data, if any, is
3244              written too. After writing, pcretest reads a new pattern.
3245    
3246        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3247              compiled pattern from the given file. There must not be any
3248              occurrences of "<" in the file name (pretty unlikely); if there are,
3249              pcretest will instead treat the initial "<" as a pattern delimiter.
3250              After reading in the pattern, pcretest goes on to read data lines as
3251              usual.
3252    
3253        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3254              and 16-bit fields in a compiled pattern, to simulate a pattern that
3255              was compiled on a host of opposite endianness.
3256    
3257    21. The pcre-exec() function can now cope with patterns that were compiled on
3258        hosts of opposite endianness, with this restriction:
3259    
3260          As for any compiled expression that is saved and used later, the tables
3261          pointer field cannot be preserved; the extra_data field in the arguments
3262          to pcre_exec() should be used to pass in a tables address if a value
3263          other than the default internal tables were used at compile time.
3264    
3265    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3266        now diagnosed as an error. Previously, most of the time, a negative number
3267        would have been treated as zero, but if in addition "ovector" was passed as
3268        NULL, a crash could occur.
3269    
3270    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3271        new versions from the libtool 1.5 distribution (the last one is a copy of
3272        a file called libtool.m4). This seems to have fixed the need to patch
3273        "configure" to support Darwin 1.3 (which I used to do). However, I still
3274        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3275        workstation).
3276    
3277    24. Changed the PCRE licence to be the more standard "BSD" licence.
3278    
3279    
3280    Version 4.5 01-Dec-03
3281    ---------------------
3282    
3283     1. There has been some re-arrangement of the code for the match() function so
3284        that it can be compiled in a version that does not call itself recursively.
3285        Instead, it keeps those local variables that need separate instances for
3286        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3287        needs to "recurse". Keeping track of where control must go is done by means
3288        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3289        hide most of the details from the main code, and operates only if
3290        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3291        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3292        operating.
3293    
3294        To make it easier for callers to provide specially tailored get/free
3295        functions for this usage, two new functions, pcre_stack_malloc, and
3296        pcre_stack_free, are used. They are always called in strict stacking order,
3297        and the size of block requested is always the same.
3298    
3299        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3300        PCRE has been compiled to use the stack or the heap for recursion. The
3301        -C option of pcretest uses this to show which version is compiled.
3302    
3303        A new data escape \S, is added to pcretest; it causes the amounts of store
3304        obtained and freed by both kinds of malloc/free at match time to be added
3305        to the output.
3306    
3307     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3308        what's available on my current Linux desktop machine.
3309    
3310     3. When matching a UTF-8 string, the test for a valid string at the start has
3311        been extended. If start_offset is not zero, PCRE now checks that it points
3312        to a byte that is the start of a UTF-8 character. If not, it returns
3313        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3314        this is necessary because there may be backward assertions in the pattern.
3315        When matching the same subject several times, it may save resources to use
3316        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3317    
3318     4. The code for checking the validity of UTF-8 strings has been tightened so
3319        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3320        containing "overlong sequences".
3321    
3322     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3323        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3324        should have been "&", but it just so happened that all the cases this let
3325        through by mistake were picked up later in the function.
3326    
3327     6. I had used a variable called "isblank" - this is a C99 function, causing
3328        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3329    
3330     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3331        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3332        so the version is shown; (c) stop "make test" echoing "./RunTest".
3333    
3334     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3335    
3336     9. The prototype for memmove() for systems that don't have it was using
3337        size_t, but the inclusion of the header that defines size_t was later. I've
3338        moved the #includes for the C headers earlier to avoid this.
3339    
3340    10. Added some adjustments to the code to make it easier to compiler on certain
3341        special systems:
3342    
3343          (a) Some "const" qualifiers were missing.
3344          (b) Added the macro EXPORT before all exported functions; by default this
3345              is defined to be empty.
3346          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3347              that it reads its output file name as an argument instead of writing
3348              to the standard output and assuming this can be redirected.
3349    
3350    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3351        class containing characters with values greater than 255, PCRE compilation
3352        went into a loop.
3353    
3354    12. A recursive reference to a subpattern that was within another subpattern
3355        that had a minimum quantifier of zero caused PCRE to crash. For example,
3356        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3357        recursion. If the recursively-called subpattern itself had a zero repeat,
3358        that was OK.
3359    
3360    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3361        buffer into which it was copied (for escape processing) was still set at
3362        1024, so long lines caused crashes.
3363    
3364    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3365        "internal error: code overflow...". This applied to any character class
3366        that was followed by a possessive quantifier.
3367    
3368    15. Modified the Makefile to add libpcre.la as a prerequisite for
3369        libpcreposix.la because I was told this is needed for a parallel build to
3370        work.
3371    
3372    16. If a pattern that contained .* following optional items at the start was
3373        studied, the wrong optimizing data was generated, leading to matching
3374        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3375        matching string must start with a or b or c. The correct conclusion for
3376        this pattern is that a match can start with any character.
3377    
3378    
3379    Version 4.4 13-Aug-03
3380    ---------------------
3381    
3382     1. In UTF-8 mode, a character class containing characters with values between
3383        127 and 255 was not handled correctly if the compiled pattern was studied.
3384        In fixing this, I have also improved the studying algorithm for such
3385        classes (slightly).
3386    
3387     2. Three internal functions had redundant arguments passed to them. Removal
3388        might give a very teeny performance improvement.
3389    
3390     3. Documentation bug: the value of the capture_top field in a callout is *one
3391        more than* the number of the hightest numbered captured substring.
3392    
3393     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3394        in incorrectly linking with a previously installed version. They now link
3395        explicitly with libpcre.la.
3396    
3397     5. configure.in no longer needs to recognize Cygwin specially.
3398    
3399     6. A problem in pcre.in for Windows platforms is fixed.
3400    
3401     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3402        pcretest, it used to include the size of the study block as part of its
3403        output. Unfortunately, the structure contains a field that has a different
3404        size on different hardware architectures. This meant that the tests that
3405        showed this size failed. As the block is currently always of a fixed size,
3406        this information isn't actually particularly useful in pcretest output, so
3407        I have just removed it.
3408    
3409     8. Three pre-processor statements accidentally did not start in column 1.
3410        Sadly, there are *still* compilers around that complain, even though
3411        standard C has not required this for well over a decade. Sigh.
3412    
3413     9. In pcretest, the code for checking callouts passed small integers in the
3414        callout_data field, which is a void * field. However, some picky compilers
3415        complained about the casts involved for this on 64-bit systems. Now
3416        pcretest passes the address of the small integer instead, which should get
3417        rid of the warnings.
3418    
3419    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3420        both compile and run time, and gives an error if an invalid UTF-8 sequence
3421        is found. There is a option for disabling this check in cases where the
3422        string is known to be correct and/or the maximum performance is wanted.
3423    
3424    11. In response to a bug report, I changed one line in Makefile.in from
3425    
3426            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3427        to
3428            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3429    
3430        to look similar to other lines, but I have no way of telling whether this
3431        is the right thing to do, as I do not use Windows. No doubt I'll get told
3432        if it's wrong...
3433    
3434    
3435    Version 4.3 21-May-03
3436    ---------------------
3437    
3438    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3439       Makefile.
3440    
3441    2. Some refactoring to improve the quality of the code:
3442    
3443       (i)   The utf8_table... variables are now declared "const".
3444    
3445       (ii)  The code for \cx, which used the "case flipping" table to upper case
3446             lower case letters, now just substracts 32. This is ASCII-specific,
3447             but the whole concept of \cx is ASCII-specific, so it seems
3448             reasonable.
3449    
3450       (iii) PCRE was using its character types table to recognize decimal and
3451             hexadecimal digits in the pattern. This is silly, because it handles
3452             only 0-9, a-f, and A-F, but the character types table is locale-
3453             specific, which means strange things might happen. A private
3454             table is now used for this - though it costs 256 bytes, a table is
3455             much faster than multiple explicit tests. Of course, the standard
3456             character types table is still used for matching digits in subject
3457             strings against \d.
3458    
3459       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3460             ending in _t are). So I've renamed it as ESC_tee.
3461    
3462    3. The first argument for regexec() in the POSIX wrapper should have been
3463       defined as "const".
3464    
3465    4. Changed pcretest to use malloc() for its buffers so that they can be
3466       Electric Fenced for debugging.
3467    
3468    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3469       to read one or more bytes before the start of the subject string. Often this
3470       had no effect on PCRE's behaviour, but in some circumstances it could
3471       provoke a segmentation fault.
3472    
3473    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3474       to try to read one or more bytes before the start of the subject string.
3475    
3476    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3477       UTF-8 support could misbehave in various ways if the subject string
3478       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3479       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3480       back over UTF-8 characters.)
3481    
3482    
3483    Version 4.2 14-Apr-03
3484    ---------------------
3485    
3486    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3487    
3488    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3489         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3490         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3491         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3492         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3493           and BUILD_EXEEXT
3494         Note: automatic setting of the BUILD variables is not yet working
3495         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3496           compile-time but not at link-time
3497         [LINK]: use for linking executables only
3498         make different versions for Windows and non-Windows
3499         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3500           libraries
3501         [LINK_FOR_BUILD]: new variable
3502         [OBJEXT]: use throughout