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

Diff of /code/trunk/ChangeLog

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

revision 11 by nigel, Sat Feb 24 21:38:17 2007 UTC revision 934 by ph10, Sat Feb 25 12:30:36 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. (*COMMIT) is now correctly confined to within a recursive subpattern call.
59    
60    
61    Version 8.30 04-February-2012
62    -----------------------------
63    
64    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
65        name is defined in ctype.h.
66    
67    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
68        only in quite long subpatterns.
69    
70    3.  Removed the function pcre_info(), which has been obsolete and deprecated
71        since it was replaced by pcre_fullinfo() in February 2000.
72    
73    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
74        match a (*MARK), and the match failed at the start of the subject, a
75        reference to memory before the start of the subject could occur. This bug
76        was introduced by fix 17 of release 8.21.
77    
78    5.  A reference to an unset group with zero minimum repetition was giving
79        totally wrong answers (in non-JavaScript-compatibility mode). For example,
80        /(another)?(\1?)test/ matched against "hello world test". This bug was
81        introduced in release 8.13.
82    
83    6.  Add support for 16-bit character strings (a large amount of work involving
84        many changes and refactorings).
85    
86    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
87        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
88        from a file.
89    
90    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
91        rounding is not applied in this particular case).
92    
93    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
94        if they appear, or are escaped, in patterns.
95    
96    10. Get rid of a number of -Wunused-but-set-variable warnings.
97    
98    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
99        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
100        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
101        also returns the mark "x". This bug applied to capturing parentheses,
102        non-capturing parentheses, and atomic parentheses. It also applied to some
103        assertions.
104    
105    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
106        information out of configure.ac instead of relying on pcre.h.generic, which
107        is not stored in the repository.
108    
109    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
110        -lreadline.
111    
112    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
113    
114    15. Applied Graycode's patch to put the top-level frame on the stack rather
115        than the heap when not using the stack for recursion. This gives a
116        performance improvement in many cases when recursion is not deep.
117    
118    16. Experimental code added to "pcretest -C" to output the stack frame size.
119    
120    
121    Version 8.21 12-Dec-2011
122    ------------------------
123    
124    1.  Updating the JIT compiler.
125    
126    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
127        are added as well.
128    
129    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
130        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
131        calling _pcre_jit_exec. Some extra comments are added.
132    
133    4.  (*MARK) settings inside atomic groups that do not contain any capturing
134        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
135        was introduced by change 18 for 8.20.
136    
137    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
138        ECMA-262 standard.
139    
140    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
141        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
142        This bug was probably introduced by change 9 of 8.13.
143    
144    7.  While fixing 6 above, I noticed that a number of other items were being
145        incorrectly rejected as "not fixed length". This arose partly because newer
146        opcodes had not been added to the fixed-length checking code. I have (a)
147        corrected the bug and added tests for these items, and (b) arranged for an
148        error to occur if an unknown opcode is encountered while checking for fixed
149        length instead of just assuming "not fixed length". The items that were
150        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
151        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
152        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
153    
154    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
155        being incorrectly compiled and would have given unpredicatble results.
156    
157    9.  A possessively repeated subpattern with minimum repeat count greater than
158        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
159        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
160        the first (A) could occur when it should not.
161    
162    10. Add a cast and remove a redundant test from the code.
163    
164    11. JIT should use pcre_malloc/pcre_free for allocation.
165    
166    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
167        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
168        is anything other than /usr, -L is still shown).
169    
170    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
171    
172    14. Perl does not support \N without a following name in a [] class; PCRE now
173        also gives an error.
174    
175    15. If a forward reference was repeated with an upper limit of around 2000,
176        it caused the error "internal error: overran compiling workspace". The
177        maximum number of forward references (including repeats) was limited by the
178        internal workspace, and dependent on the LINK_SIZE. The code has been
179        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
180        the default depends on LINK_SIZE. There is a new upper limit (for safety)
181        of around 200,000 forward references. While doing this, I also speeded up
182        the filling in of repeated forward references.
183    
184    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
185        incorrectly expecting the subject to contain another "a" after the start.
186    
187    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
188        in the match, the SKIP should be ignored. This was not happening; instead
189        the SKIP was being treated as NOMATCH. For patterns such as
190        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
191        tested.
192    
193    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
194        now much more compatible with Perl, in particular in cases where the result
195        is a non-match for a non-anchored pattern. For example, if
196        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
197        "m", where previously it did not return a name. A side effect of this
198        change is that for partial matches, the last encountered mark name is
199        returned, as for non matches. A number of tests that were previously not
200        Perl-compatible have been moved into the Perl-compatible test files. The
201        refactoring has had the pleasing side effect of removing one argument from
202        the match() function, thus reducing its stack requirements.
203    
204    19. If the /S+ option was used in pcretest to study a pattern using JIT,
205        subsequent uses of /S (without +) incorrectly behaved like /S+.
206    
207    21. Retrieve executable code size support for the JIT compiler and fixing
208        some warnings.
209    
210    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
211        not work when the shorter character appeared right at the end of the
212        subject string.
213    
214    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
215        systems.
216    
217    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
218        output it when the /M option is used in pcretest.
219    
220    25. The CheckMan script was not being included in the distribution. Also, added
221        an explicit "perl" to run Perl scripts from the PrepareRelease script
222        because this is reportedly needed in Windows.
223    
224    26. If study data was being save in a file and studying had not found a set of
225        "starts with" bytes for the pattern, the data written to the file (though
226        never used) was taken from uninitialized memory and so caused valgrind to
227        complain.
228    
229    27. Updated RunTest.bat as provided by Sheri Pierce.
230    
231    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
232    
233    29. Computation of memory usage for the table of capturing group names was
234        giving an unnecessarily large value.
235    
236    
237    Version 8.20 21-Oct-2011
238    ------------------------
239    
240    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
241        a POSIX class. After further experiments with Perl, which convinced me that
242        Perl has bugs and confusions, a closing square bracket is no longer allowed
243        in a POSIX name. This bug also affected patterns with classes that started
244        with full stops.
245    
246    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
247        captured substring, but while checking the failing first alternative,
248        substring 1 is temporarily captured. If the output vector supplied to
249        pcre_exec() was not big enough for this capture, the yield of the function
250        was still zero ("insufficient space for captured substrings"). This cannot
251        be totally fixed without adding another stack variable, which seems a lot
252        of expense for a edge case. However, I have improved the situation in cases
253        such as /(a)(b)x|abc/ matched against "abc", where the return code
254        indicates that fewer than the maximum number of slots in the ovector have
255        been set.
256    
257    3.  Related to (2) above: when there are more back references in a pattern than
258        slots in the output vector, pcre_exec() uses temporary memory during
259        matching, and copies in the captures as far as possible afterwards. It was
260        using the entire output vector, but this conflicts with the specification
261        that only 2/3 is used for passing back captured substrings. Now it uses
262        only the first 2/3, for compatibility. This is, of course, another edge
263        case.
264    
265    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
266        main code base, and can be used by building with --enable-jit. When this is
267        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
268        runtime --no-jit option is given.
269    
270    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
271        ovector, the return from the function was zero, implying that there were
272        other matches that did not fit. The correct "exactly full" value is now
273        returned.
274    
275    6.  If a subpattern that was called recursively or as a subroutine contained
276        (*PRUNE) or any other control that caused it to give a non-standard return,
277        invalid errors such as "Error -26 (nested recursion at the same subject
278        position)" or even infinite loops could occur.
279    
280    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
281        computing the minimum length on reaching *ACCEPT, and so ended up with the
282        wrong value of 1 rather than 0. Further investigation indicates that
283        computing a minimum subject length in the presence of *ACCEPT is difficult
284        (think back references, subroutine calls), and so I have changed the code
285        so that no minimum is registered for a pattern that contains *ACCEPT.
286    
287    8.  If (*THEN) was present in the first (true) branch of a conditional group,
288        it was not handled as intended. [But see 16 below.]
289    
290    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
291        Sheri Pierce.
292    
293    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
294        the first byte in a match must be "a".
295    
296    11. Change 17 for 8.13 increased the recursion depth for patterns like
297        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
298        pattern contains any instances of (*THEN). If it does not, the old
299        optimizations are restored. It would be nice to do this on a per-group
300        basis, but at the moment that is not feasible.
301    
302    12. In some environments, the output of pcretest -C is CRLF terminated. This
303        broke RunTest's code that checks for the link size. A single white space
304        character after the value is now allowed for.
305    
306    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
307        For "fr", it uses the Windows-specific input and output files.
308    
309    14. If (*THEN) appeared in a group that was called recursively or as a
310        subroutine, it did not work as intended. [But see next item.]
311    
312    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
313        pattern fragments (but not containing any | characters). If A and B are
314        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
315        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
316        D. In other words, Perl considers parentheses that do not contain any |
317        characters to be part of a surrounding alternative, whereas PCRE was
318        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
319        differently. PCRE now behaves in the same way as Perl, except in the case
320        of subroutine/recursion calls such as (?1) which have in any case always
321        been different (but PCRE had them first :-).
322    
323    16. Related to 15 above: Perl does not treat the | in a conditional group as
324        creating alternatives. Such a group is treated in the same way as an
325        ordinary group without any | characters when processing (*THEN). PCRE has
326        been changed to match Perl's behaviour.
327    
328    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
329        RunGrepTest script failed.
330    
331    18. Change 22 for version 13 caused atomic groups to use more stack. This is
332        inevitable for groups that contain captures, but it can lead to a lot of
333        stack use in large patterns. The old behaviour has been restored for atomic
334        groups that do not contain any capturing parentheses.
335    
336    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
337        suppress the check for a minimum subject length at run time. (If it was
338        given to pcre_exec() or pcre_dfa_exec() it did work.)
339    
340    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
341        fail to work when decoding hex characters in data strings in EBCDIC
342        environments.
343    
344    21. It appears that in at least one Mac OS environment, the isxdigit() function
345        is implemented as a macro that evaluates to its argument more than once,
346        contravening the C 90 Standard (I haven't checked a later standard). There
347        was an instance in pcretest which caused it to go wrong when processing
348        \x{...} escapes in subject strings. The has been rewritten to avoid using
349        things like p++ in the argument of isxdigit().
350    
351    
352    Version 8.13 16-Aug-2011
353    ------------------------
354    
355    1.  The Unicode data tables have been updated to Unicode 6.0.0.
356    
357    2.  Two minor typos in pcre_internal.h have been fixed.
358    
359    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
360        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
361        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
362    
363    4.  There were a number of related bugs in the code for matching backrefences
364        caselessly in UTF-8 mode when codes for the characters concerned were
365        different numbers of bytes. For example, U+023A and U+2C65 are an upper
366        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
367        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
368        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
369        2-byte code at the end of the subject (it thought there wasn't enough data
370        left).
371    
372    5.  Comprehensive information about what went wrong is now returned by
373        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
374        as the output vector has at least 2 elements. The offset of the start of
375        the failing character and a reason code are placed in the vector.
376    
377    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
378        now returned is for the first byte of the failing character, instead of the
379        last byte inspected. This is an incompatible change, but I hope it is small
380        enough not to be a problem. It makes the returned offset consistent with
381        pcre_exec() and pcre_dfa_exec().
382    
383    7.  pcretest now gives a text phrase as well as the error number when
384        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
385        failure, the offset and reason code are output.
386    
387    8.  When \R was used with a maximizing quantifier it failed to skip backwards
388        over a \r\n pair if the subsequent match failed. Instead, it just skipped
389        back over a single character (\n). This seems wrong (because it treated the
390        two characters as a single entity when going forwards), conflicts with the
391        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
392        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
393        has been changed.
394    
395    9.  Some internal refactoring has changed the processing so that the handling
396        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
397        time (the PCRE_DOTALL option was changed this way some time ago: version
398        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
399        which was always a bit of a fudge. It also means that there is one less
400        argument for the match() function, which reduces its stack requirements
401        slightly. This change also fixes an incompatibility with Perl: the pattern
402        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
403    
404    10. More internal refactoring has drastically reduced the number of recursive
405        calls to match() for possessively repeated groups such as (abc)++ when
406        using pcre_exec().
407    
408    11. While implementing 10, a number of bugs in the handling of groups were
409        discovered and fixed:
410    
411        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
412        (a|)*(?1) gave a compile-time internal error.
413        ((a|)+)+  did not notice that the outer group could match an empty string.
414        (^a|^)+   was not marked as anchored.
415        (.*a|.*)+ was not marked as matching at start or after a newline.
416    
417    12. Yet more internal refactoring has removed another argument from the match()
418        function. Special calls to this function are now indicated by setting a
419        value in a variable in the "match data" data block.
420    
421    13. Be more explicit in pcre_study() instead of relying on "default" for
422        opcodes that mean there is no starting character; this means that when new
423        ones are added and accidentally left out of pcre_study(), testing should
424        pick them up.
425    
426    14. The -s option of pcretest has been documented for ages as being an old
427        synonym of -m (show memory usage). I have changed it to mean "force study
428        for every regex", that is, assume /S for every regex. This is similar to -i
429        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
430        using it. It makes it easier to run collections of tests with and without
431        study enabled, and thereby test pcre_study() more easily. All the standard
432        tests are now run with and without -s (but some patterns can be marked as
433        "never study" - see 20 below).
434    
435    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
436        restoration of the capturing data to the outer values was not happening
437        correctly.
438    
439    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
440        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
441        pattern had matched an empty string, and so incorrectly returned a no
442        match.
443    
444    17. There was optimizing code for the last branch of non-capturing parentheses,
445        and also for the obeyed branch of a conditional subexpression, which used
446        tail recursion to cut down on stack usage. Unfortunately, now that there is
447        the possibility of (*THEN) occurring in these branches, tail recursion is
448        no longer possible because the return has to be checked for (*THEN). These
449        two optimizations have therefore been removed. [But see 8.20/11 above.]
450    
451    18. If a pattern containing \R was studied, it was assumed that \R always
452        matched two bytes, thus causing the minimum subject length to be
453        incorrectly computed because \R can also match just one byte.
454    
455    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
456        was incorrectly computed.
457    
458    20. If /S is present twice on a test pattern in pcretest input, it now
459        *disables* studying, thereby overriding the use of -s on the command line
460        (see 14 above). This is necessary for one or two tests to keep the output
461        identical in both cases.
462    
463    21. When (*ACCEPT) was used in an assertion that matched an empty string and
464        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
465    
466    22. When an atomic group that contained a capturing parenthesis was
467        successfully matched, but the branch in which it appeared failed, the
468        capturing was not being forgotten if a higher numbered group was later
469        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
470        group 1 to "a", when in fact it should be unset. This applied to multi-
471        branched capturing and non-capturing groups, repeated or not, and also to
472        positive assertions (capturing in negative assertions does not happen
473        in PCRE) and also to nested atomic groups.
474    
475    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
476        subject after a captured substring, to make it easier to tell which of a
477        number of identical substrings has been captured.
478    
479    24. The way atomic groups are processed by pcre_exec() has been changed so that
480        if they are repeated, backtracking one repetition now resets captured
481        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
482        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
483        "aaa". Previously, it would have been "a". As part of this code
484        refactoring, the way recursive calls are handled has also been changed.
485    
486    25. If an assertion condition captured any substrings, they were not passed
487        back unless some other capturing happened later. For example, if
488        (?(?=(a))a) was matched against "a", no capturing was returned.
489    
490    26. When studying a pattern that contained subroutine calls or assertions,
491        the code for finding the minimum length of a possible match was handling
492        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
493        group 1 called group 2 while simultaneously a separate group 2 called group
494        1). A stack overflow occurred in this case. I have fixed this by limiting
495        the recursion depth to 10.
496    
497    27. Updated RunTest.bat in the distribution to the version supplied by Tom
498        Fortmann. This supports explicit test numbers on the command line, and has
499        argument validation and error reporting.
500    
501    28. An instance of \X with an unlimited repeat could fail if at any point the
502        first character it looked at was a mark character.
503    
504    29. Some minor code refactoring concerning Unicode properties and scripts
505        should reduce the stack requirement of match() slightly.
506    
507    30. Added the '=' option to pcretest to check the setting of unused capturing
508        slots at the end of the pattern, which are documented as being -1, but are
509        not included in the return count.
510    
511    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
512        compiled something random. Now it gives a compile-time error (as does
513        Perl).
514    
515    32. A *MARK encountered during the processing of a positive assertion is now
516        recorded and passed back (compatible with Perl).
517    
518    33. If --only-matching or --colour was set on a pcregrep call whose pattern
519        had alternative anchored branches, the search for a second match in a line
520        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
521        matched the line "0102" twice. The same bug affected patterns that started
522        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
523        twice.
524    
525    34. Previously, PCRE did not allow quantification of assertions. However, Perl
526        does, and because of capturing effects, quantifying parenthesized
527        assertions may at times be useful. Quantifiers are now allowed for
528        parenthesized assertions.
529    
530    35. A minor code tidy in pcre_compile() when checking options for \R usage.
531    
532    36. \g was being checked for fancy things in a character class, when it should
533        just be a literal "g".
534    
535    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
536        appearance of a nested POSIX class supersedes an apparent external class.
537        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
538        unescaped square brackets may also appear as part of class names. For
539        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
540        more like Perl. (But see 8.20/1 above.)
541    
542    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
543        was because it thought it was \N{name}, which is not supported).
544    
545    39. Add minix to OS list not supporting the -S option in pcretest.
546    
547    40. PCRE tries to detect cases of infinite recursion at compile time, but it
548        cannot analyze patterns in sufficient detail to catch mutual recursions
549        such as ((?1))((?2)). There is now a runtime test that gives an error if a
550        subgroup is called recursively as a subpattern for a second time at the
551        same position in the subject string. In previous releases this might have
552        been caught by the recursion limit, or it might have run out of stack.
553    
554    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
555        happen only once. PCRE was, however incorrectly giving a compile time error
556        "recursive call could loop indefinitely" because it cannot analyze the
557        pattern in sufficient detail. The compile time test no longer happens when
558        PCRE is compiling a conditional subpattern, but actual runaway loops are
559        now caught at runtime (see 40 above).
560    
561    42. It seems that Perl allows any characters other than a closing parenthesis
562        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
563        has been changed to be the same.
564    
565    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
566        as not to get warnings when autogen.sh is called. Also changed
567        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
568    
569    44. To help people who use pcregrep to scan files containing exceedingly long
570        lines, the following changes have been made:
571    
572        (a) The default value of the buffer size parameter has been increased from
573            8K to 20K. (The actual buffer used is three times this size.)
574    
575        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
576            PCRE is built.
577    
578        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
579            to be set at run time.
580    
581        (d) Numerical values in pcregrep options can be followed by K or M, for
582            example --buffer-size=50K.
583    
584        (e) If a line being scanned overflows pcregrep's buffer, an error is now
585            given and the return code is set to 2.
586    
587    45. Add a pointer to the latest mark to the callout data block.
588    
589    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
590        partial match of an empty string instead of no match. This was specific to
591        the use of ".".
592    
593    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
594        complete match instead of a partial match. This bug was dependent on both
595        the PCRE_UTF8 and PCRE_DOTALL options being set.
596    
597    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
598        starting byte set, because \b was not being ignored.
599    
600    
601    Version 8.12 15-Jan-2011
602    ------------------------
603    
604    1.  Fixed some typos in the markup of the man pages, and wrote a script that
605        checks for such things as part of the documentation building process.
606    
607    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
608        --match-limit and --recursion-limit options (added for 8.11). In
609        particular, this made one of the standard tests fail. (The integer value
610        went into the wrong half of a long int.)
611    
612    3.  If the --colour option was given to pcregrep with -v (invert match), it
613        did strange things, either producing crazy output, or crashing. It should,
614        of course, ignore a request for colour when reporting lines that do not
615        match.
616    
617    4.  Another pcregrep bug caused similar problems if --colour was specified with
618        -M (multiline) and the pattern match finished with a line ending.
619    
620    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
621        matched in multiline mode, the following line was shown as part of the
622        match. This seems wrong, so I have changed it.
623    
624    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
625        the check for further matches in the same line (so they could be coloured)
626        to overrun the end of the current line. If another match was found, it was
627        incorrectly shown (and then shown again when found in the next line).
628    
629    7.  If pcregrep was compiled under Windows, there was a reference to the
630        function pcregrep_exit() before it was defined. I am assuming this was
631        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
632        reported by a user. I've moved the definition above the reference.
633    
634    
635    Version 8.11 10-Dec-2010
636    ------------------------
637    
638    1.  (*THEN) was not working properly if there were untried alternatives prior
639        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
640        backtracked to try for "b" instead of moving to the next alternative branch
641        at the same level (in this case, to look for "c"). The Perl documentation
642        is clear that when (*THEN) is backtracked onto, it goes to the "next
643        alternative in the innermost enclosing group".
644    
645    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
646        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
647        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
648        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
649        (*THEN).
650    
651    3.  If \s appeared in a character class, it removed the VT character from
652        the class, even if it had been included by some previous item, for example
653        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
654        of \s, but is part of the POSIX "space" class.)
655    
656    4.  A partial match never returns an empty string (because you can always
657        match an empty string at the end of the subject); however the checking for
658        an empty string was starting at the "start of match" point. This has been
659        changed to the "earliest inspected character" point, because the returned
660        data for a partial match starts at this character. This means that, for
661        example, /(?<=abc)def/ gives a partial match for the subject "abc"
662        (previously it gave "no match").
663    
664    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
665        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
666        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
667        has an implication that the given string is incomplete (because a partial
668        match is preferred over a full match). For this reason, these items now
669        give a partial match in this situation. [Aside: previously, the one case
670        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
671        match rather than a full match, which was wrong by the old rules, but is
672        now correct.]
673    
674    6.  There was a bug in the handling of #-introduced comments, recognized when
675        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
676        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
677        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
678        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
679        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
680        places in pcre_compile().
681    
682    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
683        comments when looking ahead for named forward references to subpatterns,
684        the only newline sequence it recognized was NL. It now handles newlines
685        according to the set newline convention.
686    
687    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
688        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
689        cater for a lack of strerror(). These oversights have been fixed.
690    
691    9.  Added --match-limit and --recursion-limit to pcregrep.
692    
693    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
694    
695    11. When the -o option was used, pcregrep was setting a return code of 1, even
696        when matches were found, and --line-buffered was not being honoured.
697    
698    12. Added an optional parentheses number to the -o and --only-matching options
699        of pcregrep.
700    
701    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
702        can match an empty string. The code to do it in pcretest and pcredemo
703        needed fixing:
704    
705        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
706            only one byte after an empty string match just before CRLF (this case
707            just got forgotten; "any" and "anycrlf" were OK).
708    
709        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
710            mode when an empty string match preceded an ASCII character followed by
711            a non-ASCII character. (The code for advancing by one character rather
712            than one byte was nonsense.)
713    
714        (c) The pcredemo.c sample program did not have any code at all to handle
715            the cases when CRLF is a valid newline sequence.
716    
717    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
718        as a starting offset was within the subject string. There is now a new
719        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
720        negative or greater than the length of the string. In order to test this,
721        pcretest is extended to allow the setting of negative starting offsets.
722    
723    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
724        starting offset points to the beginning of a UTF-8 character was
725        unnecessarily clumsy. I tidied it up.
726    
727    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
728        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
729    
730    17. Nobody had reported that the --include_dir option, which was added in
731        release 7.7 should have been called --include-dir (hyphen, not underscore)
732        for compatibility with GNU grep. I have changed it to --include-dir, but
733        left --include_dir as an undocumented synonym, and the same for
734        --exclude-dir, though that is not available in GNU grep, at least as of
735        release 2.5.4.
736    
737    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
738        characters from a string of bytes) have been redefined so as not to use
739        loops, in order to improve performance in some environments. At the same
740        time, I abstracted some of the common code into auxiliary macros to save
741        repetition (this should not affect the compiled code).
742    
743    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
744        compile-time error is now given if \c is not followed by an ASCII
745        character, that is, a byte less than 128. (In EBCDIC mode, the code is
746        different, and any byte value is allowed.)
747    
748    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
749        START_OPTIMIZE option, which is now allowed at compile time - but just
750        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
751        to pcregrep and other applications that have no direct access to PCRE
752        options. The new /Y option in pcretest sets this option when calling
753        pcre_compile().
754    
755    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
756        back references. Groups containing recursive back references were forced to
757        be atomic by that change, but in the case of named groups, the amount of
758        memory required was incorrectly computed, leading to "Failed: internal
759        error: code overflow". This has been fixed.
760    
761    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
762        pcretest.c, to avoid build problems in some Borland environments.
763    
764    
765    Version 8.10 25-Jun-2010
766    ------------------------
767    
768    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
769        THEN.
770    
771    2.  (*ACCEPT) was not working when inside an atomic group.
772    
773    3.  Inside a character class, \B is treated as a literal by default, but
774        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
775        causes the error). The code is unchanged, but I tidied the documentation.
776    
777    4.  Inside a character class, PCRE always treated \R and \X as literals,
778        whereas Perl faults them if its -w option is set. I have changed PCRE so
779        that it faults them when PCRE_EXTRA is set.
780    
781    5.  Added support for \N, which always matches any character other than
782        newline. (It is the same as "." when PCRE_DOTALL is not set.)
783    
784    6.  When compiling pcregrep with newer versions of gcc which may have
785        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
786        declared with attribute warn_unused_result" were given. Just casting the
787        result to (void) does not stop the warnings; a more elaborate fudge is
788        needed. I've used a macro to implement this.
789    
790    7.  Minor change to pcretest.c to avoid a compiler warning.
791    
792    8.  Added four artifical Unicode properties to help with an option to make
793        \s etc use properties (see next item). The new properties are: Xan
794        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
795    
796    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
797        use Unicode properties. (*UCP) at the start of a pattern can be used to set
798        this option. Modified pcretest to add /W to test this facility. Added
799        REG_UCP to make it available via the POSIX interface.
800    
801    10. Added --line-buffered to pcregrep.
802    
803    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
804        studied, and the match started with a letter with a code point greater than
805        127 whose first byte was different to the first byte of the other case of
806        the letter, the other case of this starting letter was not recognized
807        (#976).
808    
809    12. If a pattern that was studied started with a repeated Unicode property
810        test, for example, \p{Nd}+, there was the theoretical possibility of
811        setting up an incorrect bitmap of starting bytes, but fortunately it could
812        not have actually happened in practice until change 8 above was made (it
813        added property types that matched character-matching opcodes).
814    
815    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
816        possible starting bytes for non-anchored patterns.
817    
818    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
819        \R, and also a number of cases that involve Unicode properties, both
820        explicit and implicit when PCRE_UCP is set.
821    
822    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
823        input, it could crash or give wrong results if characters with values
824        greater than 0xc0 were present in the subject string. (Detail: it assumed
825        UTF-8 input when processing these items.)
826    
827    16. Added a lot of (int) casts to avoid compiler warnings in systems where
828        size_t is 64-bit (#991).
829    
830    17. Added a check for running out of memory when PCRE is compiled with
831        --disable-stack-for-recursion (#990).
832    
833    18. If the last data line in a file for pcretest does not have a newline on
834        the end, a newline was missing in the output.
835    
836    19. The default pcre_chartables.c file recognizes only ASCII characters (values
837        less than 128) in its various bitmaps. However, there is a facility for
838        generating tables according to the current locale when PCRE is compiled. It
839        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
840        characters, are recognized by isspace() and therefore were getting set in
841        these tables, and indeed these tables seem to approximate to ISO 8859. This
842        caused a problem in UTF-8 mode when pcre_study() was used to create a list
843        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
844        which of course cannot start UTF-8 characters. I have changed the code so
845        that only real ASCII characters (less than 128) and the correct starting
846        bytes for UTF-8 encodings are set for characters greater than 127 when in
847        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
848        altogether.)
849    
850    20. Added the /T option to pcretest so as to be able to run tests with non-
851        standard character tables, thus making it possible to include the tests
852        used for 19 above in the standard set of tests.
853    
854    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
855        reference to a subpattern the other side of a comment that contains an
856        opening parenthesis caused either an internal compiling error, or a
857        reference to the wrong subpattern.
858    
859    
860    Version 8.02 19-Mar-2010
861    ------------------------
862    
863    1.  The Unicode data tables have been updated to Unicode 5.2.0.
864    
865    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
866        configured.
867    
868    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
869        original author of that file, following a query about its status.
870    
871    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
872        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
873    
874    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
875        quantifier applied to a forward-referencing subroutine call, could compile
876        incorrect code or give the error "internal error: previously-checked
877        referenced subpattern not found".
878    
879    6.  Both MS Visual Studio and Symbian OS have problems with initializing
880        variables to point to external functions. For these systems, therefore,
881        pcre_malloc etc. are now initialized to local functions that call the
882        relevant global functions.
883    
884    7.  There were two entries missing in the vectors called coptable and poptable
885        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
886        I've fixed the data, and added a kludgy way of testing at compile time that
887        the lengths are correct (equal to the number of opcodes).
888    
889    8.  Following on from 7, I added a similar kludge to check the length of the
890        eint vector in pcreposix.c.
891    
892    9.  Error texts for pcre_compile() are held as one long string to avoid too
893        much relocation at load time. To find a text, the string is searched,
894        counting zeros. There was no check for running off the end of the string,
895        which could happen if a new error number was added without updating the
896        string.
897    
898    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
899    
900    11. \K was not working if it appeared in an atomic group or in a group that
901        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
902        \K is "not well defined" if used in an assertion. PCRE now accepts it if
903        the assertion is positive, but not if it is negative.
904    
905    12. Change 11 fortuitously reduced the size of the stack frame used in the
906        "match()" function of pcre_exec.c by one pointer. Forthcoming
907        implementation of support for (*MARK) will need an extra pointer on the
908        stack; I have reserved it now, so that the stack frame size does not
909        decrease.
910    
911    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
912        item in branch that calls a recursion is a subroutine call - as in the
913        second branch in the above example - was incorrectly given the compile-
914        time error "recursive call could loop indefinitely" because pcre_compile()
915        was not correctly checking the subroutine for matching a non-empty string.
916    
917    14. The checks for overrunning compiling workspace could trigger after an
918        overrun had occurred. This is a "should never occur" error, but it can be
919        triggered by pathological patterns such as hundreds of nested parentheses.
920        The checks now trigger 100 bytes before the end of the workspace.
921    
922    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
923    
924    
925    Version 8.01 19-Jan-2010
926    ------------------------
927    
928    1.  If a pattern contained a conditional subpattern with only one branch (in
929        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
930        computed the wrong minimum data length (which is of course zero for such
931        subpatterns). This could cause incorrect "no match" results.
932    
933    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
934        the pattern is reset in the first branch, pcre_compile() failed with
935        "internal error: code overflow at offset...". This happened only when
936        the reset was to the original external option setting. (An optimization
937        abstracts leading options settings into an external setting, which was the
938        cause of this.)
939    
940    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
941        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
942        assertion pattern did not match (meaning that the assertion was true), it
943        was incorrectly treated as false if the SKIP had been reached during the
944        matching. This also applied to assertions used as conditions.
945    
946    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
947        assertion subpattern, including such a pattern used as a condition,
948        unpredictable results occurred, instead of the error return
949        PCRE_ERROR_DFA_UITEM.
950    
951    5.  The C++ GlobalReplace function was not working like Perl for the special
952        situation when an empty string is matched. It now does the fancy magic
953        stuff that is necessary.
954    
955    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
956        removed. (These were left over from very, very early versions of PCRE.)
957    
958    7.  Some cosmetic changes to the code to make life easier when compiling it
959        as part of something else:
960    
961        (a) Change DEBUG to PCRE_DEBUG.
962    
963        (b) In pcre_compile(), rename the member of the "branch_chain" structure
964            called "current" as "current_branch", to prevent a collision with the
965            Linux macro when compiled as a kernel module.
966    
967        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
968            prevent a collision with the Linux macro when compiled as a kernel
969            module.
970    
971    8.  In pcre_compile() there are some checks for integer overflows that used to
972        cast potentially large values to (double). This has been changed to that
973        when building, a check for int64_t is made, and if it is found, it is used
974        instead, thus avoiding the use of floating point arithmetic. (There is no
975        other use of FP in PCRE.) If int64_t is not found, the fallback is to
976        double.
977    
978    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
979        2005 (difference between two addresses compared to an unsigned value).
980    
981    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
982        custom one, because of the following reported problem in Windows:
983    
984          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
985              under Win32.
986          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
987              therefore missing the function definition.
988          - The compiler thus generates a "C" signature for the test function.
989          - The linker fails to find the "C" function.
990          - PCRE fails to configure if asked to do so against libbz2.
991    
992    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
993        messages were output:
994    
995          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
996          rerunning libtoolize, to keep the correct libtool macros in-tree.
997          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
998    
999        I have done both of these things.
1000    
1001    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1002        most of the time, it *can* run out if it is given a pattern that contains a
1003        runaway infinite recursion. I updated the discussion in the pcrestack man
1004        page.
1005    
1006    13. Now that we have gone to the x.xx style of version numbers, the minor
1007        version may start with zero. Using 08 or 09 is a bad idea because users
1008        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1009        interpreted as invalid octal numbers. I've updated the previous comment in
1010        configure.ac, and also added a check that gives an error if 08 or 09 are
1011        used.
1012    
1013    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1014        causing partial matching to fail when the end of the subject matched \W
1015        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1016    
1017    15. There were some discrepancies between the declarations in pcre_internal.h
1018        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1019        their definitions. The declarations used "const uschar *" and the
1020        definitions used USPTR. Even though USPTR is normally defined as "const
1021        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1022        reported that: "This difference in casting confuses some C++ compilers, for
1023        example, SunCC recognizes above declarations as different functions and
1024        generates broken code for hbpcre." I have changed the declarations to use
1025        USPTR.
1026    
1027    16. GNU libtool is named differently on some systems. The autogen.sh script now
1028        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1029        (FreeBSD).
1030    
1031    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1032        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1033        comment: "Figure out how to create a longlong from a string: strtoll and
1034        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1035        instance, but it only takes 2 args instead of 3!"
1036    
1037    18. A subtle bug concerned with back references has been fixed by a change of
1038        specification, with a corresponding code fix. A pattern such as
1039        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1040        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1041        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1042        same bug. Such groups have to be quantified to be useful, or contained
1043        inside another quantified group. (If there's no repetition, the reference
1044        can never match.) The problem arises because, having left the group and
1045        moved on to the rest of the pattern, a later failure that backtracks into
1046        the group uses the captured value from the final iteration of the group
1047        rather than the correct earlier one. I have fixed this in PCRE by forcing
1048        any group that contains a reference to itself to be an atomic group; that
1049        is, there cannot be any backtracking into it once it has completed. This is
1050        similar to recursive and subroutine calls.
1051    
1052    
1053    Version 8.00 19-Oct-09
1054    ----------------------
1055    
1056    1.  The table for translating pcre_compile() error codes into POSIX error codes
1057        was out-of-date, and there was no check on the pcre_compile() error code
1058        being within the table. This could lead to an OK return being given in
1059        error.
1060    
1061    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1062        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1063        in a Windows environment.
1064    
1065    3.  The pcregrep --count option prints the count for each file even when it is
1066        zero, as does GNU grep. However, pcregrep was also printing all files when
1067        --files-with-matches was added. Now, when both options are given, it prints
1068        counts only for those files that have at least one match. (GNU grep just
1069        prints the file name in this circumstance, but including the count seems
1070        more useful - otherwise, why use --count?) Also ensured that the
1071        combination -clh just lists non-zero counts, with no names.
1072    
1073    4.  The long form of the pcregrep -F option was incorrectly implemented as
1074        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1075        but it seems right to fix it, and I didn't think it was worth preserving
1076        the old behaviour.
1077    
1078    5.  The command line items --regex=pattern and --regexp=pattern were not
1079        recognized by pcregrep, which required --regex pattern or --regexp pattern
1080        (with a space rather than an '='). The man page documented the '=' forms,
1081        which are compatible with GNU grep; these now work.
1082    
1083    6.  No libpcreposix.pc file was created for pkg-config; there was just
1084        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1085    
1086    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1087        when UCP support is not needed, by modifying the Python script that
1088        generates it from Unicode data files. This should not matter if the module
1089        is correctly used as a library, but I received one complaint about 50K of
1090        unwanted data. My guess is that the person linked everything into his
1091        program rather than using a library. Anyway, it does no harm.
1092    
1093    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1094        was a minimum greater than 1 for a wide character in a possessive
1095        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1096        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1097        character. Chaos in the form of incorrect output or a compiling loop could
1098        result.
1099    
1100    9.  The restrictions on what a pattern can contain when partial matching is
1101        requested for pcre_exec() have been removed. All patterns can now be
1102        partially matched by this function. In addition, if there are at least two
1103        slots in the offset vector, the offset of the earliest inspected character
1104        for the match and the offset of the end of the subject are set in them when
1105        PCRE_ERROR_PARTIAL is returned.
1106    
1107    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1108        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1109        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1110        and may be more useful for multi-segment matching.
1111    
1112    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1113        used to be given if ever the end of the subject was reached; now it is
1114        given only if matching could not proceed because another character was
1115        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1116        string "Z", which now yields "no match" instead of "partial match". In the
1117        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1118        final character ended with (*FAIL).
1119    
1120    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1121        if the pattern had a "must contain" character that was already found in the
1122        earlier partial match, unless partial matching was again requested. For
1123        example, with the pattern /dog.(body)?/, the "must contain" character is
1124        "g". If the first part-match was for the string "dog", restarting with
1125        "sbody" failed. This bug has been fixed.
1126    
1127    13. The string returned by pcre_dfa_exec() after a partial match has been
1128        changed so that it starts at the first inspected character rather than the
1129        first character of the match. This makes a difference only if the pattern
1130        starts with a lookbehind assertion or \b or \B (\K is not supported by
1131        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1132        matching functions compatible, and I think it's the right thing to do.
1133    
1134    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1135        so that the demonstration program is easily available in environments where
1136        PCRE has not been installed from source.
1137    
1138    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1139        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1140        library.
1141    
1142    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1143        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1144        is not the first non-POSIX option to be added. Clearly some people find
1145        these options useful.
1146    
1147    17. If a caller to the POSIX matching function regexec() passes a non-zero
1148        value for nmatch with a NULL value for pmatch, the value of
1149        nmatch is forced to zero.
1150    
1151    18. RunGrepTest did not have a test for the availability of the -u option of
1152        the diff command, as RunTest does. It now checks in the same way as
1153        RunTest, and also checks for the -b option.
1154    
1155    19. If an odd number of negated classes containing just a single character
1156        interposed, within parentheses, between a forward reference to a named
1157        subpattern and the definition of the subpattern, compilation crashed with
1158        an internal error, complaining that it could not find the referenced
1159        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1160        [The bug was that it was starting one character too far in when skipping
1161        over the character class, thus treating the ] as data rather than
1162        terminating the class. This meant it could skip too much.]
1163    
1164    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1165        /g option in pcretest when the pattern contains \K, which makes it possible
1166        to have an empty string match not at the start, even when the pattern is
1167        anchored. Updated pcretest and pcredemo to use this option.
1168    
1169    21. If the maximum number of capturing subpatterns in a recursion was greater
1170        than the maximum at the outer level, the higher number was returned, but
1171        with unset values at the outer level. The correct (outer level) value is
1172        now given.
1173    
1174    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1175        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1176        make it do so. The string so far is captured, making this feature
1177        compatible with Perl.
1178    
1179    23. The tests have been re-organized, adding tests 11 and 12, to make it
1180        possible to check the Perl 5.10 features against Perl 5.10.
1181    
1182    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1183        pattern matches a fixed length string. PCRE did not allow this; now it
1184        does. Neither allows recursion.
1185    
1186    25. I finally figured out how to implement a request to provide the minimum
1187        length of subject string that was needed in order to match a given pattern.
1188        (It was back references and recursion that I had previously got hung up
1189        on.) This code has now been added to pcre_study(); it finds a lower bound
1190        to the length of subject needed. It is not necessarily the greatest lower
1191        bound, but using it to avoid searching strings that are too short does give
1192        some useful speed-ups. The value is available to calling programs via
1193        pcre_fullinfo().
1194    
1195    26. While implementing 25, I discovered to my embarrassment that pcretest had
1196        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1197        study optimizations had never been tested with that matching function.
1198        Oops. What is worse, even when it was passed study data, there was a bug in
1199        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1200        were also very few tests of studied patterns with pcre_dfa_exec().
1201    
1202    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1203        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1204        on the other side of the coin, they are no longer allowed to have different
1205        names, because these cannot be distinguished in PCRE, and this has caused
1206        confusion. (This is a difference from Perl.)
1207    
1208    28. When duplicate subpattern names are present (necessarily with different
1209        numbers, as required by 27 above), and a test is made by name in a
1210        conditional pattern, either for a subpattern having been matched, or for
1211        recursion in such a pattern, all the associated numbered subpatterns are
1212        tested, and the overall condition is true if the condition is true for any
1213        one of them. This is the way Perl works, and is also more like the way
1214        testing by number works.
1215    
1216    
1217    Version 7.9 11-Apr-09
1218    ---------------------
1219    
1220    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1221        (pcretest), all targets were linked against these libraries. This included
1222        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1223        libraries. This caused unwanted dependencies to be created. This problem
1224        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1225        pcretest is linked with readline.
1226    
1227    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1228        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1229        moved outside it again, because FALSE and TRUE are already defined in AIX,
1230        but BOOL is not.
1231    
1232    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1233        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1234    
1235    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1236        hyphens) following file names and line numbers when outputting matching
1237        lines. This is not true; no spaces are inserted. I have also clarified the
1238        wording for the --colour (or --color) option.
1239    
1240    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1241        was not coloured; this is different to GNU grep, so I have changed it to be
1242        the same.
1243    
1244    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1245        each matching line was coloured. Now it goes on to look for further matches
1246        of any of the test patterns, which is the same behaviour as GNU grep.
1247    
1248    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1249        doesn't make sense to accept an empty string match in pcregrep, so I have
1250        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1251        seems to be how GNU grep behaves.
1252    
1253    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1254        start or after a newline", because the conditional assertion was not being
1255        correctly handled. The rule now is that both the assertion and what follows
1256        in the first alternative must satisfy the test.
1257    
1258    9.  If auto-callout was enabled in a pattern with a conditional group whose
1259        condition was an assertion, PCRE could crash during matching, both with
1260        pcre_exec() and pcre_dfa_exec().
1261    
1262    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1263        used for matching.
1264    
1265    11. Unicode property support in character classes was not working for
1266        characters (bytes) greater than 127 when not in UTF-8 mode.
1267    
1268    12. Added the -M command line option to pcretest.
1269    
1270    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1271    
1272    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1273    
1274    16. Added comments and documentation about mis-use of no_arg in the C++
1275        wrapper.
1276    
1277    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1278        from Martin Jerabek that uses macro names for all relevant character and
1279        string constants.
1280    
1281    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1282        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1283        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1284        these, but not everybody uses configure.
1285    
1286    19. A conditional group that had only one branch was not being correctly
1287        recognized as an item that could match an empty string. This meant that an
1288        enclosing group might also not be so recognized, causing infinite looping
1289        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1290        with the subject "ab", where knowledge that the repeated group can match
1291        nothing is needed in order to break the loop.
1292    
1293    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1294        exec(), but without supplying a callout function, matching went wrong.
1295    
1296    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1297        leak if the size of the offset vector was greater than 30. When the vector
1298        is smaller, the saved offsets during recursion go onto a local stack
1299        vector, but for larger vectors malloc() is used. It was failing to free
1300        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1301        error, in fact).
1302    
1303    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1304        heapframe that is used only when UTF-8 support is enabled. This caused no
1305        problem, but was untidy.
1306    
1307    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1308        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1309        included within another project.
1310    
1311    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1312        slightly modified by me:
1313    
1314          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1315              not building pcregrep.
1316    
1317          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1318              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1319    
1320    25. Forward references, both numeric and by name, in patterns that made use of
1321        duplicate group numbers, could behave incorrectly or give incorrect errors,
1322        because when scanning forward to find the reference group, PCRE was not
1323        taking into account the duplicate group numbers. A pattern such as
1324        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1325    
1326    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1327        the feature of a custom pointer more persuasive (as requested by a user).
1328    
1329    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1330        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1331        pre-defined.
1332    
1333    28. Added support for (*UTF8) at the start of a pattern.
1334    
1335    29. Arrange for flags added by the "release type" setting in CMake to be shown
1336        in the configuration summary.
1337    
1338    
1339    Version 7.8 05-Sep-08
1340    ---------------------
1341    
1342    1.  Replaced UCP searching code with optimized version as implemented for Ad
1343        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1344        stage table and inline lookup instead of a function, giving speed ups of 2
1345        to 5 times on some simple patterns that I tested. Permission was given to
1346        distribute the MultiStage2.py script that generates the tables (it's not in
1347        the tarball, but is in the Subversion repository).
1348    
1349    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1350        scripts.
1351    
1352    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1353        a group with a zero qualifier. The result of the study could be incorrect,
1354        or the function might crash, depending on the pattern.
1355    
1356    4.  Caseless matching was not working for non-ASCII characters in back
1357        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1358        It now works when Unicode Property Support is available.
1359    
1360    5.  In pcretest, an escape such as \x{de} in the data was always generating
1361        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1362        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1363        truncation.
1364    
1365    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1366    
1367    7.  Added two (int) casts to pcregrep when printing the difference of two
1368        pointers, in case they are 64-bit values.
1369    
1370    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1371        test 2 if it fails.
1372    
1373    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1374        and a #define of that name to empty if it is not externally set. This is to
1375        allow users of MSVC to set it if necessary.
1376    
1377    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1378        the convenience functions in the pcre_get.c source file.
1379    
1380    11. An option change at the start of a pattern that had top-level alternatives
1381        could cause overwriting and/or a crash. This command provoked a crash in
1382        some environments:
1383    
1384          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1385    
1386        This potential security problem was recorded as CVE-2008-2371.
1387    
1388    12. For a pattern where the match had to start at the beginning or immediately
1389        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1390        pcre_dfa_exec() could read past the end of the passed subject if there was
1391        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1392        pcretest so that it places the subject at the end of its malloc-ed buffer.
1393    
1394    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1395        exec() might read past the end of the data buffer in UTF-8 mode.
1396    
1397    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1398        the data contained the byte 0x85 as part of a UTF-8 character within its
1399        first line. This applied both to normal and DFA matching.
1400    
1401    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1402        /^[^d]*?$/8 failed to match "abc".
1403    
1404    16. Added a missing copyright notice to pcrecpp_internal.h.
1405    
1406    17. Make it more clear in the documentation that values returned from
1407        pcre_exec() in ovector are byte offsets, not character counts.
1408    
1409    18. Tidied a few places to stop certain compilers from issuing warnings.
1410    
1411    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1412        supplied by Stefan Weber. I made a further small update for 7.8 because
1413        there is a change of source arrangements: the pcre_searchfuncs.c module is
1414        replaced by pcre_ucd.c.
1415    
1416    
1417    Version 7.7 07-May-08
1418    ---------------------
1419    
1420    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1421        a string to a long long, pretend we don't even have a long long." This is
1422        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1423    
1424    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1425        pre-7.6 versions, which defined a global no_arg variable instead of putting
1426        it in the RE class. (See also #8 below.)
1427    
1428    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1429        Lopes.
1430    
1431    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1432    
1433        (1) The include/exclude patterns were being applied to the whole pathnames
1434            of files, instead of just to the final components.
1435    
1436        (2) If there was more than one level of directory, the subdirectories were
1437            skipped unless they satisfied the include/exclude conditions. This is
1438            inconsistent with GNU grep (and could even be seen as contrary to the
1439            pcregrep specification - which I improved to make it absolutely clear).
1440            The action now is always to scan all levels of directory, and just
1441            apply the include/exclude patterns to regular files.
1442    
1443    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1444        --exclude_dir in the tests to avoid scanning .svn directories.
1445    
1446    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1447        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1448        doesn't support NULs in patterns.
1449    
1450    7.  Added some missing "const"s to declarations of static tables in
1451        pcre_compile.c and pcre_dfa_exec.c.
1452    
1453    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1454        caused by fix #2  above. (Subsequently also a second patch to fix the
1455        first patch. And a third patch - this was a messy problem.)
1456    
1457    9.  Applied Craig's patch to remove the use of push_back().
1458    
1459    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1460        matching function regexec().
1461    
1462    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1463        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1464        references. PCRE supports relative numbers with this syntax (I don't think
1465        Oniguruma does).
1466    
1467    12. Previously, a group with a zero repeat such as (...){0} was completely
1468        omitted from the compiled regex. However, this means that if the group
1469        was called as a subroutine from elsewhere in the pattern, things went wrong
1470        (an internal error was given). Such groups are now left in the compiled
1471        pattern, with a new opcode that causes them to be skipped at execution
1472        time.
1473    
1474    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1475        to the way PCRE behaves:
1476    
1477        (a) A lone ] character is dis-allowed (Perl treats it as data).
1478    
1479        (b) A back reference to an unmatched subpattern matches an empty string
1480            (Perl fails the current match path).
1481    
1482        (c) A data ] in a character class must be notated as \] because if the
1483            first data character in a class is ], it defines an empty class. (In
1484            Perl it is not possible to have an empty class.) The empty class []
1485            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1486            The negative empty class [^] matches any one character, independently
1487            of the DOTALL setting.
1488    
1489    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1490        non-existent subpattern following a character class starting with ']' and
1491        containing () gave an internal compiling error instead of "reference to
1492        non-existent subpattern". Fortunately, when the pattern did exist, the
1493        compiled code was correct. (When scanning forwards to check for the
1494        existencd of the subpattern, it was treating the data ']' as terminating
1495        the class, so got the count wrong. When actually compiling, the reference
1496        was subsequently set up correctly.)
1497    
1498    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1499        it was being rejected as not supported by pcre_dfa_exec(), even though
1500        other assertions are supported. I have made pcre_dfa_exec() support
1501        (*FAIL).
1502    
1503    16. The implementation of 13c above involved the invention of a new opcode,
1504        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1505        cannot be changed at match time, I realized I could make a small
1506        improvement to matching performance by compiling OP_ALLANY instead of
1507        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1508        on the OP_ANY path.
1509    
1510    17. Compiling pcretest on Windows with readline support failed without the
1511        following two fixes: (1) Make the unistd.h include conditional on
1512        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1513    
1514    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1515        ncurses library to be included for pcretest when ReadLine support is
1516        requested, but also to allow for it to be overridden. This patch came from
1517        Daniel Bergström.
1518    
1519    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1520        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1521        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1522        spotting this.
1523    
1524    
1525    Version 7.6 28-Jan-08
1526    ---------------------
1527    
1528    1.  A character class containing a very large number of characters with
1529        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1530        overflow.
1531    
1532    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1533        HAVE_LONG_LONG is not defined.
1534    
1535    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1536        bring them up to date and include new features. This patch includes:
1537    
1538        - Fixed PH's badly added libz and libbz2 support.
1539        - Fixed a problem with static linking.
1540        - Added pcredemo. [But later removed - see 7 below.]
1541        - Fixed dftables problem and added an option.
1542        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1543            HAVE_LONG_LONG.
1544        - Added readline support for pcretest.
1545        - Added an listing of the option settings after cmake has run.
1546    
1547    4.  A user submitted a patch to Makefile that makes it easy to create
1548        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1549        Makefile.am that cause it to include this special target, without
1550        affecting anything else. Note that the same mingw target plus all
1551        the other distribution libraries and programs are now supported
1552        when configuring with CMake (see 6 below) instead of with
1553        Configure/Make.
1554    
1555    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1556        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1557        exported in the Windows port". It has not yet been confirmed that the patch
1558        solves the problem, but it does no harm.
1559    
1560    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1561        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1562        with CMake, and also correct the comment about stack recursion.
1563    
1564    7.  Remove the automatic building of pcredemo from the ./configure system and
1565        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1566        of a program that users should build themselves after PCRE is installed, so
1567        building it automatically is not really right. What is more, it gave
1568        trouble in some build environments.
1569    
1570    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1571    
1572    
1573    Version 7.5 10-Jan-08
1574    ---------------------
1575    
1576    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1577        values in parens when parsing an RE using the C++ wrapper."
1578    
1579    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1580        Characters greater than 255 were excluded from the class instead of being
1581        included.
1582    
1583    3.  The same bug as (2) above applied to negated POSIX classes such as
1584        [:^space:].
1585    
1586    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1587        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1588        I have changed it.
1589    
1590    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1591        first named subpattern) and a construct such as (?&a) would reference the
1592        first named subpattern whose name started with "a" (in other words, the
1593        length check was missing). Both these problems are fixed. "Subpattern name
1594        expected" is now given for (?&) (a zero-length name), and this patch also
1595        makes it give the same error for \k'' (previously it complained that that
1596        was a reference to a non-existent subpattern).
1597    
1598    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1599        this is right because (?- can be followed by option settings as well as by
1600        digits. I have, however, made the messages clearer.
1601    
1602    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1603        than the number used in the conditional) now cause a compile-time error.
1604        This is actually not compatible with Perl, which accepts such patterns, but
1605        treats the conditional as always being FALSE (as PCRE used to), but it
1606        seems to me that giving a diagnostic is better.
1607    
1608    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1609        and messages.
1610    
1611    9.  Fix two occurrences of "backslash" in comments that should have been
1612        "backspace".
1613    
1614    10. Remove two redundant lines of code that can never be obeyed (their function
1615        was moved elsewhere).
1616    
1617    11. The program that makes PCRE's Unicode character property table had a bug
1618        which caused it to generate incorrect table entries for sequences of
1619        characters that have the same character type, but are in different scripts.
1620        It amalgamated them into a single range, with the script of the first of
1621        them. In other words, some characters were in the wrong script. There were
1622        thirteen such cases, affecting characters in the following ranges:
1623    
1624          U+002b0 - U+002c1
1625          U+0060c - U+0060d
1626          U+0061e - U+00612
1627          U+0064b - U+0065e
1628          U+0074d - U+0076d
1629          U+01800 - U+01805
1630          U+01d00 - U+01d77
1631          U+01d9b - U+01dbf
1632          U+0200b - U+0200f
1633          U+030fc - U+030fe
1634          U+03260 - U+0327f
1635          U+0fb46 - U+0fbb1
1636          U+10450 - U+1049d
1637    
1638    12. The -o option (show only the matching part of a line) for pcregrep was not
1639        compatible with GNU grep in that, if there was more than one match in a
1640        line, it showed only the first of them. It now behaves in the same way as
1641        GNU grep.
1642    
1643    13. If the -o and -v options were combined for pcregrep, it printed a blank
1644        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1645        does the same. The return code can be used to tell if there were any
1646        non-matching lines.
1647    
1648    14. Added --file-offsets and --line-offsets to pcregrep.
1649    
1650    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1651        infinitely looping recursion. The bug was that positive lookaheads were not
1652        being skipped when checking for a possible empty match (negative lookaheads
1653        and both kinds of lookbehind were skipped).
1654    
1655    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1656        inclusion of <windows.h> to before rather than after the definition of
1657        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1658    
1659    17. Specifying a possessive quantifier with a specific limit for a Unicode
1660        character property caused pcre_compile() to compile bad code, which led at
1661        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1662        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1663        caused the error; without that there was no problem.
1664    
1665    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1666    
1667    19. Added --enable-pcretest-libreadline.
1668    
1669    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1670        RE::GlobalReplace(). As a result, the number of replacements returned was
1671        double what it should be. I removed one of the increments, but Craig sent a
1672        later patch that removed the other one (the right fix) and added unit tests
1673        that check the return values (which was not done before).
1674    
1675    21. Several CMake things:
1676    
1677        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1678            the names libpcre and libpcreposix, not just pcre and pcreposix.
1679    
1680        (2) The above change means that pcretest and pcregrep are now correctly
1681            linked with the newly-built libraries, not previously installed ones.
1682    
1683        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1684    
1685    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1686        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1687        UTF-8 newline character). The key issue is that the pattern starts .*;
1688        this means that the match must be either at the beginning, or after a
1689        newline. The bug was in the code for advancing after a failed match and
1690        checking that the new position followed a newline. It was not taking
1691        account of UTF-8 characters correctly.
1692    
1693    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1694        character classes. PCRE was not treating the sequence [:...:] as a
1695        character class unless the ... were all letters. Perl, however, seems to
1696        allow any characters between [: and :], though of course it rejects as
1697        unknown any "names" that contain non-letters, because all the known class
1698        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1699        for example, whereas PCRE did not - it did not recognize a POSIX character
1700        class. This seemed a bit dangerous, so the code has been changed to be
1701        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1702        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1703        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1704        Perl does, and where it didn't before.
1705    
1706    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1707        Windows environments %n is disabled by default.
1708    
1709    
1710    Version 7.4 21-Sep-07
1711    ---------------------
1712    
1713    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1714        means that a class such as [\s] counted as "explicit reference to CR or
1715        LF". That isn't really right - the whole point of the change was to try to
1716        help when there was an actual mention of one of the two characters. So now
1717        the change happens only if \r or \n (or a literal CR or LF) character is
1718        encountered.
1719    
1720    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1721        of both had grown to the point where there were only 3 bits left.
1722        Fortunately, there was spare space in the data structure, and so I have
1723        moved the internal flags into a new 16-bit field to free up more option
1724        bits.
1725    
1726    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1727        but did not set the internal JCHANGED flag - either of these is enough to
1728        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1729        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1730        start sets both bits.
1731    
1732    4.  Added options (at build time, compile time, exec time) to change \R from
1733        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1734    
1735    5.  doc/pcresyntax.html was missing from the distribution.
1736    
1737    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1738        compatibility, even though it is no longer used.
1739    
1740    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1741        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1742        windows.h file is present (where different names are used). [This was
1743        reversed later after testing - see 16 below.]
1744    
1745    8.  Changed all #include <config.h> to #include "config.h". There were also
1746        some further <pcre.h> cases that I changed to "pcre.h".
1747    
1748    9.  When pcregrep was used with the --colour option, it missed the line ending
1749        sequence off the lines that it output.
1750    
1751    10. It was pointed out to me that arrays of string pointers cause lots of
1752        relocations when a shared library is dynamically loaded. A technique of
1753        using a single long string with a table of offsets can drastically reduce
1754        these. I have refactored PCRE in four places to do this. The result is
1755        dramatic:
1756    
1757          Originally:                          290
1758          After changing UCP table:            187
1759          After changing error message table:   43
1760          After changing table of "verbs"       36
1761          After changing table of Posix names   22
1762    
1763        Thanks to the folks working on Gregex for glib for this insight.
1764    
1765    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1766        unicode-properties was also set.
1767    
1768    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1769    
1770    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1771        checked only for CRLF.
1772    
1773    14. Added casts to pcretest.c to avoid compiler warnings.
1774    
1775    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1776    
1777    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1778        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1779        entirely. This removes changes made in 7 above.
1780    
1781    17. The CMake files have been updated, and there is now more information about
1782        building with CMake in the NON-UNIX-USE document.
1783    
1784    
1785    Version 7.3 28-Aug-07
1786    ---------------------
1787    
1788     1. In the rejigging of the build system that eventually resulted in 7.1, the
1789        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1790        brackets there is not right, since it causes compilers to look for an
1791        installed pcre.h, not the version that is in the source that is being
1792        compiled (which of course may be different). I have changed it back to:
1793    
1794          #include "pcre.h"
1795    
1796        I have a vague recollection that the change was concerned with compiling in
1797        different directories, but in the new build system, that is taken care of
1798        by the VPATH setting the Makefile.
1799    
1800     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1801        when the subject happened to end in the byte 0x85 (e.g. if the last
1802        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1803        characters but of course it shouldn't be taken as a newline when it is part
1804        of another character. The bug was that, for an unlimited repeat of . in
1805        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1806        characters when looking for a newline.
1807    
1808     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1809    
1810     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1811        in debug output.
1812    
1813     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1814        long printing in the pcrecpp unittest when running under MinGW.
1815    
1816     6. ESC_K was left out of the EBCDIC table.
1817    
1818     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1819        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1820        limit also applies to "virtual nesting" when a pattern is recursive, and in
1821        this case 1000 isn't so big. I have been able to remove this limit at the
1822        expense of backing off one optimization in certain circumstances. Normally,
1823        when pcre_exec() would call its internal match() function recursively and
1824        immediately return the result unconditionally, it uses a "tail recursion"
1825        feature to save stack. However, when a subpattern that can match an empty
1826        string has an unlimited repetition quantifier, it no longer makes this
1827        optimization. That gives it a stack frame in which to save the data for
1828        checking that an empty string has been matched. Previously this was taken
1829        from the 1000-entry workspace that had been reserved. So now there is no
1830        explicit limit, but more stack is used.
1831    
1832     8. Applied Daniel's patches to solve problems with the import/export magic
1833        syntax that is required for Windows, and which was going wrong for the
1834        pcreposix and pcrecpp parts of the library. These were overlooked when this
1835        problem was solved for the main library.
1836    
1837     9. There were some crude static tests to avoid integer overflow when computing
1838        the size of patterns that contain repeated groups with explicit upper
1839        limits. As the maximum quantifier is 65535, the maximum group length was
1840        set at 30,000 so that the product of these two numbers did not overflow a
1841        32-bit integer. However, it turns out that people want to use groups that
1842        are longer than 30,000 bytes (though not repeat them that many times).
1843        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1844        made it possible to implement the integer overflow checks in a much more
1845        dynamic way, which I have now done. The artificial limitation on group
1846        length has been removed - we now have only the limit on the total length of
1847        the compiled pattern, which depends on the LINK_SIZE setting.
1848    
1849    10. Fixed a bug in the documentation for get/copy named substring when
1850        duplicate names are permitted. If none of the named substrings are set, the
1851        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1852        empty string.
1853    
1854    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1855        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1856        because the ] is interpreted as the first data character and the
1857        terminating ] is not found. PCRE has been made compatible with Perl in this
1858        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1859        cause memory overwriting.
1860    
1861    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1862        string has been matched (to stop an infinite loop). It was not recognizing
1863        a conditional subpattern that could match an empty string if that
1864        subpattern was within another subpattern. For example, it looped when
1865        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1866        condition was not nested. This bug has been fixed.
1867    
1868    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1869        past the start of the subject in the presence of bytes with the top bit
1870        set, for example "\x8aBCD".
1871    
1872    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1873        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1874    
1875    14. Optimized (?!) to (*FAIL).
1876    
1877    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1878        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1879        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1880        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1881        does: it's just the validity check that is more restrictive.
1882    
1883    16. Inserted checks for integer overflows during escape sequence (backslash)
1884        processing, and also fixed erroneous offset values for syntax errors during
1885        backslash processing.
1886    
1887    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1888        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1889    
1890    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1891        caused an overrun.
1892    
1893    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1894        something other than just ASCII characters) inside a group that had an
1895        unlimited repeat caused a loop at compile time (while checking to see
1896        whether the group could match an empty string).
1897    
1898    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1899        [\P{Any}] did so. (Error in the code for printing property names.)
1900    
1901    21. An orphan \E inside a character class could cause a crash.
1902    
1903    22. A repeated capturing bracket such as (A)? could cause a wild memory
1904        reference during compilation.
1905    
1906    23. There are several functions in pcre_compile() that scan along a compiled
1907        expression for various reasons (e.g. to see if it's fixed length for look
1908        behind). There were bugs in these functions when a repeated \p or \P was
1909        present in the pattern. These operators have additional parameters compared
1910        with \d, etc, and these were not being taken into account when moving along
1911        the compiled data. Specifically:
1912    
1913        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1914            length.
1915    
1916        (b) An item such as \pL+ within a repeated group could cause crashes or
1917            loops.
1918    
1919        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1920            "reference to non-existent subpattern" error.
1921    
1922        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1923    
1924    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1925        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1926    
1927    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1928    
1929    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1930        character were causing crashes (broken optimization).
1931    
1932    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1933        \p or \P) caused a compile-time loop.
1934    
1935    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1936        break. For example, the unstudied pattern [\r\n]A does not match the string
1937        "\r\nA" because change 7.0/46 below moves the current point on by two
1938        characters after failing to match at the start. However, the pattern \nA
1939        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1940        the same is true. There doesn't seem any very clean way out of this, but
1941        what I have chosen to do makes the common cases work: PCRE now takes note
1942        of whether there can be an explicit match for \r or \n anywhere in the
1943        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1944        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1945        pattern has explicit CR or LF references.
1946    
1947    29. Added (*CR) etc for changing newline setting at start of pattern.
1948    
1949    
1950    Version 7.2 19-Jun-07
1951    ---------------------
1952    
1953     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1954        which is apparently normally available under Windows.
1955    
1956     2. Re-jig the pcregrep tests with different newline settings in an attempt
1957        to make them independent of the local environment's newline setting.
1958    
1959     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1960    
1961     4. Some of the "internals" tests were previously cut out when the link size
1962        was not 2, because the output contained actual offsets. The recent new
1963        "Z" feature of pcretest means that these can be cut out, making the tests
1964        usable with all link sizes.
1965    
1966     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1967        stack recursion. This gives a massive performance boost under BSD, but just
1968        a small improvement under Linux. However, it saves one field in the frame
1969        in all cases.
1970    
1971     6. Added more features from the forthcoming Perl 5.10:
1972    
1973        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1974            recursion call. It refers to the nth most recently opened parentheses.
1975    
1976        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1977            to be opened parentheses.
1978    
1979        (c) Conditions that refer to capturing parentheses can be specified
1980            relatively, for example, (?(-2)... or (?(+3)...
1981    
1982        (d) \K resets the start of the current match so that everything before
1983            is not part of it.
1984    
1985        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1986    
1987        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1988            reference syntax.
1989    
1990        (g) (?| introduces a group in which the numbering of parentheses in each
1991            alternative starts with the same number.
1992    
1993        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1994    
1995     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1996        PCRE_INFO_JCHANGED.
1997    
1998     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1999        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2000        for detecting groups that can match an empty string.
2001    
2002     9. A pattern with a very large number of alternatives (more than several
2003        hundred) was running out of internal workspace during the pre-compile
2004        phase, where pcre_compile() figures out how much memory will be needed. A
2005        bit of new cunning has reduced the workspace needed for groups with
2006        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2007        workspace instead of running out of the 4096 that are available.
2008    
2009    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2010    
2011    11. Applied patch from Google to remove an optimization that didn't quite work.
2012        The report of the bug said:
2013    
2014          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2015          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2016          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2017    
2018    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2019        it matched the wrong number of bytes.
2020    
2021    
2022    Version 7.1 24-Apr-07
2023    ---------------------
2024    
2025     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2026        that is more "standard", making use of automake and other Autotools. There
2027        is some re-arrangement of the files and adjustment of comments consequent
2028        on this.
2029    
2030     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2031        for recursive directory scanning broke on some systems because the files
2032        are not scanned in any specific order and on different systems the order
2033        was different. A call to "sort" has been inserted into RunGrepTest for the
2034        approprate test as a short-term fix. In the longer term there may be an
2035        alternative.
2036    
2037     3. I had an email from Eric Raymond about problems translating some of PCRE's
2038        man pages to HTML (despite the fact that I distribute HTML pages, some
2039        people do their own conversions for various reasons). The problems
2040        concerned the use of low-level troff macros .br and .in. I have therefore
2041        removed all such uses from the man pages (some were redundant, some could
2042        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2043        HTML has been updated to handle .nf/.fi and to complain if it encounters
2044        .br or .in.
2045    
2046     4. Updated comments in configure.ac that get placed in config.h.in and also
2047        arranged for config.h to be included in the distribution, with the name
2048        config.h.generic, for the benefit of those who have to compile without
2049        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2050    
2051     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2052        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2053        makevp.bat for the current PCRE, using the additional files
2054        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2055    
2056     6. A Windows user reported a minor discrepancy with test 2, which turned out
2057        to be caused by a trailing space on an input line that had got lost in his
2058        copy. The trailing space was an accident, so I've just removed it.
2059    
2060     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2061        that is needed.
2062    
2063     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2064        as "const" (a) because they are and (b) because it helps the PHP
2065        maintainers who have recently made a script to detect big data structures
2066        in the php code that should be moved to the .rodata section. I remembered
2067        to update Builducptable as well, so it won't revert if ucptable.h is ever
2068        re-created.
2069    
2070     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2071        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2072        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2073        support is not required. This saves 1.5-2K of code, which is important in
2074        some applications.
2075    
2076        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2077        so as not to refer to the tables, even though these functions will never be
2078        called when UTF-8 support is disabled. Otherwise there are problems with a
2079        shared library.
2080    
2081    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2082    
2083        (a) It was defining its arguments as char * instead of void *.
2084    
2085        (b) It was assuming that all moves were upwards in memory; this was true
2086            a long time ago when I wrote it, but is no longer the case.
2087    
2088        The emulated memove() is provided for those environments that have neither
2089        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2090        is clearly not the case, as these two bugs were recently reported.
2091    
2092    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2093        and Detrail to create the HTML documentation, the .txt form of the man
2094        pages, and it removes trailing spaces from listed files. It also creates
2095        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2096        case, it wraps all the #defines with #ifndefs. This script should be run
2097        before "make dist".
2098    
2099    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2100        with Unicode property support.
2101    
2102        (a) For a maximizing quantifier, if the two different cases of the
2103            character were of different lengths in their UTF-8 codings (there are
2104            some cases like this - I found 11), and the matching function had to
2105            back up over a mixture of the two cases, it incorrectly assumed they
2106            were both the same length.
2107    
2108        (b) When PCRE was configured to use the heap rather than the stack for
2109            recursion during matching, it was not correctly preserving the data for
2110            the other case of a UTF-8 character when checking ahead for a match
2111            while processing a minimizing repeat. If the check also involved
2112            matching a wide character, but failed, corruption could cause an
2113            erroneous result when trying to check for a repeat of the original
2114            character.
2115    
2116    13. Some tidying changes to the testing mechanism:
2117    
2118        (a) The RunTest script now detects the internal link size and whether there
2119            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2120            values substituted by "configure". (The RunGrepTest script already did
2121            this for UTF-8.) The configure.ac script no longer substitutes the
2122            relevant variables.
2123    
2124        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2125            with length and offset values. This means that the output is different
2126            for different internal link sizes. Test 2 is skipped for link sizes
2127            other than 2 because of this, bypassing the problem. Unfortunately,
2128            there was also a test in test 3 (the locale tests) that used /B and
2129            failed for link sizes other than 2. Rather than cut the whole test out,
2130            I have added a new /Z option to pcretest that replaces the length and
2131            offset values with spaces. This is now used to make test 3 independent
2132            of link size. (Test 2 will be tidied up later.)
2133    
2134    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2135        segmentation fault instead of returning the appropriate error message.
2136    
2137    15. In multiline mode when the newline sequence was set to "any", the pattern
2138        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2139        This doesn't seem right; it now treats the CRLF combination as the line
2140        ending, and so does not match in that case. It's only a pattern such as ^$
2141        that would hit this one: something like ^ABC$ would have failed after \r
2142        and then tried again after \r\n.
2143    
2144    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2145        in an attempt to make files that differ only in their line terminators
2146        compare equal. This works on Linux.
2147    
2148    17. Under certain error circumstances pcregrep might try to free random memory
2149        as it exited. This is now fixed, thanks to valgrind.
2150    
2151    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2152        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2153        was because its rules for how to advance for /g after matching an empty
2154        string at the end of a line did not allow for this case. They now check for
2155        it specially.
2156    
2157    20. pcretest is supposed to handle patterns and data of any length, by
2158        extending its buffers when necessary. It was getting this wrong when the
2159        buffer for a data line had to be extended.
2160    
2161    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2162        CRLF as a newline sequence.
2163    
2164    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2165        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2166        I have nevertheless tidied it up.
2167    
2168    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2169    
2170    24. Added a man page for pcre-config.
2171    
2172    
2173    Version 7.0 19-Dec-06
2174    ---------------------
2175    
2176     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2177        moving to gcc 4.1.1.
2178    
2179     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2180        sys/time.h, which is documented as needed for this function. It doesn't
2181        seem to matter on Linux, but it showed up on some releases of OS X.
2182    
2183     3. It seems that there are systems where bytes whose values are greater than
2184        127 match isprint() in the "C" locale. The "C" locale should be the
2185        default when a C program starts up. In most systems, only ASCII printing
2186        characters match isprint(). This difference caused the output from pcretest
2187        to vary, making some of the tests fail. I have changed pcretest so that:
2188    
2189        (a) When it is outputting text in the compiled version of a pattern, bytes
2190            other than 32-126 are always shown as hex escapes.
2191    
2192        (b) When it is outputting text that is a matched part of a subject string,
2193            it does the same, unless a different locale has been set for the match
2194            (using the /L modifier). In this case, it uses isprint() to decide.
2195    
2196     4. Fixed a major bug that caused incorrect computation of the amount of memory
2197        required for a compiled pattern when options that changed within the
2198        pattern affected the logic of the preliminary scan that determines the
2199        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2200        that the computed length was too small. The symptoms of this bug were
2201        either the PCRE error "internal error: code overflow" from pcre_compile(),
2202        or a glibc crash with a message such as "pcretest: free(): invalid next
2203        size (fast)". Examples of patterns that provoked this bug (shown in
2204        pcretest format) are:
2205    
2206          /(?-x: )/x
2207          /(?x)(?-x: \s*#\s*)/
2208          /((?i)[\x{c0}])/8
2209          /(?i:[\x{c0}])/8
2210    
2211        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2212        is now done differently.
2213    
2214     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2215        wrapper classes; (b) implement a new function in the C++ scanner that is
2216        more efficient than the old way of doing things because it avoids levels of
2217        recursion in the regex matching; (c) add a paragraph to the documentation
2218        for the FullMatch() function.
2219    
2220     6. The escape sequence \n was being treated as whatever was defined as
2221        "newline". Not only was this contrary to the documentation, which states
2222        that \n is character 10 (hex 0A), but it also went horribly wrong when
2223        "newline" was defined as CRLF. This has been fixed.
2224    
2225     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2226        was being set to -1 for the "end of line" case (supposedly a value that no
2227        character can have). Though this value is never used (the check for end of
2228        line is "zero bytes in current character"), it caused compiler complaints.
2229        I've changed it to 0xffffffff.
2230    
2231     8. In pcre_version.c, the version string was being built by a sequence of
2232        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2233        string (as it is for production releases) called a macro with an empty
2234        argument. The C standard says the result of this is undefined. The gcc
2235        compiler treats it as an empty string (which was what was wanted) but it is
2236        reported that Visual C gives an error. The source has been hacked around to
2237        avoid this problem.
2238    
2239     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2240        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2241        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2242        of them did).
2243    
2244    10. Originally, pcretest opened its input and output without "b"; then I was
2245        told that "b" was needed in some environments, so it was added for release
2246        5.0 to both the input and output. (It makes no difference on Unix-like
2247        systems.) Later I was told that it is wrong for the input on Windows. I've
2248        now abstracted the modes into two macros, to make it easier to fiddle with
2249        them, and removed "b" from the input mode under Windows.
2250    
2251    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2252    
2253    12. Added -help and --help to pcretest as an official way of being reminded
2254        of the options.
2255    
2256    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2257        and pcrecpp.cc because they annoy compilers at high warning levels.
2258    
2259    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2260    
2261    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2262        scripts are not C programs :-) and which was not noticed because it works
2263        on Linux.
2264    
2265    16. pcretest is supposed to handle any length of pattern and data line (as one
2266        line or as a continued sequence of lines) by extending its input buffer if
2267        necessary. This feature was broken for very long pattern lines, leading to
2268        a string of junk being passed to pcre_compile() if the pattern was longer
2269        than about 50K.
2270    
2271    17. I have done a major re-factoring of the way pcre_compile() computes the
2272        amount of memory needed for a compiled pattern. Previously, there was code
2273        that made a preliminary scan of the pattern in order to do this. That was
2274        OK when PCRE was new, but as the facilities have expanded, it has become
2275        harder and harder to keep it in step with the real compile phase, and there
2276        have been a number of bugs (see for example, 4 above). I have now found a
2277        cunning way of running the real compile function in a "fake" mode that
2278        enables it to compute how much memory it would need, while actually only
2279        ever using a few hundred bytes of working memory and without too many
2280        tests of the mode. This should make future maintenance and development
2281        easier. A side effect of this work is that the limit of 200 on the nesting
2282        depth of parentheses has been removed (though this was never a serious
2283        limitation, I suspect). However, there is a downside: pcre_compile() now
2284        runs more slowly than before (30% or more, depending on the pattern). I
2285        hope this isn't a big issue. There is no effect on runtime performance.
2286    
2287    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2288        newline (only possible for the last line of a file) and it was a
2289        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2290    
2291    19. Added additional timing features to pcretest. (1) The -tm option now times
2292        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2293        separate command line item, by a number that specifies the number of
2294        repeats to use when timing. The default is 50000; this gives better
2295        precision, but takes uncomfortably long for very large patterns.
2296    
2297    20. Extended pcre_study() to be more clever in cases where a branch of a
2298        subpattern has no definite first character. For example, (a*|b*)[cd] would
2299        previously give no result from pcre_study(). Now it recognizes that the
2300        first character must be a, b, c, or d.
2301    
2302    21. There was an incorrect error "recursive call could loop indefinitely" if
2303        a subpattern (or the entire pattern) that was being tested for matching an
2304        empty string contained only one non-empty item after a nested subpattern.
2305        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2306        incorrectly, because the \d was being skipped in the check.
2307    
2308    22. The pcretest program now has a new pattern option /B and a command line
2309        option -b, which is equivalent to adding /B to every pattern. This causes
2310        it to show the compiled bytecode, without the additional information that
2311        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2312        is the same as /B/I).
2313    
2314    23. A new optimization is now able automatically to treat some sequences such
2315        as a*b as a*+b. More specifically, if something simple (such as a character
2316        or a simple class like \d) has an unlimited quantifier, and is followed by
2317        something that cannot possibly match the quantified thing, the quantifier
2318        is automatically "possessified".
2319    
2320    24. A recursive reference to a subpattern whose number was greater than 39
2321        went wrong under certain circumstances in UTF-8 mode. This bug could also
2322        have affected the operation of pcre_study().
2323    
2324    25. Realized that a little bit of performance could be had by replacing
2325        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2326    
2327    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2328    
2329    27. Possessive quantifiers such as a++ were previously implemented by turning
2330        them into atomic groups such as ($>a+). Now they have their own opcodes,
2331        which improves performance. This includes the automatically created ones
2332        from 23 above.
2333    
2334    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2335        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2336        the first matched character to be a colon. This applied both to named and
2337        numbered groups.
2338    
2339    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2340    
2341    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2342        building PCRE on Windows easier, so I have included it in the distribution.
2343    
2344    31. There is now a check in pcretest against a ridiculously large number being
2345        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2346        loop, the loop is abandoned.
2347    
2348    32. Forward references to subpatterns in conditions such as (?(2)...) where
2349        subpattern 2 is defined later cause pcre_compile() to search forwards in
2350        the pattern for the relevant set of parentheses. This search went wrong
2351        when there were unescaped parentheses in a character class, parentheses
2352        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2353    
2354    33. "Subroutine" calls and backreferences were previously restricted to
2355        referencing subpatterns earlier in the regex. This restriction has now
2356        been removed.
2357    
2358    34. Added a number of extra features that are going to be in Perl 5.10. On the
2359        whole, these are just syntactic alternatives for features that PCRE had
2360        previously implemented using the Python syntax or my own invention. The
2361        other formats are all retained for compatibility.
2362    
2363        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2364            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2365            also .NET compatible.
2366    
2367        (b) A recursion or subroutine call to a named group can now be defined as
2368            (?&name) as well as (?P>name).
2369    
2370        (c) A backreference to a named group can now be defined as \k<name> or
2371            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2372            5.10, are also .NET compatible.
2373    
2374        (d) A conditional reference to a named group can now use the syntax
2375            (?(<name>) or (?('name') as well as (?(name).
2376    
2377        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2378            groups (named and numbered) that are never evaluated inline, but can be
2379            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2380            is always false. There may be only one alternative in such a group.
2381    
2382        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2383            as the simple (?(R). The condition is true only if the most recent
2384            recursion is that of the given number or name. It does not search out
2385            through the entire recursion stack.
2386    
2387        (g) The escape \gN or \g{N} has been added, where N is a positive or
2388            negative number, specifying an absolute or relative reference.
2389    
2390    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2391        some "unreachable code" warnings.
2392    
2393    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2394        things, this adds five new scripts.
2395    
2396    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2397        There were also incompatibilities regarding the handling of \Q..\E inside
2398        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2399        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2400    
2401    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2402        matches an empty string, and forcibly breaks the loop. There were bugs in
2403        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2404        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2405        separate and independent bugs (that affected different cases) have been
2406        fixed.
2407    
2408    39. Refactored the code to abolish the use of different opcodes for small
2409        capturing bracket numbers. This is a tidy that I avoided doing when I
2410        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2411        The new approach is not only tidier, it makes it possible to reduce the
2412        memory needed to fix the previous bug (38).
2413    
2414    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2415        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2416        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2417        mode.
2418    
2419    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2420        report.
2421    
2422    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2423        copy construction and assignment in the C++ wrapper.
2424    
2425    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2426        couple of bugs that could have given wrong results in the "--newline=crlf"
2427        case.
2428    
2429    44. Added a number of casts and did some reorganization of signed/unsigned int
2430        variables following suggestions from Dair Grant. Also renamed the variable
2431        "this" as "item" because it is a C++ keyword.
2432    
2433    45. Arranged for dftables to add
2434    
2435          #include "pcre_internal.h"
2436    
2437        to pcre_chartables.c because without it, gcc 4.x may remove the array
2438        definition from the final binary if PCRE is built into a static library and
2439        dead code stripping is activated.
2440    
2441    46. For an unanchored pattern, if a match attempt fails at the start of a
2442        newline sequence, and the newline setting is CRLF or ANY, and the next two
2443        characters are CRLF, advance by two characters instead of one.
2444    
2445    
2446    Version 6.7 04-Jul-06
2447    ---------------------
2448    
2449     1. In order to handle tests when input lines are enormously long, pcretest has
2450        been re-factored so that it automatically extends its buffers when
2451        necessary. The code is crude, but this _is_ just a test program. The
2452        default size has been increased from 32K to 50K.
2453    
2454     2. The code in pcre_study() was using the value of the re argument before
2455        testing it for NULL. (Of course, in any sensible call of the function, it
2456        won't be NULL.)
2457    
2458     3. The memmove() emulation function in pcre_internal.h, which is used on
2459        systems that lack both memmove() and bcopy() - that is, hardly ever -
2460        was missing a "static" storage class specifier.
2461    
2462     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2463        containing an extended class (one that cannot be represented by a bitmap
2464        because it contains high-valued characters or Unicode property items, e.g.
2465        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2466        pattern, but PCRE should not loop if you do not (it no longer does).
2467        [Detail: two cases were found: (a) a repeated subpattern containing an
2468        extended class; (b) a recursive reference to a subpattern that followed a
2469        previous extended class. It wasn't skipping over the extended class
2470        correctly when UTF-8 mode was not set.]
2471    
2472     5. A negated single-character class was not being recognized as fixed-length
2473        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2474        compile error "lookbehind assertion is not fixed length".
2475    
2476     6. The RunPerlTest auxiliary script was showing an unexpected difference
2477        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2478        write a Perl script that can interpret lines of an input file either as
2479        byte characters or as UTF-8, which is what "perltest" was being required to
2480        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2481        can't do is switch easily at run time between having the "use utf8;" pragma
2482        or not. In the end, I fudged it by using the RunPerlTest script to insert
2483        "use utf8;" explicitly for the UTF-8 tests.
2484    
2485     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2486        the end of the subject string, contrary to the documentation and to what
2487        Perl does. This was true of both matching functions. Now it matches only at
2488        the start of the subject and immediately after *internal* newlines.
2489    
2490     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2491        a pointer to an int instead of a pointer to an unsigned long int. This
2492        caused problems on 64-bit systems.
2493    
2494     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2495        instance of the 'standard' template library not being so standard".
2496    
2497    10. There was no check on the number of named subpatterns nor the maximum
2498        length of a subpattern name. The product of these values is used to compute
2499        the size of the memory block for a compiled pattern. By supplying a very
2500        long subpattern name and a large number of named subpatterns, the size
2501        computation could be caused to overflow. This is now prevented by limiting
2502        the length of names to 32 characters, and the number of named subpatterns
2503        to 10,000.
2504    
2505    11. Subpatterns that are repeated with specific counts have to be replicated in
2506        the compiled pattern. The size of memory for this was computed from the
2507        length of the subpattern and the repeat count. The latter is limited to
2508        65535, but there was no limit on the former, meaning that integer overflow
2509        could in principle occur. The compiled length of a repeated subpattern is
2510        now limited to 30,000 bytes in order to prevent this.
2511    
2512    12. Added the optional facility to have named substrings with the same name.
2513    
2514    13. Added the ability to use a named substring as a condition, using the
2515        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2516        are numbers (not recommended). Forward references are permitted.
2517    
2518    14. Added forward references in named backreferences (if you see what I mean).
2519    
2520    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2521        pattern could run off the end of the subject. For example, the pattern
2522        "(?s)(.{1,5})"8 did this with the subject "ab".
2523    
2524    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2525        PCRE_CASELESS was set when matching characters that were quantified with ?
2526        or *.
2527    
2528    17. A character class other than a single negated character that had a minimum
2529        but no maximum quantifier - for example [ab]{6,} - was not handled
2530        correctly by pce_dfa_exec(). It would match only one character.
2531    
2532    18. A valid (though odd) pattern that looked like a POSIX character
2533        class but used an invalid character after [ (for example [[,abc,]]) caused
2534        pcre_compile() to give the error "Failed: internal error: code overflow" or
2535        in some cases to crash with a glibc free() error. This could even happen if
2536        the pattern terminated after [[ but there just happened to be a sequence of
2537        letters, a binary zero, and a closing ] in the memory that followed.
2538    
2539    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2540        over the years. Originally (before any Unicode support), just the bottom 8
2541        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2542        output from "man perlunicode" includes this:
2543    
2544          The regular expression compiler produces polymorphic opcodes.  That
2545          is, the pattern adapts to the data and automatically switches to
2546          the Unicode character scheme when presented with Unicode data--or
2547          instead uses a traditional byte scheme when presented with byte
2548          data.
2549    
2550        Sadly, a wide octal escape does not cause a switch, and in a string with
2551        no other multibyte characters, these octal escapes are treated as before.
2552        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2553        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2554        Unicode string.
2555    
2556        I have not perpetrated such confusion in PCRE. Up till now, it took just
2557        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2558        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2559        translate to the appropriate multibyte character.
2560    
2561    29. Applied some refactoring to reduce the number of warnings from Microsoft
2562        and Borland compilers. This has included removing the fudge introduced
2563        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2564        a warning about an unused variable.
2565    
2566    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2567        characters since release 4.0, because Perl (from release 5.004) does not.
2568        [Or at least, is documented not to: some releases seem to be in conflict
2569        with the documentation.] However, when a pattern was studied with
2570        pcre_study() and all its branches started with \s, PCRE still included VT
2571        as a possible starting character. Of course, this did no harm; it just
2572        caused an unnecessary match attempt.
2573    
2574    22. Removed a now-redundant internal flag bit that recorded the fact that case
2575        dependency changed within the pattern. This was once needed for "required
2576        byte" processing, but is no longer used. This recovers a now-scarce options
2577        bit. Also moved the least significant internal flag bit to the most-
2578        significant bit of the word, which was not previously used (hangover from
2579        the days when it was an int rather than a uint) to free up another bit for
2580        the future.
2581    
2582    23. Added support for CRLF line endings as well as CR and LF. As well as the
2583        default being selectable at build time, it can now be changed at runtime
2584        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2585        specify that it is scanning data with non-default line endings.
2586    
2587    24. Changed the definition of CXXLINK to make it agree with the definition of
2588        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2589    
2590    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2591        recursions. This makes a big different to stack usage for some patterns.
2592    
2593    26. If a subpattern containing a named recursion or subroutine reference such
2594        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2595        the space required for the compiled pattern went wrong and gave too small a
2596        value. Depending on the environment, this could lead to "Failed: internal
2597        error: code overflow at offset 49" or "glibc detected double free or
2598        corruption" errors.
2599    
2600    27. Applied patches from Google (a) to support the new newline modes and (b) to
2601        advance over multibyte UTF-8 characters in GlobalReplace.
2602    
2603    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2604        difference for some implementation of PCRE in some Windows version.
2605    
2606    29. Added some extra testing facilities to pcretest:
2607    
2608        \q<number>   in a data line sets the "match limit" value
2609        \Q<number>   in a data line sets the "match recursion limt" value
2610        -S <number>  sets the stack size, where <number> is in megabytes
2611    
2612        The -S option isn't available for Windows.
2613    
2614    
2615    Version 6.6 06-Feb-06
2616    ---------------------
2617    
2618     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2619        in pcreposix.h. I have copied the definition from pcre.h.
2620    
2621     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2622        because pcre.h is no longer a built file.
2623    
2624     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2625        not normally included in the compiled code.
2626    
2627    
2628    Version 6.5 01-Feb-06
2629    ---------------------
2630    
2631     1. When using the partial match feature with pcre_dfa_exec(), it was not
2632        anchoring the second and subsequent partial matches at the new starting
2633        point. This could lead to incorrect results. For example, with the pattern
2634        /1234/, partially matching against "123" and then "a4" gave a match.
2635    
2636     2. Changes to pcregrep:
2637    
2638        (a) All non-match returns from pcre_exec() were being treated as failures
2639            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2640            error message is output. Some extra information is given for the
2641            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2642            probably the only errors that are likely to be caused by users (by
2643            specifying a regex that has nested indefinite repeats, for instance).
2644            If there are more than 20 of these errors, pcregrep is abandoned.
2645    
2646        (b) A binary zero was treated as data while matching, but terminated the
2647            output line if it was written out. This has been fixed: binary zeroes
2648            are now no different to any other data bytes.
2649    
2650        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2651            used to set a locale for matching. The --locale=xxxx long option has
2652            been added (no short equivalent) to specify a locale explicitly on the
2653            pcregrep command, overriding the environment variables.
2654    
2655        (d) When -B was used with -n, some line numbers in the output were one less
2656            than they should have been.
2657    
2658        (e) Added the -o (--only-matching) option.
2659    
2660        (f) If -A or -C was used with -c (count only), some lines of context were
2661            accidentally printed for the final match.
2662    
2663        (g) Added the -H (--with-filename) option.
2664    
2665        (h) The combination of options -rh failed to suppress file names for files
2666            that were found from directory arguments.
2667    
2668        (i) Added the -D (--devices) and -d (--directories) options.
2669    
2670        (j) Added the -F (--fixed-strings) option.
2671    
2672        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2673    
2674        (l) Added the --colo(u)r option.
2675    
2676        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2677            is not present by default.
2678    
2679     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2680        items such as (?R) or (?1), when the recursion could match a number of
2681        alternatives. If it matched one of the alternatives, but subsequently,
2682        outside the recursion, there was a failure, the code tried to back up into
2683        the recursion. However, because of the way PCRE is implemented, this is not
2684        possible, and the result was an incorrect result from the match.
2685    
2686        In order to prevent this happening, the specification of recursion has
2687        been changed so that all such subpatterns are automatically treated as
2688        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2689    
2690     4. I had overlooked the fact that, in some locales, there are characters for
2691        which isalpha() is true but neither isupper() nor islower() are true. In
2692        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2693        and ordfeminine) are like this. This affected the treatment of \w and \W
2694        when they appeared in character classes, but not when they appeared outside
2695        a character class. The bit map for "word" characters is now created
2696        separately from the results of isalnum() instead of just taking it from the
2697        upper, lower, and digit maps. (Plus the underscore character, of course.)
2698    
2699     5. The above bug also affected the handling of POSIX character classes such as
2700        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2701        permanent tables. Instead, the bit maps for such a class were previously
2702        created as the appropriate unions of the upper, lower, and digit bitmaps.
2703        Now they are created by subtraction from the [[:word:]] class, which has
2704        its own bitmap.
2705    
2706     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2707        It is created by subtracting the vertical space characters (\x09, \x0a,
2708        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2709        subtraction was done in the overall bitmap for a character class, meaning
2710        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2711        be recognized. This bug has been fixed.
2712    
2713     7. Patches from the folks at Google:
2714    
2715          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2716          real life, but is still worth protecting against".
2717    
2718          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2719          regular expressions".
2720    
2721          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2722          have it.
2723    
2724          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2725          "configure" and the latter not, in order to fix a problem somebody had
2726          with compiling the Arg class on HP-UX.
2727    
2728          (e) Improve the error-handling of the C++ wrapper a little bit.
2729    
2730          (f) New tests for checking recursion limiting.
2731    
2732     8. The pcre_memmove() function, which is used only if the environment does not
2733        have a standard memmove() function (and is therefore rarely compiled),
2734        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2735        returning a result (though PCRE never actually uses the result).
2736    
2737     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2738        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2739        returned instead of calling malloc() with an overflowing number that would
2740        most likely cause subsequent chaos.
2741    
2742    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2743    
2744    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2745        with this option is matched, the nmatch and pmatch options of regexec() are
2746        ignored.
2747    
2748    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2749        provided in case anyone wants to the the POSIX interface with UTF-8
2750        strings.
2751    
2752    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2753        C++ linking (needed for some HP-UX environments).
2754    
2755    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2756        (unused parameter) and in the pcre_printint() function (omitted "default"
2757        switch label when the default is to do nothing).
2758    
2759    15. Added some code to make it possible, when PCRE is compiled as a C++
2760        library, to replace subject pointers for pcre_exec() with a smart pointer
2761        class, thus making it possible to process discontinuous strings.
2762    
2763    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2764        much the same function. They were added by different people who were trying
2765        to make PCRE easy to compile on non-Unix systems. It has been suggested
2766        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2767        for compiling on Windows systems. I have therefore replaced it with
2768        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2769        defaults to "extern" for C or "extern C" for C++, which works fine on
2770        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2771        SCOPE with something explicit in config.h. In addition:
2772    
2773        (a) pcreposix.h still had just "extern" instead of either of these macros;
2774            I have replaced it with PCRE_DATA_SCOPE.
2775    
2776        (b) Functions such as _pcre_xclass(), which are internal to the library,
2777            but external in the C sense, all had PCRE_EXPORT in their definitions.
2778            This is apparently wrong for the Windows case, so I have removed it.
2779            (It makes no difference on Unix-like systems.)
2780    
2781    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2782        of recursive calls to match(). This is different to MATCH_LIMIT because
2783        that limits the total number of calls to match(), not all of which increase
2784        the depth of recursion. Limiting the recursion depth limits the amount of
2785        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2786        when PCRE is compiled, and changed at run time. A patch from Google adds
2787        this functionality to the C++ interface.
2788    
2789    18. Changes to the handling of Unicode character properties:
2790    
2791        (a) Updated the table to Unicode 4.1.0.
2792    
2793        (b) Recognize characters that are not in the table as "Cn" (undefined).
2794    
2795        (c) I revised the way the table is implemented to a much improved format
2796            which includes recognition of ranges. It now supports the ranges that
2797            are defined in UnicodeData.txt, and it also amalgamates other
2798            characters into ranges. This has reduced the number of entries in the
2799            table from around 16,000 to around 3,000, thus reducing its size
2800            considerably. I realized I did not need to use a tree structure after
2801            all - a binary chop search is just as efficient. Having reduced the
2802            number of entries, I extended their size from 6 bytes to 8 bytes to
2803            allow for more data.
2804    
2805        (d) Added support for Unicode script names via properties such as \p{Han}.
2806    
2807    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2808        matching that character.
2809    
2810    20. When matching a repeated Unicode property with a minimum greater than zero,
2811        (for example \pL{2,}), PCRE could look past the end of the subject if it
2812        reached it while seeking the minimum number of characters. This could
2813        happen only if some of the characters were more than one byte long, because
2814        there is a check for at least the minimum number of bytes.
2815    
2816    21. Refactored the implementation of \p and \P so as to be more general, to
2817        allow for more different types of property in future. This has changed the
2818        compiled form incompatibly. Anybody with saved compiled patterns that use
2819        \p or \P will have to recompile them.
2820    
2821    22. Added "Any" and "L&" to the supported property types.
2822    
2823    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2824        but give a compile time error if the value is greater than 0xff.
2825    
2826    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2827        accidentally not being installed or uninstalled.
2828    
2829    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2830        made were to insert the current release number. This seemed silly, because
2831        it made things harder for people building PCRE on systems that don't run
2832        "configure". I have turned pcre.h into a distributed file, no longer built
2833        by "configure", with the version identification directly included. There is
2834        no longer a pcre.h.in file.
2835    
2836        However, this change necessitated a change to the pcre-config script as
2837        well. It is built from pcre-config.in, and one of the substitutions was the
2838        release number. I have updated configure.ac so that ./configure now finds
2839        the release number by grepping pcre.h.
2840    
2841    26. Added the ability to run the tests under valgrind.
2842    
2843    
2844    Version 6.4 05-Sep-05
2845    ---------------------
2846    
2847     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2848        "--" to be printed when multiple files were scanned, even when none of the
2849        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2850        consider it to be a bug, and have restored the previous behaviour.
2851    
2852     2. A couple of code tidies to get rid of compiler warnings.
2853    
2854     3. The pcretest program used to cheat by referring to symbols in the library
2855        whose names begin with _pcre_. These are internal symbols that are not
2856        really supposed to be visible externally, and in some environments it is
2857        possible to suppress them. The cheating is now confined to including
2858        certain files from the library's source, which is a bit cleaner.
2859    
2860     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2861        file's purpose clearer.
2862    
2863     5. Reorganized pcre_ucp_findchar().
2864    
2865    
2866    Version 6.3 15-Aug-05
2867    ---------------------
2868    
2869     1. The file libpcre.pc.in did not have general read permission in the tarball.
2870    
2871     2. There were some problems when building without C++ support:
2872    
2873        (a) If C++ support was not built, "make install" and "make test" still
2874            tried to test it.
2875    
2876        (b) There were problems when the value of CXX was explicitly set. Some
2877            changes have been made to try to fix these, and ...
2878    
2879        (c) --disable-cpp can now be used to explicitly disable C++ support.
2880    
2881        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2882            backslash in a target when C++ was disabled. This confuses some
2883            versions of "make", apparently. Using an intermediate variable solves
2884            this. (Same for CPP_LOBJ.)
2885    
2886     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2887        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2888        necessary on certain architectures.
2889    
2890     4. Added a setting of -export-symbols-regex to the link command to remove
2891        those symbols that are exported in the C sense, but actually are local
2892        within the library, and not documented. Their names all begin with
2893        "_pcre_". This is not a perfect job, because (a) we have to except some
2894        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2895        available (and never for static libraries). I have made a note to try to
2896        find a way round (a) in the future.
2897    
2898    
2899    Version 6.2 01-Aug-05
2900    ---------------------
2901    
2902     1. There was no test for integer overflow of quantifier values. A construction
2903        such as {1111111111111111} would give undefined results. What is worse, if
2904        a minimum quantifier for a parenthesized subpattern overflowed and became
2905        negative, the calculation of the memory size went wrong. This could have
2906        led to memory overwriting.
2907    
2908     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2909    
2910     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2911        operating environments where this matters.
2912    
2913     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2914        PCRE options from within the C++ wrapper.
2915    
2916     5. Named capturing subpatterns were not being correctly counted when a pattern
2917        was compiled. This caused two problems: (a) If there were more than 100
2918        such subpatterns, the calculation of the memory needed for the whole
2919        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2920        back references of the form \12, where the number was greater than 9, were
2921        not recognized as back references, even though there were sufficient
2922        previous subpatterns.
2923    
2924     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2925        versions of gcc, e.g. 2.95.4.
2926    
2927    
2928    Version 6.1 21-Jun-05
2929    ---------------------
2930    
2931     1. There was one reference to the variable "posix" in pcretest.c that was not
2932        surrounded by "#if !defined NOPOSIX".
2933    
2934     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2935        the cross-check on the old pcre_info() function, for the benefit of the
2936        cut-down version of PCRE that is currently imported into Exim.
2937    
2938     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2939        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2940        patterns that start (?i) but I don't think that matters. The use of (?i) is
2941        just an example; this all applies to the other options as well.
2942    
2943     4. Since libtool seems to echo the compile commands it is issuing, the output
2944        from "make" can be reduced a bit by putting "@" in front of each libtool
2945        compile command.
2946    
2947     5. Patch from the folks at Google for configure.in to be a bit more thorough
2948        in checking for a suitable C++ installation before trying to compile the
2949        C++ stuff. This should fix a reported problem when a compiler was present,
2950        but no suitable headers.
2951    
2952     6. The man pages all had just "PCRE" as their title. I have changed them to
2953        be the relevant file name. I have also arranged that these names are
2954        retained in the file doc/pcre.txt, which is a concatenation in text format
2955        of all the man pages except the little individual ones for each function.
2956    
2957     7. The NON-UNIX-USE file had not been updated for the different set of source
2958        files that come with release 6. I also added a few comments about the C++
2959        wrapper.
2960    
2961    
2962    Version 6.0 07-Jun-05
2963    ---------------------
2964    
2965     1. Some minor internal re-organization to help with my DFA experiments.
2966    
2967     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2968        didn't matter for the library itself when fully configured, but did matter
2969        when compiling without UCP support, or within Exim, where the ucp files are
2970        not imported.
2971    
2972     3. Refactoring of the library code to split up the various functions into
2973        different source modules. The addition of the new DFA matching code (see
2974        below) to a single monolithic source would have made it really too
2975        unwieldy, quite apart from causing all the code to be include in a
2976        statically linked application, when only some functions are used. This is
2977        relevant even without the DFA addition now that patterns can be compiled in
2978        one application and matched in another.
2979    
2980        The downside of splitting up is that there have to be some external
2981        functions and data tables that are used internally in different modules of
2982        the library but which are not part of the API. These have all had their
2983        names changed to start with "_pcre_" so that they are unlikely to clash
2984        with other external names.
2985    
2986     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2987        a different (DFA) algorithm. Although it is slower than the original
2988        function, it does have some advantages for certain types of matching
2989        problem.
2990    
2991     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2992        including restarting after a partial match.
2993    
2994     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2995        defined when compiling for Windows was sent to me. I have put it into the
2996        code, though I have no means of testing or verifying it.
2997    
2998     7. Added the pcre_refcount() auxiliary function.
2999    
3000     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3001        match before or at the first newline in the subject string. In pcretest,
3002        the /f option on a pattern can be used to set this.
3003    
3004     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3005        would behave wrongly. This has been present in PCRE since release 4.0.
3006    
3007    10. A number of changes to the pcregrep command:
3008    
3009        (a) Refactored how -x works; insert ^(...)$ instead of setting
3010            PCRE_ANCHORED and checking the length, in preparation for adding
3011            something similar for -w.
3012    
3013        (b) Added the -w (match as a word) option.
3014    
3015        (c) Refactored the way lines are read and buffered so as to have more
3016            than one at a time available.
3017    
3018        (d) Implemented a pcregrep test script.
3019    
3020        (e) Added the -M (multiline match) option. This allows patterns to match
3021            over several lines of the subject. The buffering ensures that at least
3022            8K, or the rest of the document (whichever is the shorter) is available
3023            for matching (and similarly the previous 8K for lookbehind assertions).
3024    
3025        (f) Changed the --help output so that it now says
3026    
3027              -w, --word-regex(p)
3028    
3029            instead of two lines, one with "regex" and the other with "regexp"
3030            because that confused at least one person since the short forms are the
3031            same. (This required a bit of code, as the output is generated
3032            automatically from a table. It wasn't just a text change.)
3033    
3034        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3035            option but starts with a hyphen. Could be a pattern or a path name
3036            starting with a hyphen, for instance.
3037    
3038        (h) "-" can be given as a file name to represent stdin.
3039    
3040        (i) When file names are being printed, "(standard input)" is used for
3041            the standard input, for compatibility with GNU grep. Previously
3042            "<stdin>" was used.
3043    
3044        (j) The option --label=xxx can be used to supply a name to be used for
3045            stdin when file names are being printed. There is no short form.
3046    
3047        (k) Re-factored the options decoding logic because we are going to add
3048            two more options that take data. Such options can now be given in four
3049            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3050    
3051        (l) Added the -A, -B, and -C options for requesting that lines of context
3052            around matches be printed.
3053    
3054        (m) Added the -L option to print the names of files that do not contain
3055            any matching lines, that is, the complement of -l.
3056    
3057        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3058            continue to scan other files.
3059    
3060        (o) The -s option was incorrectly implemented. For compatibility with other
3061            greps, it now suppresses the error message for a non-existent or non-
3062            accessible file (but not the return code). There is a new option called
3063            -q that suppresses the output of matching lines, which was what -s was
3064            previously doing.
3065    
3066        (p) Added --include and --exclude options to specify files for inclusion
3067            and exclusion when recursing.
3068    
3069    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3070        Hopefully, it now does.
3071    
3072    12. Missing cast in pcre_study().
3073    
3074    13. Added an "uninstall" target to the makefile.
3075    
3076    14. Replaced "extern" in the function prototypes in Makefile.in with
3077        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3078        world, but is set differently for Windows.
3079    
3080    15. Added a second compiling function called pcre_compile2(). The only
3081        difference is that it has an extra argument, which is a pointer to an
3082        integer error code. When there is a compile-time failure, this is set
3083        non-zero, in addition to the error test pointer being set to point to an
3084        error message. The new argument may be NULL if no error number is required
3085        (but then you may as well call pcre_compile(), which is now just a
3086        wrapper). This facility is provided because some applications need a
3087        numeric error indication, but it has also enabled me to tidy up the way
3088        compile-time errors are handled in the POSIX wrapper.
3089    
3090    16. Added VPATH=.libs to the makefile; this should help when building with one
3091        prefix path and installing with another. (Or so I'm told by someone who
3092        knows more about this stuff than I do.)
3093    
3094    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3095        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3096        match everything, including newlines. This is not POSIX-compatible, but
3097        somebody wanted the feature. From pcretest it can be activated by using
3098        both the P and the s flags.
3099    
3100    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3101    
3102    19. libpcre.pc was being incorrectly installed as executable.
3103    
3104    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3105        it now also looks for '\r' so that it will work unmodified on Windows.
3106    
3107    21. Added Google's contributed C++ wrapper to the distribution.
3108    
3109    22. Added some untidy missing memory free() calls in pcretest, to keep
3110        Electric Fence happy when testing.
3111    
3112    
3113    
3114    Version 5.0 13-Sep-04
3115    ---------------------
3116    
3117     1. Internal change: literal characters are no longer packed up into items
3118        containing multiple characters in a single byte-string. Each character
3119        is now matched using a separate opcode. However, there may be more than one
3120        byte in the character in UTF-8 mode.
3121    
3122     2. The pcre_callout_block structure has two new fields: pattern_position and
3123        next_item_length. These contain the offset in the pattern to the next match
3124        item, and its length, respectively.
3125    
3126     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3127        insertion of callouts before each pattern item. Added the /C option to
3128        pcretest to make use of this.
3129    
3130     4. On the advice of a Windows user, the lines
3131    
3132          #if defined(_WIN32) || defined(WIN32)
3133          _setmode( _fileno( stdout ), 0x8000 );
3134          #endif  /* defined(_WIN32) || defined(WIN32) */
3135    
3136        have been added to the source of pcretest. This apparently does useful
3137        magic in relation to line terminators.
3138    
3139     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3140        for the benefit of those environments where the "b" makes a difference.
3141    
3142     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3143        to know about it. I have put a hack into configure.in that adds in code
3144        to set GCC=yes if CC=icc. This seems to end up at a point in the
3145        generated configure script that is early enough to affect the setting of
3146        compiler options, which is what is needed, but I have no means of testing
3147        whether it really works. (The user who reported this had patched the
3148        generated configure script, which of course I cannot do.)
3149    
3150        LATER: After change 22 below (new libtool files), the configure script
3151        seems to know about icc (and also ecc). Therefore, I have commented out
3152        this hack in configure.in.
3153    
3154     7. Added support for pkg-config (2 patches were sent in).
3155    
3156     8. Negated POSIX character classes that used a combination of internal tables
3157        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3158        [[:^ascii]]. Typically, they would match almost any characters. The other
3159        POSIX classes were not broken in this way.
3160    
3161     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3162        to find the match, as PCRE was deluded into thinking that the match had to
3163        start at the start point or following a newline. The same bug applied to
3164        patterns with negative forward assertions or any backward assertions
3165        preceding ".*" at the start, unless the pattern required a fixed first
3166        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3167    
3168    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3169        starting at the last subject character, bytes beyond the end of the subject
3170        string were read.
3171    
3172    11. Renamed the variable "class" as "classbits" to make life easier for C++
3173        users. (Previously there was a macro definition, but it apparently wasn't
3174        enough.)
3175    
3176    12. Added the new field "tables" to the extra data so that tables can be passed
3177        in at exec time, or the internal tables can be re-selected. This allows
3178        a compiled regex to be saved and re-used at a later time by a different
3179        program that might have everything at different addresses.
3180    
3181    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3182        -R library as well as a -L library.
3183    
3184    14. The debugging options of pcretest (-d on the command line or D on a
3185        pattern) showed incorrect output for anything following an extended class
3186        that contained multibyte characters and which was followed by a quantifier.
3187    
3188    15. Added optional support for general category Unicode character properties
3189        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3190        support. It adds about 90K to the size of the library. The meanings of the
3191        inbuilt class escapes such as \d and \s have NOT been changed.
3192    
3193    16. Updated pcredemo.c to include calls to free() to release the memory for the
3194        compiled pattern.
3195    
3196    17. The generated file chartables.c was being created in the source directory
3197        instead of in the building directory. This caused the build to fail if the
3198        source directory was different from the building directory, and was
3199        read-only.
3200    
3201    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3202        file. No doubt somebody will tell me if they don't make sense... Also added
3203        Dan Mooney's comments about building on OpenVMS.
3204    
3205    19. Added support for partial matching via the PCRE_PARTIAL option for
3206        pcre_exec() and the \P data escape in pcretest.
3207    
3208    20. Extended pcretest with 3 new pattern features:
3209    
3210        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3211              write the compiled pattern to the file whose name is "rest-of-line".
3212              This is a straight binary dump of the data, with the saved pointer to
3213              the character tables forced to be NULL. The study data, if any, is
3214              written too. After writing, pcretest reads a new pattern.
3215    
3216        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3217              compiled pattern from the given file. There must not be any
3218              occurrences of "<" in the file name (pretty unlikely); if there are,
3219              pcretest will instead treat the initial "<" as a pattern delimiter.
3220              After reading in the pattern, pcretest goes on to read data lines as
3221              usual.
3222    
3223        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3224              and 16-bit fields in a compiled pattern, to simulate a pattern that
3225              was compiled on a host of opposite endianness.
3226    
3227    21. The pcre-exec() function can now cope with patterns that were compiled on
3228        hosts of opposite endianness, with this restriction:
3229    
3230          As for any compiled expression that is saved and used later, the tables
3231          pointer field cannot be preserved; the extra_data field in the arguments
3232          to pcre_exec() should be used to pass in a tables address if a value
3233          other than the default internal tables were used at compile time.
3234    
3235    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3236        now diagnosed as an error. Previously, most of the time, a negative number
3237        would have been treated as zero, but if in addition "ovector" was passed as
3238        NULL, a crash could occur.
3239    
3240    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3241        new versions from the libtool 1.5 distribution (the last one is a copy of
3242        a file called libtool.m4). This seems to have fixed the need to patch
3243        "configure" to support Darwin 1.3 (which I used to do). However, I still
3244        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3245        workstation).
3246    
3247    24. Changed the PCRE licence to be the more standard "BSD" licence.
3248    
3249    
3250    Version 4.5 01-Dec-03
3251    ---------------------
3252    
3253     1. There has been some re-arrangement of the code for the match() function so
3254        that it can be compiled in a version that does not call itself recursively.
3255        Instead, it keeps those local variables that need separate instances for
3256        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3257        needs to "recurse". Keeping track of where control must go is done by means
3258        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3259        hide most of the details from the main code, and operates only if
3260        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3261        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3262        operating.
3263    
3264        To make it easier for callers to provide specially tailored get/free
3265        functions for this usage, two new functions, pcre_stack_malloc, and
3266        pcre_stack_free, are used. They are always called in strict stacking order,
3267        and the size of block requested is always the same.
3268    
3269        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3270        PCRE has been compiled to use the stack or the heap for recursion. The
3271        -C option of pcretest uses this to show which version is compiled.
3272    
3273        A new data escape \S, is added to pcretest; it causes the amounts of store
3274        obtained and freed by both kinds of malloc/free at match time to be added
3275        to the output.
3276    
3277     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3278        what's available on my current Linux desktop machine.
3279    
3280     3. When matching a UTF-8 string, the test for a valid string at the start has
3281        been extended. If start_offset is not zero, PCRE now checks that it points
3282        to a byte that is the start of a UTF-8 character. If not, it returns
3283        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3284        this is necessary because there may be backward assertions in the pattern.
3285        When matching the same subject several times, it may save resources to use
3286        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3287    
3288     4. The code for checking the validity of UTF-8 strings has been tightened so
3289        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3290        containing "overlong sequences".
3291    
3292     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3293        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3294        should have been "&", but it just so happened that all the cases this let
3295        through by mistake were picked up later in the function.
3296    
3297     6. I had used a variable called "isblank" - this is a C99 function, causing
3298        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3299    
3300     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3301        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3302        so the version is shown; (c) stop "make test" echoing "./RunTest".
3303    
3304     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3305    
3306     9. The prototype for memmove() for systems that don't have it was using
3307        size_t, but the inclusion of the header that defines size_t was later. I've
3308        moved the #includes for the C headers earlier to avoid this.
3309    
3310    10. Added some adjustments to the code to make it easier to compiler on certain
3311        special systems:
3312    
3313          (a) Some "const" qualifiers were missing.
3314          (b) Added the macro EXPORT before all exported functions; by default this
3315              is defined to be empty.
3316          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3317              that it reads its output file name as an argument instead of writing
3318              to the standard output and assuming this can be redirected.
3319    
3320    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3321        class containing characters with values greater than 255, PCRE compilation
3322        went into a loop.
3323    
3324    12. A recursive reference to a subpattern that was within another subpattern
3325        that had a minimum quantifier of zero caused PCRE to crash. For example,
3326        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3327        recursion. If the recursively-called subpattern itself had a zero repeat,
3328        that was OK.
3329    
3330    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3331        buffer into which it was copied (for escape processing) was still set at
3332        1024, so long lines caused crashes.
3333    
3334    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3335        "internal error: code overflow...". This applied to any character class
3336        that was followed by a possessive quantifier.
3337    
3338    15. Modified the Makefile to add libpcre.la as a prerequisite for
3339        libpcreposix.la because I was told this is needed for a parallel build to
3340        work.
3341    
3342    16. If a pattern that contained .* following optional items at the start was
3343        studied, the wrong optimizing data was generated, leading to matching
3344        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3345        matching string must start with a or b or c. The correct conclusion for
3346        this pattern is that a match can start with any character.
3347    
3348    
3349    Version 4.4 13-Aug-03
3350    ---------------------
3351    
3352     1. In UTF-8 mode, a character class containing characters with values between
3353        127 and 255 was not handled correctly if the compiled pattern was studied.
3354        In fixing this, I have also improved the studying algorithm for such
3355        classes (slightly).
3356    
3357     2. Three internal functions had redundant arguments passed to them. Removal
3358        might give a very teeny performance improvement.
3359    
3360     3. Documentation bug: the value of the capture_top field in a callout is *one
3361        more than* the number of the hightest numbered captured substring.
3362    
3363     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3364        in incorrectly linking with a previously installed version. They now link
3365        explicitly with libpcre.la.
3366    
3367     5. configure.in no longer needs to recognize Cygwin specially.
3368    
3369     6. A problem in pcre.in for Windows platforms is fixed.
3370    
3371     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3372        pcretest, it used to include the size of the study block as part of its
3373        output. Unfortunately, the structure contains a field that has a different
3374        size on different hardware architectures. This meant that the tests that
3375        showed this size failed. As the block is currently always of a fixed size,
3376        this information isn't actually particularly useful in pcretest output, so
3377        I have just removed it.
3378    
3379     8. Three pre-processor statements accidentally did not start in column 1.
3380        Sadly, there are *still* compilers around that complain, even though
3381        standard C has not required this for well over a decade. Sigh.
3382    
3383     9. In pcretest, the code for checking callouts passed small integers in the
3384        callout_data field, which is a void * field. However, some picky compilers
3385        complained about the casts involved for this on 64-bit systems. Now
3386        pcretest passes the address of the small integer instead, which should get
3387        rid of the warnings.
3388    
3389    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3390        both compile and run time, and gives an error if an invalid UTF-8 sequence
3391        is found. There is a option for disabling this check in cases where the
3392        string is known to be correct and/or the maximum performance is wanted.
3393    
3394    11. In response to a bug report, I changed one line in Makefile.in from
3395    
3396            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3397        to
3398            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3399    
3400        to look similar to other lines, but I have no way of telling whether this
3401        is the right thing to do, as I do not use Windows. No doubt I'll get told
3402        if it's wrong...
3403    
3404    
3405    Version 4.3 21-May-03
3406    ---------------------
3407    
3408    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3409       Makefile.
3410    
3411    2. Some refactoring to improve the quality of the code:
3412    
3413       (i)   The utf8_table... variables are now declared "const".
3414    
3415       (ii)  The code for \cx, which used the "case flipping" table to upper case
3416             lower case letters, now just substracts 32. This is ASCII-specific,
3417             but the whole concept of \cx is ASCII-specific, so it seems
3418             reasonable.
3419    
3420       (iii) PCRE was using its character types table to recognize decimal and
3421             hexadecimal digits in the pattern. This is silly, because it handles
3422             only 0-9, a-f, and A-F, but the character types table is locale-
3423             specific, which means strange things might happen. A private
3424             table is now used for this - though it costs 256 bytes, a table is
3425             much faster than multiple explicit tests. Of course, the standard
3426             character types table is still used for matching digits in subject
3427             strings against \d.
3428    
3429       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3430             ending in _t are). So I've renamed it as ESC_tee.
3431    
3432    3. The first argument for regexec() in the POSIX wrapper should have been
3433       defined as "const".
3434    
3435    4. Changed pcretest to use malloc() for its buffers so that they can be
3436       Electric Fenced for debugging.
3437    
3438    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3439       to read one or more bytes before the start of the subject string. Often this
3440       had no effect on PCRE's behaviour, but in some circumstances it could
3441       provoke a segmentation fault.
3442    
3443    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3444       to try to read one or more bytes before the start of the subject string.
3445    
3446    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3447       UTF-8 support could misbehave in various ways if the subject string
3448       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3449       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3450       back over UTF-8 characters.)
3451    
3452    
3453    Version 4.2 14-Apr-03
3454    ---------------------
3455    
3456    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3457    
3458    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3459         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3460         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3461         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3462         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3463           and BUILD_EXEEXT
3464         Note: automatic setting of the BUILD variables is not yet working
3465         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3466           compile-time but not at link-time
3467         [LINK]: use for linking executables only
3468         make different versions for Windows and non-Windows
3469         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3470           libraries
3471         [LINK_FOR_BUILD]: new variable
3472         [OBJEXT]: use throughout
3473         [EXEEXT]: use throughout
3474         <winshared>: new target
3475         <wininstall>: new target
3476         <dftables.o>: use native compiler
3477         <dftables>: use native linker
3478         <install>: handle Windows platform correctly
3479         <clean>: ditto
3480         <check>: ditto
3481         copy DLL to top builddir before testing
3482    
3483       As part of these changes, -no-undefined was removed again. This was reported
3484       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3485       in any case.
3486    
3487    3. Some tidies to get rid of compiler warnings:
3488    
3489       . In the match_data structure, match_limit was an unsigned long int, whereas
3490         match_call_count was an int. I've made them both unsigned long ints.
3491    
3492       . In pcretest the fact that a const uschar * doesn't automatically cast to
3493         a void * provoked a warning.
3494    
3495       . Turning on some more compiler warnings threw up some "shadow" variables
3496         and a few more missing casts.
3497    
3498    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3499       option, a class that contained a single character with a value between 128
3500       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3501    
3502    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3503       option, a class that contained several characters, but with at least one
3504       whose value was between 128 and 255 caused PCRE to crash.
3505    
3506    
3507    Version 4.1 12-Mar-03
3508    ---------------------
3509    
3510    1. Compiling with gcc -pedantic found a couple of places where casts were
3511    needed, and a string in dftables.c that was longer than standard compilers are
3512    required to support.
3513    
3514    2. Compiling with Sun's compiler found a few more places where the code could
3515    be tidied up in order to avoid warnings.
3516    
3517    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3518    first of these names is deprecated in the latest Autoconf in favour of the name
3519    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3520    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3521    analogy I have changed it to CFLAGS_FOR_BUILD.
3522    
3523    4. Added -no-undefined to the linking command in the Makefile, because this is
3524    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3525    linking step for the pcreposix library.
3526    
3527    5. PCRE was failing to diagnose the case of two named groups with the same
3528    name.
3529    
3530    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3531    literal character that is needed in the subject for a match, and scans along to
3532    ensure that it is present before embarking on the full matching process. This
3533    saves time in cases of nested unlimited repeats that are never going to match.
3534    Problem: the scan can take a lot of time if the subject is very long (e.g.
3535    megabytes), thus penalizing straightforward matches. It is now done only if the
3536    amount of subject to be scanned is less than 1000 bytes.
3537    
3538    7. A lesser problem with the same optimization is that it was recording the
3539    first character of an anchored pattern as "needed", thus provoking a search
3540    right along the subject, even when the first match of the pattern was going to
3541    fail. The "needed" character is now not set for anchored patterns, unless it
3542    follows something in the pattern that is of non-fixed length. Thus, it still
3543    fulfils its original purpose of finding quick non-matches in cases of nested
3544    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3545    
3546    
3547    Version 4.0 17-Feb-03
3548    ---------------------
3549    
3550    1. If a comment in an extended regex that started immediately after a meta-item
3551    extended to the end of string, PCRE compiled incorrect data. This could lead to
3552    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3553    
3554    2. Moved to autoconf 2.53 and libtool 1.4.2.
3555    
3556    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3557    the special perltest8 script is no longer needed - all the tests can be run
3558    from a single perltest script.
3559    
3560    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3561    by \s. It has now been removed in PCRE. This means it isn't recognized as
3562    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3563    class [:space:] *does* include VT, thereby creating a mess.
3564    
3565    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3566    space and tab.
3567    
3568    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3569    its new features into the main test script, reducing the number of scripts.
3570    
3571    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3572    were backward compatible, and made the (?i) apply to the whole pattern, as if
3573    /i were given. Now it behaves more logically, and applies the option setting
3574    only to what follows. PCRE has been changed to follow suit. However, if it
3575    finds options settings right at the start of the pattern, it extracts them into
3576    the global options, as before. Thus, they show up in the info data.
3577    
3578    8. Added support for the \Q...\E escape sequence. Characters in between are
3579    treated as literals. This is slightly different from Perl in that $ and @ are
3580    also handled as literals inside the quotes. In Perl, they will cause variable
3581    interpolation. Note the following examples:
3582    
3583        Pattern            PCRE matches      Perl matches
3584    
3585        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3586        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3587        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3588    
3589    For compatibility with Perl, \Q...\E sequences are recognized inside character
3590    classes as well as outside them.
3591    
3592    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3593    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3594    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3595    signed/unsigned warnings.
3596    
3597    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3598    option for pcretest, so I've replaced it by a simple function that does just
3599    that job.
3600    
3601    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3602    "pcregrep -".
3603    
3604    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3605    Java package. This provides some syntactic sugar for simple cases of what my
3606    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3607    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3608    item, you can use this simplified notation. Note that only makes sense with
3609    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3610    greediness, whatever the setting of the PCRE_UNGREEDY option.
3611    
3612    13. A change of greediness default within a pattern was not taking effect at
3613    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3614    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3615    was abstracted outside.
3616    
3617    14. PCRE now supports the \G assertion. It is true when the current matching
3618    position is at the start point of the match. This differs from \A when the
3619    starting offset is non-zero. Used with the /g option of pcretest (or similar
3620    code), it works in the same way as it does for Perl's /g option. If all
3621    alternatives of a regex begin with \G, the expression is anchored to the start
3622    match position, and the "anchored" flag is set in the compiled expression.
3623    
3624    15. Some bugs concerning the handling of certain option changes within patterns
3625    have been fixed. These applied to options other than (?ims). For example,
3626    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3627    been the other way round. Some of this was related to change 7 above.
3628    
3629    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3630    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3631    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3632    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3633    
3634    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3635    mode. Unlike ".", it always matches newline, whatever the setting of
3636    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3637    assertions. Perl allows it, but it doesn't (in general) work because it can't
3638    calculate the length of the lookbehind. At least, that's the case for Perl
3639    5.8.0 - I've been told they are going to document that it doesn't work in
3640    future.
3641    
3642    18. Added an error diagnosis for escapes that PCRE does not support: these are
3643    \L, \l, \N, \P, \p, \U, \u, and \X.
3644    
3645    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3646    reading past the end of the pattern in cases such as /[abcd/.
3647    
3648    20. PCRE was getting more memory than necessary for patterns with classes that
3649    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3650    
3651    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3652    compiling PCRE for use with Virtual Pascal.
3653    
3654    22. Small fix to the Makefile to make it work properly if the build is done
3655    outside the source tree.
3656    
3657    23. Added a new extension: a condition to go with recursion. If a conditional
3658    subpattern starts with (?(R) the "true" branch is used if recursion has
3659    happened, whereas the "false" branch is used only at the top level.
3660    
3661    24. When there was a very long string of literal characters (over 255 bytes
3662    without UTF support, over 250 bytes with UTF support), the computation of how
3663    much memory was required could be incorrect, leading to segfaults or other
3664    strange effects.
3665    
3666    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3667    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3668    there was a subsequent back reference to those brackets. This meant that, for
3669    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3670    possible to check for precisely this case. All we can do is abandon the
3671    optimization if .* occurs inside capturing brackets when there are any back
3672    references whatsoever. (See below for a better fix that came later.)
3673    
3674    26. The handling of the optimization for finding the first character of a
3675    non-anchored pattern, and for finding a character that is required later in the
3676    match were failing in some cases. This didn't break the matching; it just
3677    failed to optimize when it could. The way this is done has been re-implemented.
3678    
3679    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3680    
3681    28. Added a new feature that provides some of the functionality that Perl
3682    provides with (?{...}). The facility is termed a "callout". The way it is done
3683    in PCRE is for the caller to provide an optional function, by setting
3684    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3685    global variable. By default it is unset, which disables all calling out. To get
3686    the function called, the regex must include (?C) at appropriate points. This
3687    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3688    This provides a means of identifying different callout points. When PCRE
3689    reaches such a point in the regex, if pcre_callout has been set, the external
3690    function is called. It is provided with data in a structure called
3691    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3692    matching continues; if it returns a non-zero value, the match at the current
3693    point fails. However, backtracking will occur if possible. [This was changed
3694    later and other features added - see item 49 below.]
3695    
3696    29. pcretest is upgraded to test the callout functionality. It provides a
3697    callout function that displays information. By default, it shows the start of
3698    the match and the current position in the text. There are some new data escapes
3699    to vary what happens:
3700    
3701        \C+         in addition, show current contents of captured substrings
3702        \C-         do not supply a callout function
3703        \C!n        return 1 when callout number n is reached
3704        \C!n!m      return 1 when callout number n is reached for the mth time
3705    
3706    30. If pcregrep was called with the -l option and just a single file name, it
3707    output "<stdin>" if a match was found, instead of the file name.
3708    
3709    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3710    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3711    pcre_exec(). This saves a malloc/free per call. The default value of
3712    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3713    when configuring.
3714    
3715    32. The default maximum size of a compiled pattern is 64K. There have been a
3716    few cases of people hitting this limit. The code now uses macros to handle the
3717    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3718    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3719    configuring. Tests 2 and 5 work only with 2-byte links because they output
3720    debugging information about compiled patterns.
3721    
3722    33. Internal code re-arrangements:
3723    
3724    (a) Moved the debugging function for printing out a compiled regex into
3725        its own source file (printint.c) and used #include to pull it into
3726        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3727        separate copies.
3728    
3729    (b) Defined the list of op-code names for debugging as a macro in
3730        internal.h so that it is next to the definition of the opcodes.
3731    
3732    (c) Defined a table of op-code lengths for simpler skipping along compiled
3733        code. This is again a macro in internal.h so that it is next to the
3734        definition of the opcodes.
3735    
3736    34. Added support for recursive calls to individual subpatterns, along the
3737    lines of Robin Houston's patch (but implemented somewhat differently).
3738    
3739    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3740    allow it to read and process whole directories in Win32. This code was
3741    contributed by Lionel Fourquaux; it has not been tested by me.
3742    
3743    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3744    used to name a group. Names consist of alphanumerics and underscores, and must
3745    be unique. Back references use the syntax (?P=name) and recursive calls use
3746    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3747    numbers. The function pcre_fullinfo() can be used after compilation to extract
3748    a name/number map. There are three relevant calls:
3749    
3750      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3751      PCRE_INFO_NAMECOUNT            yields the number of entries
3752      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3753    
3754    The map is a vector of fixed-size entries. The size of each entry depends on
3755    the length of the longest name used. The first two bytes of each entry are the
3756    group number, most significant byte first. There follows the corresponding
3757    name, zero terminated. The names are in alphabetical order.
3758    
3759    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3760    case instead of 255. Making it the same both with and without UTF-8 support
3761    means that the same test output works with both.
3762    
3763    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3764    calling malloc() with a zero argument.
3765    
3766    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3767    optimization. I've improved things by keeping a bitmap of backreferences with
3768    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3769    fact referenced, the optimization can be applied. It is unlikely that a
3770    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3771    the match to follow \n) will appear inside brackets with a number greater than
3772    31, but if it does, any back reference > 31 suppresses the optimization.
3773    
3774    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3775    of disabling numbered capturing parentheses. Any opening parenthesis that is
3776    not followed by ? behaves as if it were followed by ?: but named parentheses
3777    can still be used for capturing (and they will acquire numbers in the usual
3778    way).
3779    
3780    41. Redesigned the return codes from the match() function into yes/no/error so
3781    that errors can be passed back from deep inside the nested calls. A malloc
3782    failure while inside a recursive subpattern call now causes the
3783    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3784    
3785    42. It is now possible to set a limit on the number of times the match()
3786    function is called in a call to pcre_exec(). This facility makes it possible to
3787    limit the amount of recursion and backtracking, though not in a directly
3788    obvious way, because the match() function is used in a number of different
3789    circumstances. The count starts from zero for each position in the subject
3790    string (for non-anchored patterns). The default limit is, for compatibility, a
3791    large number, namely 10 000 000. You can change this in two ways:
3792    
3793    (a) When configuring PCRE before making, you can use --with-match-limit=n
3794        to set a default value for the compiled library.
3795    
3796    (b) For each call to pcre_exec(), you can pass a p