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

Diff of /code/trunk/ChangeLog

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

revision 61 by nigel, Sat Feb 24 21:39:58 2007 UTC revision 929 by zherczeg, Fri Feb 24 11:07:47 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4  Version 3.0 02-Jan-02  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    
48    Version 8.30 04-February-2012
49    -----------------------------
50    
51    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
52        name is defined in ctype.h.
53    
54    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
55        only in quite long subpatterns.
56    
57    3.  Removed the function pcre_info(), which has been obsolete and deprecated
58        since it was replaced by pcre_fullinfo() in February 2000.
59    
60    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
61        match a (*MARK), and the match failed at the start of the subject, a
62        reference to memory before the start of the subject could occur. This bug
63        was introduced by fix 17 of release 8.21.
64    
65    5.  A reference to an unset group with zero minimum repetition was giving
66        totally wrong answers (in non-JavaScript-compatibility mode). For example,
67        /(another)?(\1?)test/ matched against "hello world test". This bug was
68        introduced in release 8.13.
69    
70    6.  Add support for 16-bit character strings (a large amount of work involving
71        many changes and refactorings).
72    
73    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
74        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
75        from a file.
76    
77    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
78        rounding is not applied in this particular case).
79    
80    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
81        if they appear, or are escaped, in patterns.
82    
83    10. Get rid of a number of -Wunused-but-set-variable warnings.
84    
85    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
86        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
87        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
88        also returns the mark "x". This bug applied to capturing parentheses,
89        non-capturing parentheses, and atomic parentheses. It also applied to some
90        assertions.
91    
92    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
93        information out of configure.ac instead of relying on pcre.h.generic, which
94        is not stored in the repository.
95    
96    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
97        -lreadline.
98    
99    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
100    
101    15. Applied Graycode's patch to put the top-level frame on the stack rather
102        than the heap when not using the stack for recursion. This gives a
103        performance improvement in many cases when recursion is not deep.
104    
105    16. Experimental code added to "pcretest -C" to output the stack frame size.
106    
107    
108    Version 8.21 12-Dec-2011
109    ------------------------
110    
111    1.  Updating the JIT compiler.
112    
113    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
114        are added as well.
115    
116    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
117        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
118        calling _pcre_jit_exec. Some extra comments are added.
119    
120    4.  (*MARK) settings inside atomic groups that do not contain any capturing
121        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
122        was introduced by change 18 for 8.20.
123    
124    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
125        ECMA-262 standard.
126    
127    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
128        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
129        This bug was probably introduced by change 9 of 8.13.
130    
131    7.  While fixing 6 above, I noticed that a number of other items were being
132        incorrectly rejected as "not fixed length". This arose partly because newer
133        opcodes had not been added to the fixed-length checking code. I have (a)
134        corrected the bug and added tests for these items, and (b) arranged for an
135        error to occur if an unknown opcode is encountered while checking for fixed
136        length instead of just assuming "not fixed length". The items that were
137        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
138        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
139        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
140    
141    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
142        being incorrectly compiled and would have given unpredicatble results.
143    
144    9.  A possessively repeated subpattern with minimum repeat count greater than
145        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
146        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
147        the first (A) could occur when it should not.
148    
149    10. Add a cast and remove a redundant test from the code.
150    
151    11. JIT should use pcre_malloc/pcre_free for allocation.
152    
153    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
154        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
155        is anything other than /usr, -L is still shown).
156    
157    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
158    
159    14. Perl does not support \N without a following name in a [] class; PCRE now
160        also gives an error.
161    
162    15. If a forward reference was repeated with an upper limit of around 2000,
163        it caused the error "internal error: overran compiling workspace". The
164        maximum number of forward references (including repeats) was limited by the
165        internal workspace, and dependent on the LINK_SIZE. The code has been
166        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
167        the default depends on LINK_SIZE. There is a new upper limit (for safety)
168        of around 200,000 forward references. While doing this, I also speeded up
169        the filling in of repeated forward references.
170    
171    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
172        incorrectly expecting the subject to contain another "a" after the start.
173    
174    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
175        in the match, the SKIP should be ignored. This was not happening; instead
176        the SKIP was being treated as NOMATCH. For patterns such as
177        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
178        tested.
179    
180    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
181        now much more compatible with Perl, in particular in cases where the result
182        is a non-match for a non-anchored pattern. For example, if
183        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
184        "m", where previously it did not return a name. A side effect of this
185        change is that for partial matches, the last encountered mark name is
186        returned, as for non matches. A number of tests that were previously not
187        Perl-compatible have been moved into the Perl-compatible test files. The
188        refactoring has had the pleasing side effect of removing one argument from
189        the match() function, thus reducing its stack requirements.
190    
191    19. If the /S+ option was used in pcretest to study a pattern using JIT,
192        subsequent uses of /S (without +) incorrectly behaved like /S+.
193    
194    21. Retrieve executable code size support for the JIT compiler and fixing
195        some warnings.
196    
197    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
198        not work when the shorter character appeared right at the end of the
199        subject string.
200    
201    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
202        systems.
203    
204    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
205        output it when the /M option is used in pcretest.
206    
207    25. The CheckMan script was not being included in the distribution. Also, added
208        an explicit "perl" to run Perl scripts from the PrepareRelease script
209        because this is reportedly needed in Windows.
210    
211    26. If study data was being save in a file and studying had not found a set of
212        "starts with" bytes for the pattern, the data written to the file (though
213        never used) was taken from uninitialized memory and so caused valgrind to
214        complain.
215    
216    27. Updated RunTest.bat as provided by Sheri Pierce.
217    
218    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
219    
220    29. Computation of memory usage for the table of capturing group names was
221        giving an unnecessarily large value.
222    
223    
224    Version 8.20 21-Oct-2011
225    ------------------------
226    
227    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
228        a POSIX class. After further experiments with Perl, which convinced me that
229        Perl has bugs and confusions, a closing square bracket is no longer allowed
230        in a POSIX name. This bug also affected patterns with classes that started
231        with full stops.
232    
233    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
234        captured substring, but while checking the failing first alternative,
235        substring 1 is temporarily captured. If the output vector supplied to
236        pcre_exec() was not big enough for this capture, the yield of the function
237        was still zero ("insufficient space for captured substrings"). This cannot
238        be totally fixed without adding another stack variable, which seems a lot
239        of expense for a edge case. However, I have improved the situation in cases
240        such as /(a)(b)x|abc/ matched against "abc", where the return code
241        indicates that fewer than the maximum number of slots in the ovector have
242        been set.
243    
244    3.  Related to (2) above: when there are more back references in a pattern than
245        slots in the output vector, pcre_exec() uses temporary memory during
246        matching, and copies in the captures as far as possible afterwards. It was
247        using the entire output vector, but this conflicts with the specification
248        that only 2/3 is used for passing back captured substrings. Now it uses
249        only the first 2/3, for compatibility. This is, of course, another edge
250        case.
251    
252    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
253        main code base, and can be used by building with --enable-jit. When this is
254        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
255        runtime --no-jit option is given.
256    
257    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
258        ovector, the return from the function was zero, implying that there were
259        other matches that did not fit. The correct "exactly full" value is now
260        returned.
261    
262    6.  If a subpattern that was called recursively or as a subroutine contained
263        (*PRUNE) or any other control that caused it to give a non-standard return,
264        invalid errors such as "Error -26 (nested recursion at the same subject
265        position)" or even infinite loops could occur.
266    
267    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
268        computing the minimum length on reaching *ACCEPT, and so ended up with the
269        wrong value of 1 rather than 0. Further investigation indicates that
270        computing a minimum subject length in the presence of *ACCEPT is difficult
271        (think back references, subroutine calls), and so I have changed the code
272        so that no minimum is registered for a pattern that contains *ACCEPT.
273    
274    8.  If (*THEN) was present in the first (true) branch of a conditional group,
275        it was not handled as intended. [But see 16 below.]
276    
277    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
278        Sheri Pierce.
279    
280    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
281        the first byte in a match must be "a".
282    
283    11. Change 17 for 8.13 increased the recursion depth for patterns like
284        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
285        pattern contains any instances of (*THEN). If it does not, the old
286        optimizations are restored. It would be nice to do this on a per-group
287        basis, but at the moment that is not feasible.
288    
289    12. In some environments, the output of pcretest -C is CRLF terminated. This
290        broke RunTest's code that checks for the link size. A single white space
291        character after the value is now allowed for.
292    
293    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
294        For "fr", it uses the Windows-specific input and output files.
295    
296    14. If (*THEN) appeared in a group that was called recursively or as a
297        subroutine, it did not work as intended. [But see next item.]
298    
299    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
300        pattern fragments (but not containing any | characters). If A and B are
301        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
302        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
303        D. In other words, Perl considers parentheses that do not contain any |
304        characters to be part of a surrounding alternative, whereas PCRE was
305        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
306        differently. PCRE now behaves in the same way as Perl, except in the case
307        of subroutine/recursion calls such as (?1) which have in any case always
308        been different (but PCRE had them first :-).
309    
310    16. Related to 15 above: Perl does not treat the | in a conditional group as
311        creating alternatives. Such a group is treated in the same way as an
312        ordinary group without any | characters when processing (*THEN). PCRE has
313        been changed to match Perl's behaviour.
314    
315    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
316        RunGrepTest script failed.
317    
318    18. Change 22 for version 13 caused atomic groups to use more stack. This is
319        inevitable for groups that contain captures, but it can lead to a lot of
320        stack use in large patterns. The old behaviour has been restored for atomic
321        groups that do not contain any capturing parentheses.
322    
323    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
324        suppress the check for a minimum subject length at run time. (If it was
325        given to pcre_exec() or pcre_dfa_exec() it did work.)
326    
327    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
328        fail to work when decoding hex characters in data strings in EBCDIC
329        environments.
330    
331    21. It appears that in at least one Mac OS environment, the isxdigit() function
332        is implemented as a macro that evaluates to its argument more than once,
333        contravening the C 90 Standard (I haven't checked a later standard). There
334        was an instance in pcretest which caused it to go wrong when processing
335        \x{...} escapes in subject strings. The has been rewritten to avoid using
336        things like p++ in the argument of isxdigit().
337    
338    
339    Version 8.13 16-Aug-2011
340    ------------------------
341    
342    1.  The Unicode data tables have been updated to Unicode 6.0.0.
343    
344    2.  Two minor typos in pcre_internal.h have been fixed.
345    
346    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
347        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
348        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
349    
350    4.  There were a number of related bugs in the code for matching backrefences
351        caselessly in UTF-8 mode when codes for the characters concerned were
352        different numbers of bytes. For example, U+023A and U+2C65 are an upper
353        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
354        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
355        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
356        2-byte code at the end of the subject (it thought there wasn't enough data
357        left).
358    
359    5.  Comprehensive information about what went wrong is now returned by
360        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
361        as the output vector has at least 2 elements. The offset of the start of
362        the failing character and a reason code are placed in the vector.
363    
364    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
365        now returned is for the first byte of the failing character, instead of the
366        last byte inspected. This is an incompatible change, but I hope it is small
367        enough not to be a problem. It makes the returned offset consistent with
368        pcre_exec() and pcre_dfa_exec().
369    
370    7.  pcretest now gives a text phrase as well as the error number when
371        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
372        failure, the offset and reason code are output.
373    
374    8.  When \R was used with a maximizing quantifier it failed to skip backwards
375        over a \r\n pair if the subsequent match failed. Instead, it just skipped
376        back over a single character (\n). This seems wrong (because it treated the
377        two characters as a single entity when going forwards), conflicts with the
378        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
379        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
380        has been changed.
381    
382    9.  Some internal refactoring has changed the processing so that the handling
383        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
384        time (the PCRE_DOTALL option was changed this way some time ago: version
385        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
386        which was always a bit of a fudge. It also means that there is one less
387        argument for the match() function, which reduces its stack requirements
388        slightly. This change also fixes an incompatibility with Perl: the pattern
389        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
390    
391    10. More internal refactoring has drastically reduced the number of recursive
392        calls to match() for possessively repeated groups such as (abc)++ when
393        using pcre_exec().
394    
395    11. While implementing 10, a number of bugs in the handling of groups were
396        discovered and fixed:
397    
398        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
399        (a|)*(?1) gave a compile-time internal error.
400        ((a|)+)+  did not notice that the outer group could match an empty string.
401        (^a|^)+   was not marked as anchored.
402        (.*a|.*)+ was not marked as matching at start or after a newline.
403    
404    12. Yet more internal refactoring has removed another argument from the match()
405        function. Special calls to this function are now indicated by setting a
406        value in a variable in the "match data" data block.
407    
408    13. Be more explicit in pcre_study() instead of relying on "default" for
409        opcodes that mean there is no starting character; this means that when new
410        ones are added and accidentally left out of pcre_study(), testing should
411        pick them up.
412    
413    14. The -s option of pcretest has been documented for ages as being an old
414        synonym of -m (show memory usage). I have changed it to mean "force study
415        for every regex", that is, assume /S for every regex. This is similar to -i
416        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
417        using it. It makes it easier to run collections of tests with and without
418        study enabled, and thereby test pcre_study() more easily. All the standard
419        tests are now run with and without -s (but some patterns can be marked as
420        "never study" - see 20 below).
421    
422    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
423        restoration of the capturing data to the outer values was not happening
424        correctly.
425    
426    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
427        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
428        pattern had matched an empty string, and so incorrectly returned a no
429        match.
430    
431    17. There was optimizing code for the last branch of non-capturing parentheses,
432        and also for the obeyed branch of a conditional subexpression, which used
433        tail recursion to cut down on stack usage. Unfortunately, now that there is
434        the possibility of (*THEN) occurring in these branches, tail recursion is
435        no longer possible because the return has to be checked for (*THEN). These
436        two optimizations have therefore been removed. [But see 8.20/11 above.]
437    
438    18. If a pattern containing \R was studied, it was assumed that \R always
439        matched two bytes, thus causing the minimum subject length to be
440        incorrectly computed because \R can also match just one byte.
441    
442    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
443        was incorrectly computed.
444    
445    20. If /S is present twice on a test pattern in pcretest input, it now
446        *disables* studying, thereby overriding the use of -s on the command line
447        (see 14 above). This is necessary for one or two tests to keep the output
448        identical in both cases.
449    
450    21. When (*ACCEPT) was used in an assertion that matched an empty string and
451        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
452    
453    22. When an atomic group that contained a capturing parenthesis was
454        successfully matched, but the branch in which it appeared failed, the
455        capturing was not being forgotten if a higher numbered group was later
456        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
457        group 1 to "a", when in fact it should be unset. This applied to multi-
458        branched capturing and non-capturing groups, repeated or not, and also to
459        positive assertions (capturing in negative assertions does not happen
460        in PCRE) and also to nested atomic groups.
461    
462    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
463        subject after a captured substring, to make it easier to tell which of a
464        number of identical substrings has been captured.
465    
466    24. The way atomic groups are processed by pcre_exec() has been changed so that
467        if they are repeated, backtracking one repetition now resets captured
468        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
469        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
470        "aaa". Previously, it would have been "a". As part of this code
471        refactoring, the way recursive calls are handled has also been changed.
472    
473    25. If an assertion condition captured any substrings, they were not passed
474        back unless some other capturing happened later. For example, if
475        (?(?=(a))a) was matched against "a", no capturing was returned.
476    
477    26. When studying a pattern that contained subroutine calls or assertions,
478        the code for finding the minimum length of a possible match was handling
479        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
480        group 1 called group 2 while simultaneously a separate group 2 called group
481        1). A stack overflow occurred in this case. I have fixed this by limiting
482        the recursion depth to 10.
483    
484    27. Updated RunTest.bat in the distribution to the version supplied by Tom
485        Fortmann. This supports explicit test numbers on the command line, and has
486        argument validation and error reporting.
487    
488    28. An instance of \X with an unlimited repeat could fail if at any point the
489        first character it looked at was a mark character.
490    
491    29. Some minor code refactoring concerning Unicode properties and scripts
492        should reduce the stack requirement of match() slightly.
493    
494    30. Added the '=' option to pcretest to check the setting of unused capturing
495        slots at the end of the pattern, which are documented as being -1, but are
496        not included in the return count.
497    
498    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
499        compiled something random. Now it gives a compile-time error (as does
500        Perl).
501    
502    32. A *MARK encountered during the processing of a positive assertion is now
503        recorded and passed back (compatible with Perl).
504    
505    33. If --only-matching or --colour was set on a pcregrep call whose pattern
506        had alternative anchored branches, the search for a second match in a line
507        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
508        matched the line "0102" twice. The same bug affected patterns that started
509        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
510        twice.
511    
512    34. Previously, PCRE did not allow quantification of assertions. However, Perl
513        does, and because of capturing effects, quantifying parenthesized
514        assertions may at times be useful. Quantifiers are now allowed for
515        parenthesized assertions.
516    
517    35. A minor code tidy in pcre_compile() when checking options for \R usage.
518    
519    36. \g was being checked for fancy things in a character class, when it should
520        just be a literal "g".
521    
522    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
523        appearance of a nested POSIX class supersedes an apparent external class.
524        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
525        unescaped square brackets may also appear as part of class names. For
526        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
527        more like Perl. (But see 8.20/1 above.)
528    
529    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
530        was because it thought it was \N{name}, which is not supported).
531    
532    39. Add minix to OS list not supporting the -S option in pcretest.
533    
534    40. PCRE tries to detect cases of infinite recursion at compile time, but it
535        cannot analyze patterns in sufficient detail to catch mutual recursions
536        such as ((?1))((?2)). There is now a runtime test that gives an error if a
537        subgroup is called recursively as a subpattern for a second time at the
538        same position in the subject string. In previous releases this might have
539        been caught by the recursion limit, or it might have run out of stack.
540    
541    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
542        happen only once. PCRE was, however incorrectly giving a compile time error
543        "recursive call could loop indefinitely" because it cannot analyze the
544        pattern in sufficient detail. The compile time test no longer happens when
545        PCRE is compiling a conditional subpattern, but actual runaway loops are
546        now caught at runtime (see 40 above).
547    
548    42. It seems that Perl allows any characters other than a closing parenthesis
549        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
550        has been changed to be the same.
551    
552    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
553        as not to get warnings when autogen.sh is called. Also changed
554        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
555    
556    44. To help people who use pcregrep to scan files containing exceedingly long
557        lines, the following changes have been made:
558    
559        (a) The default value of the buffer size parameter has been increased from
560            8K to 20K. (The actual buffer used is three times this size.)
561    
562        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
563            PCRE is built.
564    
565        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
566            to be set at run time.
567    
568        (d) Numerical values in pcregrep options can be followed by K or M, for
569            example --buffer-size=50K.
570    
571        (e) If a line being scanned overflows pcregrep's buffer, an error is now
572            given and the return code is set to 2.
573    
574    45. Add a pointer to the latest mark to the callout data block.
575    
576    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
577        partial match of an empty string instead of no match. This was specific to
578        the use of ".".
579    
580    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
581        complete match instead of a partial match. This bug was dependent on both
582        the PCRE_UTF8 and PCRE_DOTALL options being set.
583    
584    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
585        starting byte set, because \b was not being ignored.
586    
587    
588    Version 8.12 15-Jan-2011
589    ------------------------
590    
591    1.  Fixed some typos in the markup of the man pages, and wrote a script that
592        checks for such things as part of the documentation building process.
593    
594    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
595        --match-limit and --recursion-limit options (added for 8.11). In
596        particular, this made one of the standard tests fail. (The integer value
597        went into the wrong half of a long int.)
598    
599    3.  If the --colour option was given to pcregrep with -v (invert match), it
600        did strange things, either producing crazy output, or crashing. It should,
601        of course, ignore a request for colour when reporting lines that do not
602        match.
603    
604    4.  Another pcregrep bug caused similar problems if --colour was specified with
605        -M (multiline) and the pattern match finished with a line ending.
606    
607    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
608        matched in multiline mode, the following line was shown as part of the
609        match. This seems wrong, so I have changed it.
610    
611    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
612        the check for further matches in the same line (so they could be coloured)
613        to overrun the end of the current line. If another match was found, it was
614        incorrectly shown (and then shown again when found in the next line).
615    
616    7.  If pcregrep was compiled under Windows, there was a reference to the
617        function pcregrep_exit() before it was defined. I am assuming this was
618        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
619        reported by a user. I've moved the definition above the reference.
620    
621    
622    Version 8.11 10-Dec-2010
623    ------------------------
624    
625    1.  (*THEN) was not working properly if there were untried alternatives prior
626        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
627        backtracked to try for "b" instead of moving to the next alternative branch
628        at the same level (in this case, to look for "c"). The Perl documentation
629        is clear that when (*THEN) is backtracked onto, it goes to the "next
630        alternative in the innermost enclosing group".
631    
632    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
633        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
634        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
635        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
636        (*THEN).
637    
638    3.  If \s appeared in a character class, it removed the VT character from
639        the class, even if it had been included by some previous item, for example
640        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
641        of \s, but is part of the POSIX "space" class.)
642    
643    4.  A partial match never returns an empty string (because you can always
644        match an empty string at the end of the subject); however the checking for
645        an empty string was starting at the "start of match" point. This has been
646        changed to the "earliest inspected character" point, because the returned
647        data for a partial match starts at this character. This means that, for
648        example, /(?<=abc)def/ gives a partial match for the subject "abc"
649        (previously it gave "no match").
650    
651    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
652        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
653        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
654        has an implication that the given string is incomplete (because a partial
655        match is preferred over a full match). For this reason, these items now
656        give a partial match in this situation. [Aside: previously, the one case
657        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
658        match rather than a full match, which was wrong by the old rules, but is
659        now correct.]
660    
661    6.  There was a bug in the handling of #-introduced comments, recognized when
662        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
663        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
664        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
665        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
666        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
667        places in pcre_compile().
668    
669    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
670        comments when looking ahead for named forward references to subpatterns,
671        the only newline sequence it recognized was NL. It now handles newlines
672        according to the set newline convention.
673    
674    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
675        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
676        cater for a lack of strerror(). These oversights have been fixed.
677    
678    9.  Added --match-limit and --recursion-limit to pcregrep.
679    
680    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
681    
682    11. When the -o option was used, pcregrep was setting a return code of 1, even
683        when matches were found, and --line-buffered was not being honoured.
684    
685    12. Added an optional parentheses number to the -o and --only-matching options
686        of pcregrep.
687    
688    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
689        can match an empty string. The code to do it in pcretest and pcredemo
690        needed fixing:
691    
692        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
693            only one byte after an empty string match just before CRLF (this case
694            just got forgotten; "any" and "anycrlf" were OK).
695    
696        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
697            mode when an empty string match preceded an ASCII character followed by
698            a non-ASCII character. (The code for advancing by one character rather
699            than one byte was nonsense.)
700    
701        (c) The pcredemo.c sample program did not have any code at all to handle
702            the cases when CRLF is a valid newline sequence.
703    
704    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
705        as a starting offset was within the subject string. There is now a new
706        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
707        negative or greater than the length of the string. In order to test this,
708        pcretest is extended to allow the setting of negative starting offsets.
709    
710    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
711        starting offset points to the beginning of a UTF-8 character was
712        unnecessarily clumsy. I tidied it up.
713    
714    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
715        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
716    
717    17. Nobody had reported that the --include_dir option, which was added in
718        release 7.7 should have been called --include-dir (hyphen, not underscore)
719        for compatibility with GNU grep. I have changed it to --include-dir, but
720        left --include_dir as an undocumented synonym, and the same for
721        --exclude-dir, though that is not available in GNU grep, at least as of
722        release 2.5.4.
723    
724    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
725        characters from a string of bytes) have been redefined so as not to use
726        loops, in order to improve performance in some environments. At the same
727        time, I abstracted some of the common code into auxiliary macros to save
728        repetition (this should not affect the compiled code).
729    
730    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
731        compile-time error is now given if \c is not followed by an ASCII
732        character, that is, a byte less than 128. (In EBCDIC mode, the code is
733        different, and any byte value is allowed.)
734    
735    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
736        START_OPTIMIZE option, which is now allowed at compile time - but just
737        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
738        to pcregrep and other applications that have no direct access to PCRE
739        options. The new /Y option in pcretest sets this option when calling
740        pcre_compile().
741    
742    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
743        back references. Groups containing recursive back references were forced to
744        be atomic by that change, but in the case of named groups, the amount of
745        memory required was incorrectly computed, leading to "Failed: internal
746        error: code overflow". This has been fixed.
747    
748    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
749        pcretest.c, to avoid build problems in some Borland environments.
750    
751    
752    Version 8.10 25-Jun-2010
753    ------------------------
754    
755    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
756        THEN.
757    
758    2.  (*ACCEPT) was not working when inside an atomic group.
759    
760    3.  Inside a character class, \B is treated as a literal by default, but
761        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
762        causes the error). The code is unchanged, but I tidied the documentation.
763    
764    4.  Inside a character class, PCRE always treated \R and \X as literals,
765        whereas Perl faults them if its -w option is set. I have changed PCRE so
766        that it faults them when PCRE_EXTRA is set.
767    
768    5.  Added support for \N, which always matches any character other than
769        newline. (It is the same as "." when PCRE_DOTALL is not set.)
770    
771    6.  When compiling pcregrep with newer versions of gcc which may have
772        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
773        declared with attribute warn_unused_result" were given. Just casting the
774        result to (void) does not stop the warnings; a more elaborate fudge is
775        needed. I've used a macro to implement this.
776    
777    7.  Minor change to pcretest.c to avoid a compiler warning.
778    
779    8.  Added four artifical Unicode properties to help with an option to make
780        \s etc use properties (see next item). The new properties are: Xan
781        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
782    
783    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
784        use Unicode properties. (*UCP) at the start of a pattern can be used to set
785        this option. Modified pcretest to add /W to test this facility. Added
786        REG_UCP to make it available via the POSIX interface.
787    
788    10. Added --line-buffered to pcregrep.
789    
790    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
791        studied, and the match started with a letter with a code point greater than
792        127 whose first byte was different to the first byte of the other case of
793        the letter, the other case of this starting letter was not recognized
794        (#976).
795    
796    12. If a pattern that was studied started with a repeated Unicode property
797        test, for example, \p{Nd}+, there was the theoretical possibility of
798        setting up an incorrect bitmap of starting bytes, but fortunately it could
799        not have actually happened in practice until change 8 above was made (it
800        added property types that matched character-matching opcodes).
801    
802    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
803        possible starting bytes for non-anchored patterns.
804    
805    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
806        \R, and also a number of cases that involve Unicode properties, both
807        explicit and implicit when PCRE_UCP is set.
808    
809    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
810        input, it could crash or give wrong results if characters with values
811        greater than 0xc0 were present in the subject string. (Detail: it assumed
812        UTF-8 input when processing these items.)
813    
814    16. Added a lot of (int) casts to avoid compiler warnings in systems where
815        size_t is 64-bit (#991).
816    
817    17. Added a check for running out of memory when PCRE is compiled with
818        --disable-stack-for-recursion (#990).
819    
820    18. If the last data line in a file for pcretest does not have a newline on
821        the end, a newline was missing in the output.
822    
823    19. The default pcre_chartables.c file recognizes only ASCII characters (values
824        less than 128) in its various bitmaps. However, there is a facility for
825        generating tables according to the current locale when PCRE is compiled. It
826        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
827        characters, are recognized by isspace() and therefore were getting set in
828        these tables, and indeed these tables seem to approximate to ISO 8859. This
829        caused a problem in UTF-8 mode when pcre_study() was used to create a list
830        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
831        which of course cannot start UTF-8 characters. I have changed the code so
832        that only real ASCII characters (less than 128) and the correct starting
833        bytes for UTF-8 encodings are set for characters greater than 127 when in
834        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
835        altogether.)
836    
837    20. Added the /T option to pcretest so as to be able to run tests with non-
838        standard character tables, thus making it possible to include the tests
839        used for 19 above in the standard set of tests.
840    
841    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
842        reference to a subpattern the other side of a comment that contains an
843        opening parenthesis caused either an internal compiling error, or a
844        reference to the wrong subpattern.
845    
846    
847    Version 8.02 19-Mar-2010
848    ------------------------
849    
850    1.  The Unicode data tables have been updated to Unicode 5.2.0.
851    
852    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
853        configured.
854    
855    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
856        original author of that file, following a query about its status.
857    
858    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
859        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
860    
861    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
862        quantifier applied to a forward-referencing subroutine call, could compile
863        incorrect code or give the error "internal error: previously-checked
864        referenced subpattern not found".
865    
866    6.  Both MS Visual Studio and Symbian OS have problems with initializing
867        variables to point to external functions. For these systems, therefore,
868        pcre_malloc etc. are now initialized to local functions that call the
869        relevant global functions.
870    
871    7.  There were two entries missing in the vectors called coptable and poptable
872        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
873        I've fixed the data, and added a kludgy way of testing at compile time that
874        the lengths are correct (equal to the number of opcodes).
875    
876    8.  Following on from 7, I added a similar kludge to check the length of the
877        eint vector in pcreposix.c.
878    
879    9.  Error texts for pcre_compile() are held as one long string to avoid too
880        much relocation at load time. To find a text, the string is searched,
881        counting zeros. There was no check for running off the end of the string,
882        which could happen if a new error number was added without updating the
883        string.
884    
885    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
886    
887    11. \K was not working if it appeared in an atomic group or in a group that
888        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
889        \K is "not well defined" if used in an assertion. PCRE now accepts it if
890        the assertion is positive, but not if it is negative.
891    
892    12. Change 11 fortuitously reduced the size of the stack frame used in the
893        "match()" function of pcre_exec.c by one pointer. Forthcoming
894        implementation of support for (*MARK) will need an extra pointer on the
895        stack; I have reserved it now, so that the stack frame size does not
896        decrease.
897    
898    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
899        item in branch that calls a recursion is a subroutine call - as in the
900        second branch in the above example - was incorrectly given the compile-
901        time error "recursive call could loop indefinitely" because pcre_compile()
902        was not correctly checking the subroutine for matching a non-empty string.
903    
904    14. The checks for overrunning compiling workspace could trigger after an
905        overrun had occurred. This is a "should never occur" error, but it can be
906        triggered by pathological patterns such as hundreds of nested parentheses.
907        The checks now trigger 100 bytes before the end of the workspace.
908    
909    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
910    
911    
912    Version 8.01 19-Jan-2010
913    ------------------------
914    
915    1.  If a pattern contained a conditional subpattern with only one branch (in
916        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
917        computed the wrong minimum data length (which is of course zero for such
918        subpatterns). This could cause incorrect "no match" results.
919    
920    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
921        the pattern is reset in the first branch, pcre_compile() failed with
922        "internal error: code overflow at offset...". This happened only when
923        the reset was to the original external option setting. (An optimization
924        abstracts leading options settings into an external setting, which was the
925        cause of this.)
926    
927    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
928        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
929        assertion pattern did not match (meaning that the assertion was true), it
930        was incorrectly treated as false if the SKIP had been reached during the
931        matching. This also applied to assertions used as conditions.
932    
933    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
934        assertion subpattern, including such a pattern used as a condition,
935        unpredictable results occurred, instead of the error return
936        PCRE_ERROR_DFA_UITEM.
937    
938    5.  The C++ GlobalReplace function was not working like Perl for the special
939        situation when an empty string is matched. It now does the fancy magic
940        stuff that is necessary.
941    
942    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
943        removed. (These were left over from very, very early versions of PCRE.)
944    
945    7.  Some cosmetic changes to the code to make life easier when compiling it
946        as part of something else:
947    
948        (a) Change DEBUG to PCRE_DEBUG.
949    
950        (b) In pcre_compile(), rename the member of the "branch_chain" structure
951            called "current" as "current_branch", to prevent a collision with the
952            Linux macro when compiled as a kernel module.
953    
954        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
955            prevent a collision with the Linux macro when compiled as a kernel
956            module.
957    
958    8.  In pcre_compile() there are some checks for integer overflows that used to
959        cast potentially large values to (double). This has been changed to that
960        when building, a check for int64_t is made, and if it is found, it is used
961        instead, thus avoiding the use of floating point arithmetic. (There is no
962        other use of FP in PCRE.) If int64_t is not found, the fallback is to
963        double.
964    
965    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
966        2005 (difference between two addresses compared to an unsigned value).
967    
968    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
969        custom one, because of the following reported problem in Windows:
970    
971          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
972              under Win32.
973          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
974              therefore missing the function definition.
975          - The compiler thus generates a "C" signature for the test function.
976          - The linker fails to find the "C" function.
977          - PCRE fails to configure if asked to do so against libbz2.
978    
979    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
980        messages were output:
981    
982          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
983          rerunning libtoolize, to keep the correct libtool macros in-tree.
984          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
985    
986        I have done both of these things.
987    
988    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
989        most of the time, it *can* run out if it is given a pattern that contains a
990        runaway infinite recursion. I updated the discussion in the pcrestack man
991        page.
992    
993    13. Now that we have gone to the x.xx style of version numbers, the minor
994        version may start with zero. Using 08 or 09 is a bad idea because users
995        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
996        interpreted as invalid octal numbers. I've updated the previous comment in
997        configure.ac, and also added a check that gives an error if 08 or 09 are
998        used.
999    
1000    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1001        causing partial matching to fail when the end of the subject matched \W
1002        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1003    
1004    15. There were some discrepancies between the declarations in pcre_internal.h
1005        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1006        their definitions. The declarations used "const uschar *" and the
1007        definitions used USPTR. Even though USPTR is normally defined as "const
1008        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1009        reported that: "This difference in casting confuses some C++ compilers, for
1010        example, SunCC recognizes above declarations as different functions and
1011        generates broken code for hbpcre." I have changed the declarations to use
1012        USPTR.
1013    
1014    16. GNU libtool is named differently on some systems. The autogen.sh script now
1015        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1016        (FreeBSD).
1017    
1018    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1019        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1020        comment: "Figure out how to create a longlong from a string: strtoll and
1021        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1022        instance, but it only takes 2 args instead of 3!"
1023    
1024    18. A subtle bug concerned with back references has been fixed by a change of
1025        specification, with a corresponding code fix. A pattern such as
1026        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1027        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1028        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1029        same bug. Such groups have to be quantified to be useful, or contained
1030        inside another quantified group. (If there's no repetition, the reference
1031        can never match.) The problem arises because, having left the group and
1032        moved on to the rest of the pattern, a later failure that backtracks into
1033        the group uses the captured value from the final iteration of the group
1034        rather than the correct earlier one. I have fixed this in PCRE by forcing
1035        any group that contains a reference to itself to be an atomic group; that
1036        is, there cannot be any backtracking into it once it has completed. This is
1037        similar to recursive and subroutine calls.
1038    
1039    
1040    Version 8.00 19-Oct-09
1041    ----------------------
1042    
1043    1.  The table for translating pcre_compile() error codes into POSIX error codes
1044        was out-of-date, and there was no check on the pcre_compile() error code
1045        being within the table. This could lead to an OK return being given in
1046        error.
1047    
1048    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1049        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1050        in a Windows environment.
1051    
1052    3.  The pcregrep --count option prints the count for each file even when it is
1053        zero, as does GNU grep. However, pcregrep was also printing all files when
1054        --files-with-matches was added. Now, when both options are given, it prints
1055        counts only for those files that have at least one match. (GNU grep just
1056        prints the file name in this circumstance, but including the count seems
1057        more useful - otherwise, why use --count?) Also ensured that the
1058        combination -clh just lists non-zero counts, with no names.
1059    
1060    4.  The long form of the pcregrep -F option was incorrectly implemented as
1061        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1062        but it seems right to fix it, and I didn't think it was worth preserving
1063        the old behaviour.
1064    
1065    5.  The command line items --regex=pattern and --regexp=pattern were not
1066        recognized by pcregrep, which required --regex pattern or --regexp pattern
1067        (with a space rather than an '='). The man page documented the '=' forms,
1068        which are compatible with GNU grep; these now work.
1069    
1070    6.  No libpcreposix.pc file was created for pkg-config; there was just
1071        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1072    
1073    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1074        when UCP support is not needed, by modifying the Python script that
1075        generates it from Unicode data files. This should not matter if the module
1076        is correctly used as a library, but I received one complaint about 50K of
1077        unwanted data. My guess is that the person linked everything into his
1078        program rather than using a library. Anyway, it does no harm.
1079    
1080    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1081        was a minimum greater than 1 for a wide character in a possessive
1082        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1083        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1084        character. Chaos in the form of incorrect output or a compiling loop could
1085        result.
1086    
1087    9.  The restrictions on what a pattern can contain when partial matching is
1088        requested for pcre_exec() have been removed. All patterns can now be
1089        partially matched by this function. In addition, if there are at least two
1090        slots in the offset vector, the offset of the earliest inspected character
1091        for the match and the offset of the end of the subject are set in them when
1092        PCRE_ERROR_PARTIAL is returned.
1093    
1094    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1095        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1096        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1097        and may be more useful for multi-segment matching.
1098    
1099    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1100        used to be given if ever the end of the subject was reached; now it is
1101        given only if matching could not proceed because another character was
1102        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1103        string "Z", which now yields "no match" instead of "partial match". In the
1104        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1105        final character ended with (*FAIL).
1106    
1107    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1108        if the pattern had a "must contain" character that was already found in the
1109        earlier partial match, unless partial matching was again requested. For
1110        example, with the pattern /dog.(body)?/, the "must contain" character is
1111        "g". If the first part-match was for the string "dog", restarting with
1112        "sbody" failed. This bug has been fixed.
1113    
1114    13. The string returned by pcre_dfa_exec() after a partial match has been
1115        changed so that it starts at the first inspected character rather than the
1116        first character of the match. This makes a difference only if the pattern
1117        starts with a lookbehind assertion or \b or \B (\K is not supported by
1118        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1119        matching functions compatible, and I think it's the right thing to do.
1120    
1121    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1122        so that the demonstration program is easily available in environments where
1123        PCRE has not been installed from source.
1124    
1125    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1126        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1127        library.
1128    
1129    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1130        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1131        is not the first non-POSIX option to be added. Clearly some people find
1132        these options useful.
1133    
1134    17. If a caller to the POSIX matching function regexec() passes a non-zero
1135        value for nmatch with a NULL value for pmatch, the value of
1136        nmatch is forced to zero.
1137    
1138    18. RunGrepTest did not have a test for the availability of the -u option of
1139        the diff command, as RunTest does. It now checks in the same way as
1140        RunTest, and also checks for the -b option.
1141    
1142    19. If an odd number of negated classes containing just a single character
1143        interposed, within parentheses, between a forward reference to a named
1144        subpattern and the definition of the subpattern, compilation crashed with
1145        an internal error, complaining that it could not find the referenced
1146        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1147        [The bug was that it was starting one character too far in when skipping
1148        over the character class, thus treating the ] as data rather than
1149        terminating the class. This meant it could skip too much.]
1150    
1151    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1152        /g option in pcretest when the pattern contains \K, which makes it possible
1153        to have an empty string match not at the start, even when the pattern is
1154        anchored. Updated pcretest and pcredemo to use this option.
1155    
1156    21. If the maximum number of capturing subpatterns in a recursion was greater
1157        than the maximum at the outer level, the higher number was returned, but
1158        with unset values at the outer level. The correct (outer level) value is
1159        now given.
1160    
1161    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1162        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1163        make it do so. The string so far is captured, making this feature
1164        compatible with Perl.
1165    
1166    23. The tests have been re-organized, adding tests 11 and 12, to make it
1167        possible to check the Perl 5.10 features against Perl 5.10.
1168    
1169    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1170        pattern matches a fixed length string. PCRE did not allow this; now it
1171        does. Neither allows recursion.
1172    
1173    25. I finally figured out how to implement a request to provide the minimum
1174        length of subject string that was needed in order to match a given pattern.
1175        (It was back references and recursion that I had previously got hung up
1176        on.) This code has now been added to pcre_study(); it finds a lower bound
1177        to the length of subject needed. It is not necessarily the greatest lower
1178        bound, but using it to avoid searching strings that are too short does give
1179        some useful speed-ups. The value is available to calling programs via
1180        pcre_fullinfo().
1181    
1182    26. While implementing 25, I discovered to my embarrassment that pcretest had
1183        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1184        study optimizations had never been tested with that matching function.
1185        Oops. What is worse, even when it was passed study data, there was a bug in
1186        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1187        were also very few tests of studied patterns with pcre_dfa_exec().
1188    
1189    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1190        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1191        on the other side of the coin, they are no longer allowed to have different
1192        names, because these cannot be distinguished in PCRE, and this has caused
1193        confusion. (This is a difference from Perl.)
1194    
1195    28. When duplicate subpattern names are present (necessarily with different
1196        numbers, as required by 27 above), and a test is made by name in a
1197        conditional pattern, either for a subpattern having been matched, or for
1198        recursion in such a pattern, all the associated numbered subpatterns are
1199        tested, and the overall condition is true if the condition is true for any
1200        one of them. This is the way Perl works, and is also more like the way
1201        testing by number works.
1202    
1203    
1204    Version 7.9 11-Apr-09
1205    ---------------------
1206    
1207    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1208        (pcretest), all targets were linked against these libraries. This included
1209        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1210        libraries. This caused unwanted dependencies to be created. This problem
1211        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1212        pcretest is linked with readline.
1213    
1214    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1215        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1216        moved outside it again, because FALSE and TRUE are already defined in AIX,
1217        but BOOL is not.
1218    
1219    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1220        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1221    
1222    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1223        hyphens) following file names and line numbers when outputting matching
1224        lines. This is not true; no spaces are inserted. I have also clarified the
1225        wording for the --colour (or --color) option.
1226    
1227    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1228        was not coloured; this is different to GNU grep, so I have changed it to be
1229        the same.
1230    
1231    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1232        each matching line was coloured. Now it goes on to look for further matches
1233        of any of the test patterns, which is the same behaviour as GNU grep.
1234    
1235    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1236        doesn't make sense to accept an empty string match in pcregrep, so I have
1237        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1238        seems to be how GNU grep behaves.
1239    
1240    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1241        start or after a newline", because the conditional assertion was not being
1242        correctly handled. The rule now is that both the assertion and what follows
1243        in the first alternative must satisfy the test.
1244    
1245    9.  If auto-callout was enabled in a pattern with a conditional group whose
1246        condition was an assertion, PCRE could crash during matching, both with
1247        pcre_exec() and pcre_dfa_exec().
1248    
1249    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1250        used for matching.
1251    
1252    11. Unicode property support in character classes was not working for
1253        characters (bytes) greater than 127 when not in UTF-8 mode.
1254    
1255    12. Added the -M command line option to pcretest.
1256    
1257    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1258    
1259    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1260    
1261    16. Added comments and documentation about mis-use of no_arg in the C++
1262        wrapper.
1263    
1264    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1265        from Martin Jerabek that uses macro names for all relevant character and
1266        string constants.
1267    
1268    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1269        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1270        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1271        these, but not everybody uses configure.
1272    
1273    19. A conditional group that had only one branch was not being correctly
1274        recognized as an item that could match an empty string. This meant that an
1275        enclosing group might also not be so recognized, causing infinite looping
1276        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1277        with the subject "ab", where knowledge that the repeated group can match
1278        nothing is needed in order to break the loop.
1279    
1280    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1281        exec(), but without supplying a callout function, matching went wrong.
1282    
1283    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1284        leak if the size of the offset vector was greater than 30. When the vector
1285        is smaller, the saved offsets during recursion go onto a local stack
1286        vector, but for larger vectors malloc() is used. It was failing to free
1287        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1288        error, in fact).
1289    
1290    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1291        heapframe that is used only when UTF-8 support is enabled. This caused no
1292        problem, but was untidy.
1293    
1294    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1295        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1296        included within another project.
1297    
1298    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1299        slightly modified by me:
1300    
1301          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1302              not building pcregrep.
1303    
1304          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1305              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1306    
1307    25. Forward references, both numeric and by name, in patterns that made use of
1308        duplicate group numbers, could behave incorrectly or give incorrect errors,
1309        because when scanning forward to find the reference group, PCRE was not
1310        taking into account the duplicate group numbers. A pattern such as
1311        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1312    
1313    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1314        the feature of a custom pointer more persuasive (as requested by a user).
1315    
1316    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1317        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1318        pre-defined.
1319    
1320    28. Added support for (*UTF8) at the start of a pattern.
1321    
1322    29. Arrange for flags added by the "release type" setting in CMake to be shown
1323        in the configuration summary.
1324    
1325    
1326    Version 7.8 05-Sep-08
1327    ---------------------
1328    
1329    1.  Replaced UCP searching code with optimized version as implemented for Ad
1330        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1331        stage table and inline lookup instead of a function, giving speed ups of 2
1332        to 5 times on some simple patterns that I tested. Permission was given to
1333        distribute the MultiStage2.py script that generates the tables (it's not in
1334        the tarball, but is in the Subversion repository).
1335    
1336    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1337        scripts.
1338    
1339    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1340        a group with a zero qualifier. The result of the study could be incorrect,
1341        or the function might crash, depending on the pattern.
1342    
1343    4.  Caseless matching was not working for non-ASCII characters in back
1344        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1345        It now works when Unicode Property Support is available.
1346    
1347    5.  In pcretest, an escape such as \x{de} in the data was always generating
1348        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1349        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1350        truncation.
1351    
1352    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1353    
1354    7.  Added two (int) casts to pcregrep when printing the difference of two
1355        pointers, in case they are 64-bit values.
1356    
1357    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1358        test 2 if it fails.
1359    
1360    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1361        and a #define of that name to empty if it is not externally set. This is to
1362        allow users of MSVC to set it if necessary.
1363    
1364    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1365        the convenience functions in the pcre_get.c source file.
1366    
1367    11. An option change at the start of a pattern that had top-level alternatives
1368        could cause overwriting and/or a crash. This command provoked a crash in
1369        some environments:
1370    
1371          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1372    
1373        This potential security problem was recorded as CVE-2008-2371.
1374    
1375    12. For a pattern where the match had to start at the beginning or immediately
1376        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1377        pcre_dfa_exec() could read past the end of the passed subject if there was
1378        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1379        pcretest so that it places the subject at the end of its malloc-ed buffer.
1380    
1381    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1382        exec() might read past the end of the data buffer in UTF-8 mode.
1383    
1384    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1385        the data contained the byte 0x85 as part of a UTF-8 character within its
1386        first line. This applied both to normal and DFA matching.
1387    
1388    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1389        /^[^d]*?$/8 failed to match "abc".
1390    
1391    16. Added a missing copyright notice to pcrecpp_internal.h.
1392    
1393    17. Make it more clear in the documentation that values returned from
1394        pcre_exec() in ovector are byte offsets, not character counts.
1395    
1396    18. Tidied a few places to stop certain compilers from issuing warnings.
1397    
1398    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1399        supplied by Stefan Weber. I made a further small update for 7.8 because
1400        there is a change of source arrangements: the pcre_searchfuncs.c module is
1401        replaced by pcre_ucd.c.
1402    
1403    
1404    Version 7.7 07-May-08
1405    ---------------------
1406    
1407    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1408        a string to a long long, pretend we don't even have a long long." This is
1409        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1410    
1411    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1412        pre-7.6 versions, which defined a global no_arg variable instead of putting
1413        it in the RE class. (See also #8 below.)
1414    
1415    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1416        Lopes.
1417    
1418    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1419    
1420        (1) The include/exclude patterns were being applied to the whole pathnames
1421            of files, instead of just to the final components.
1422    
1423        (2) If there was more than one level of directory, the subdirectories were
1424            skipped unless they satisfied the include/exclude conditions. This is
1425            inconsistent with GNU grep (and could even be seen as contrary to the
1426            pcregrep specification - which I improved to make it absolutely clear).
1427            The action now is always to scan all levels of directory, and just
1428            apply the include/exclude patterns to regular files.
1429    
1430    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1431        --exclude_dir in the tests to avoid scanning .svn directories.
1432    
1433    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1434        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1435        doesn't support NULs in patterns.
1436    
1437    7.  Added some missing "const"s to declarations of static tables in
1438        pcre_compile.c and pcre_dfa_exec.c.
1439    
1440    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1441        caused by fix #2  above. (Subsequently also a second patch to fix the
1442        first patch. And a third patch - this was a messy problem.)
1443    
1444    9.  Applied Craig's patch to remove the use of push_back().
1445    
1446    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1447        matching function regexec().
1448    
1449    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1450        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1451        references. PCRE supports relative numbers with this syntax (I don't think
1452        Oniguruma does).
1453    
1454    12. Previously, a group with a zero repeat such as (...){0} was completely
1455        omitted from the compiled regex. However, this means that if the group
1456        was called as a subroutine from elsewhere in the pattern, things went wrong
1457        (an internal error was given). Such groups are now left in the compiled
1458        pattern, with a new opcode that causes them to be skipped at execution
1459        time.
1460    
1461    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1462        to the way PCRE behaves:
1463    
1464        (a) A lone ] character is dis-allowed (Perl treats it as data).
1465    
1466        (b) A back reference to an unmatched subpattern matches an empty string
1467            (Perl fails the current match path).
1468    
1469        (c) A data ] in a character class must be notated as \] because if the
1470            first data character in a class is ], it defines an empty class. (In
1471            Perl it is not possible to have an empty class.) The empty class []
1472            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1473            The negative empty class [^] matches any one character, independently
1474            of the DOTALL setting.
1475    
1476    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1477        non-existent subpattern following a character class starting with ']' and
1478        containing () gave an internal compiling error instead of "reference to
1479        non-existent subpattern". Fortunately, when the pattern did exist, the
1480        compiled code was correct. (When scanning forwards to check for the
1481        existencd of the subpattern, it was treating the data ']' as terminating
1482        the class, so got the count wrong. When actually compiling, the reference
1483        was subsequently set up correctly.)
1484    
1485    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1486        it was being rejected as not supported by pcre_dfa_exec(), even though
1487        other assertions are supported. I have made pcre_dfa_exec() support
1488        (*FAIL).
1489    
1490    16. The implementation of 13c above involved the invention of a new opcode,
1491        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1492        cannot be changed at match time, I realized I could make a small
1493        improvement to matching performance by compiling OP_ALLANY instead of
1494        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1495        on the OP_ANY path.
1496    
1497    17. Compiling pcretest on Windows with readline support failed without the
1498        following two fixes: (1) Make the unistd.h include conditional on
1499        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1500    
1501    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1502        ncurses library to be included for pcretest when ReadLine support is
1503        requested, but also to allow for it to be overridden. This patch came from
1504        Daniel Bergström.
1505    
1506    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1507        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1508        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1509        spotting this.
1510    
1511    
1512    Version 7.6 28-Jan-08
1513    ---------------------
1514    
1515    1.  A character class containing a very large number of characters with
1516        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1517        overflow.
1518    
1519    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1520        HAVE_LONG_LONG is not defined.
1521    
1522    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1523        bring them up to date and include new features. This patch includes:
1524    
1525        - Fixed PH's badly added libz and libbz2 support.
1526        - Fixed a problem with static linking.
1527        - Added pcredemo. [But later removed - see 7 below.]
1528        - Fixed dftables problem and added an option.
1529        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1530            HAVE_LONG_LONG.
1531        - Added readline support for pcretest.
1532        - Added an listing of the option settings after cmake has run.
1533    
1534    4.  A user submitted a patch to Makefile that makes it easy to create
1535        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1536        Makefile.am that cause it to include this special target, without
1537        affecting anything else. Note that the same mingw target plus all
1538        the other distribution libraries and programs are now supported
1539        when configuring with CMake (see 6 below) instead of with
1540        Configure/Make.
1541    
1542    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1543        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1544        exported in the Windows port". It has not yet been confirmed that the patch
1545        solves the problem, but it does no harm.
1546    
1547    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1548        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1549        with CMake, and also correct the comment about stack recursion.
1550    
1551    7.  Remove the automatic building of pcredemo from the ./configure system and
1552        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1553        of a program that users should build themselves after PCRE is installed, so
1554        building it automatically is not really right. What is more, it gave
1555        trouble in some build environments.
1556    
1557    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1558    
1559    
1560    Version 7.5 10-Jan-08
1561    ---------------------
1562    
1563    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1564        values in parens when parsing an RE using the C++ wrapper."
1565    
1566    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1567        Characters greater than 255 were excluded from the class instead of being
1568        included.
1569    
1570    3.  The same bug as (2) above applied to negated POSIX classes such as
1571        [:^space:].
1572    
1573    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1574        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1575        I have changed it.
1576    
1577    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1578        first named subpattern) and a construct such as (?&a) would reference the
1579        first named subpattern whose name started with "a" (in other words, the
1580        length check was missing). Both these problems are fixed. "Subpattern name
1581        expected" is now given for (?&) (a zero-length name), and this patch also
1582        makes it give the same error for \k'' (previously it complained that that
1583        was a reference to a non-existent subpattern).
1584    
1585    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1586        this is right because (?- can be followed by option settings as well as by
1587        digits. I have, however, made the messages clearer.
1588    
1589    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1590        than the number used in the conditional) now cause a compile-time error.
1591        This is actually not compatible with Perl, which accepts such patterns, but
1592        treats the conditional as always being FALSE (as PCRE used to), but it
1593        seems to me that giving a diagnostic is better.
1594    
1595    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1596        and messages.
1597    
1598    9.  Fix two occurrences of "backslash" in comments that should have been
1599        "backspace".
1600    
1601    10. Remove two redundant lines of code that can never be obeyed (their function
1602        was moved elsewhere).
1603    
1604    11. The program that makes PCRE's Unicode character property table had a bug
1605        which caused it to generate incorrect table entries for sequences of
1606        characters that have the same character type, but are in different scripts.
1607        It amalgamated them into a single range, with the script of the first of
1608        them. In other words, some characters were in the wrong script. There were
1609        thirteen such cases, affecting characters in the following ranges:
1610    
1611          U+002b0 - U+002c1
1612          U+0060c - U+0060d
1613          U+0061e - U+00612
1614          U+0064b - U+0065e
1615          U+0074d - U+0076d
1616          U+01800 - U+01805
1617          U+01d00 - U+01d77
1618          U+01d9b - U+01dbf
1619          U+0200b - U+0200f
1620          U+030fc - U+030fe
1621          U+03260 - U+0327f
1622          U+0fb46 - U+0fbb1
1623          U+10450 - U+1049d
1624    
1625    12. The -o option (show only the matching part of a line) for pcregrep was not
1626        compatible with GNU grep in that, if there was more than one match in a
1627        line, it showed only the first of them. It now behaves in the same way as
1628        GNU grep.
1629    
1630    13. If the -o and -v options were combined for pcregrep, it printed a blank
1631        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1632        does the same. The return code can be used to tell if there were any
1633        non-matching lines.
1634    
1635    14. Added --file-offsets and --line-offsets to pcregrep.
1636    
1637    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1638        infinitely looping recursion. The bug was that positive lookaheads were not
1639        being skipped when checking for a possible empty match (negative lookaheads
1640        and both kinds of lookbehind were skipped).
1641    
1642    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1643        inclusion of <windows.h> to before rather than after the definition of
1644        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1645    
1646    17. Specifying a possessive quantifier with a specific limit for a Unicode
1647        character property caused pcre_compile() to compile bad code, which led at
1648        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1649        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1650        caused the error; without that there was no problem.
1651    
1652    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1653    
1654    19. Added --enable-pcretest-libreadline.
1655    
1656    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1657        RE::GlobalReplace(). As a result, the number of replacements returned was
1658        double what it should be. I removed one of the increments, but Craig sent a
1659        later patch that removed the other one (the right fix) and added unit tests
1660        that check the return values (which was not done before).
1661    
1662    21. Several CMake things:
1663    
1664        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1665            the names libpcre and libpcreposix, not just pcre and pcreposix.
1666    
1667        (2) The above change means that pcretest and pcregrep are now correctly
1668            linked with the newly-built libraries, not previously installed ones.
1669    
1670        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1671    
1672    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1673        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1674        UTF-8 newline character). The key issue is that the pattern starts .*;
1675        this means that the match must be either at the beginning, or after a
1676        newline. The bug was in the code for advancing after a failed match and
1677        checking that the new position followed a newline. It was not taking
1678        account of UTF-8 characters correctly.
1679    
1680    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1681        character classes. PCRE was not treating the sequence [:...:] as a
1682        character class unless the ... were all letters. Perl, however, seems to
1683        allow any characters between [: and :], though of course it rejects as
1684        unknown any "names" that contain non-letters, because all the known class
1685        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1686        for example, whereas PCRE did not - it did not recognize a POSIX character
1687        class. This seemed a bit dangerous, so the code has been changed to be
1688        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1689        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1690        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1691        Perl does, and where it didn't before.
1692    
1693    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1694        Windows environments %n is disabled by default.
1695    
1696    
1697    Version 7.4 21-Sep-07
1698    ---------------------
1699    
1700    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1701        means that a class such as [\s] counted as "explicit reference to CR or
1702        LF". That isn't really right - the whole point of the change was to try to
1703        help when there was an actual mention of one of the two characters. So now
1704        the change happens only if \r or \n (or a literal CR or LF) character is
1705        encountered.
1706    
1707    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1708        of both had grown to the point where there were only 3 bits left.
1709        Fortunately, there was spare space in the data structure, and so I have
1710        moved the internal flags into a new 16-bit field to free up more option
1711        bits.
1712    
1713    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1714        but did not set the internal JCHANGED flag - either of these is enough to
1715        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1716        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1717        start sets both bits.
1718    
1719    4.  Added options (at build time, compile time, exec time) to change \R from
1720        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1721    
1722    5.  doc/pcresyntax.html was missing from the distribution.
1723    
1724    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1725        compatibility, even though it is no longer used.
1726    
1727    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1728        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1729        windows.h file is present (where different names are used). [This was
1730        reversed later after testing - see 16 below.]
1731    
1732    8.  Changed all #include <config.h> to #include "config.h". There were also
1733        some further <pcre.h> cases that I changed to "pcre.h".
1734    
1735    9.  When pcregrep was used with the --colour option, it missed the line ending
1736        sequence off the lines that it output.
1737    
1738    10. It was pointed out to me that arrays of string pointers cause lots of
1739        relocations when a shared library is dynamically loaded. A technique of
1740        using a single long string with a table of offsets can drastically reduce
1741        these. I have refactored PCRE in four places to do this. The result is
1742        dramatic:
1743    
1744          Originally:                          290
1745          After changing UCP table:            187
1746          After changing error message table:   43
1747          After changing table of "verbs"       36
1748          After changing table of Posix names   22
1749    
1750        Thanks to the folks working on Gregex for glib for this insight.
1751    
1752    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1753        unicode-properties was also set.
1754    
1755    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1756    
1757    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1758        checked only for CRLF.
1759    
1760    14. Added casts to pcretest.c to avoid compiler warnings.
1761    
1762    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1763    
1764    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1765        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1766        entirely. This removes changes made in 7 above.
1767    
1768    17. The CMake files have been updated, and there is now more information about
1769        building with CMake in the NON-UNIX-USE document.
1770    
1771    
1772    Version 7.3 28-Aug-07
1773    ---------------------
1774    
1775     1. In the rejigging of the build system that eventually resulted in 7.1, the
1776        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1777        brackets there is not right, since it causes compilers to look for an
1778        installed pcre.h, not the version that is in the source that is being
1779        compiled (which of course may be different). I have changed it back to:
1780    
1781          #include "pcre.h"
1782    
1783        I have a vague recollection that the change was concerned with compiling in
1784        different directories, but in the new build system, that is taken care of
1785        by the VPATH setting the Makefile.
1786    
1787     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1788        when the subject happened to end in the byte 0x85 (e.g. if the last
1789        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1790        characters but of course it shouldn't be taken as a newline when it is part
1791        of another character. The bug was that, for an unlimited repeat of . in
1792        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1793        characters when looking for a newline.
1794    
1795     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1796    
1797     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1798        in debug output.
1799    
1800     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1801        long printing in the pcrecpp unittest when running under MinGW.
1802    
1803     6. ESC_K was left out of the EBCDIC table.
1804    
1805     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1806        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1807        limit also applies to "virtual nesting" when a pattern is recursive, and in
1808        this case 1000 isn't so big. I have been able to remove this limit at the
1809        expense of backing off one optimization in certain circumstances. Normally,
1810        when pcre_exec() would call its internal match() function recursively and
1811        immediately return the result unconditionally, it uses a "tail recursion"
1812        feature to save stack. However, when a subpattern that can match an empty
1813        string has an unlimited repetition quantifier, it no longer makes this
1814        optimization. That gives it a stack frame in which to save the data for
1815        checking that an empty string has been matched. Previously this was taken
1816        from the 1000-entry workspace that had been reserved. So now there is no
1817        explicit limit, but more stack is used.
1818    
1819     8. Applied Daniel's patches to solve problems with the import/export magic
1820        syntax that is required for Windows, and which was going wrong for the
1821        pcreposix and pcrecpp parts of the library. These were overlooked when this
1822        problem was solved for the main library.
1823    
1824     9. There were some crude static tests to avoid integer overflow when computing
1825        the size of patterns that contain repeated groups with explicit upper
1826        limits. As the maximum quantifier is 65535, the maximum group length was
1827        set at 30,000 so that the product of these two numbers did not overflow a
1828        32-bit integer. However, it turns out that people want to use groups that
1829        are longer than 30,000 bytes (though not repeat them that many times).
1830        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1831        made it possible to implement the integer overflow checks in a much more
1832        dynamic way, which I have now done. The artificial limitation on group
1833        length has been removed - we now have only the limit on the total length of
1834        the compiled pattern, which depends on the LINK_SIZE setting.
1835    
1836    10. Fixed a bug in the documentation for get/copy named substring when
1837        duplicate names are permitted. If none of the named substrings are set, the
1838        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1839        empty string.
1840    
1841    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1842        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1843        because the ] is interpreted as the first data character and the
1844        terminating ] is not found. PCRE has been made compatible with Perl in this
1845        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1846        cause memory overwriting.
1847    
1848    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1849        string has been matched (to stop an infinite loop). It was not recognizing
1850        a conditional subpattern that could match an empty string if that
1851        subpattern was within another subpattern. For example, it looped when
1852        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1853        condition was not nested. This bug has been fixed.
1854    
1855    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1856        past the start of the subject in the presence of bytes with the top bit
1857        set, for example "\x8aBCD".
1858    
1859    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1860        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1861    
1862    14. Optimized (?!) to (*FAIL).
1863    
1864    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1865        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1866        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1867        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1868        does: it's just the validity check that is more restrictive.
1869    
1870    16. Inserted checks for integer overflows during escape sequence (backslash)
1871        processing, and also fixed erroneous offset values for syntax errors during
1872        backslash processing.
1873    
1874    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1875        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1876    
1877    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1878        caused an overrun.
1879    
1880    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1881        something other than just ASCII characters) inside a group that had an
1882        unlimited repeat caused a loop at compile time (while checking to see
1883        whether the group could match an empty string).
1884    
1885    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1886        [\P{Any}] did so. (Error in the code for printing property names.)
1887    
1888    21. An orphan \E inside a character class could cause a crash.
1889    
1890    22. A repeated capturing bracket such as (A)? could cause a wild memory
1891        reference during compilation.
1892    
1893    23. There are several functions in pcre_compile() that scan along a compiled
1894        expression for various reasons (e.g. to see if it's fixed length for look
1895        behind). There were bugs in these functions when a repeated \p or \P was
1896        present in the pattern. These operators have additional parameters compared
1897        with \d, etc, and these were not being taken into account when moving along
1898        the compiled data. Specifically:
1899    
1900        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1901            length.
1902    
1903        (b) An item such as \pL+ within a repeated group could cause crashes or
1904            loops.
1905    
1906        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1907            "reference to non-existent subpattern" error.
1908    
1909        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1910    
1911    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1912        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1913    
1914    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1915    
1916    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1917        character were causing crashes (broken optimization).
1918    
1919    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1920        \p or \P) caused a compile-time loop.
1921    
1922    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1923        break. For example, the unstudied pattern [\r\n]A does not match the string
1924        "\r\nA" because change 7.0/46 below moves the current point on by two
1925        characters after failing to match at the start. However, the pattern \nA
1926        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1927        the same is true. There doesn't seem any very clean way out of this, but
1928        what I have chosen to do makes the common cases work: PCRE now takes note
1929        of whether there can be an explicit match for \r or \n anywhere in the
1930        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1931        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1932        pattern has explicit CR or LF references.
1933    
1934    29. Added (*CR) etc for changing newline setting at start of pattern.
1935    
1936    
1937    Version 7.2 19-Jun-07
1938    ---------------------
1939    
1940     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1941        which is apparently normally available under Windows.
1942    
1943     2. Re-jig the pcregrep tests with different newline settings in an attempt
1944        to make them independent of the local environment's newline setting.
1945    
1946     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1947    
1948     4. Some of the "internals" tests were previously cut out when the link size
1949        was not 2, because the output contained actual offsets. The recent new
1950        "Z" feature of pcretest means that these can be cut out, making the tests
1951        usable with all link sizes.
1952    
1953     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1954        stack recursion. This gives a massive performance boost under BSD, but just
1955        a small improvement under Linux. However, it saves one field in the frame
1956        in all cases.
1957    
1958     6. Added more features from the forthcoming Perl 5.10:
1959    
1960        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1961            recursion call. It refers to the nth most recently opened parentheses.
1962    
1963        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1964            to be opened parentheses.
1965    
1966        (c) Conditions that refer to capturing parentheses can be specified
1967            relatively, for example, (?(-2)... or (?(+3)...
1968    
1969        (d) \K resets the start of the current match so that everything before
1970            is not part of it.
1971    
1972        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1973    
1974        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1975            reference syntax.
1976    
1977        (g) (?| introduces a group in which the numbering of parentheses in each
1978            alternative starts with the same number.
1979    
1980        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1981    
1982     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1983        PCRE_INFO_JCHANGED.
1984    
1985     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1986        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1987        for detecting groups that can match an empty string.
1988    
1989     9. A pattern with a very large number of alternatives (more than several
1990        hundred) was running out of internal workspace during the pre-compile
1991        phase, where pcre_compile() figures out how much memory will be needed. A
1992        bit of new cunning has reduced the workspace needed for groups with
1993        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1994        workspace instead of running out of the 4096 that are available.
1995    
1996    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1997    
1998    11. Applied patch from Google to remove an optimization that didn't quite work.
1999        The report of the bug said:
2000    
2001          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2002          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2003          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2004    
2005    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2006        it matched the wrong number of bytes.
2007    
2008    
2009    Version 7.1 24-Apr-07
2010    ---------------------
2011    
2012     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2013        that is more "standard", making use of automake and other Autotools. There
2014        is some re-arrangement of the files and adjustment of comments consequent
2015        on this.
2016    
2017     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2018        for recursive directory scanning broke on some systems because the files
2019        are not scanned in any specific order and on different systems the order
2020        was different. A call to "sort" has been inserted into RunGrepTest for the
2021        approprate test as a short-term fix. In the longer term there may be an
2022        alternative.
2023    
2024     3. I had an email from Eric Raymond about problems translating some of PCRE's
2025        man pages to HTML (despite the fact that I distribute HTML pages, some
2026        people do their own conversions for various reasons). The problems
2027        concerned the use of low-level troff macros .br and .in. I have therefore
2028        removed all such uses from the man pages (some were redundant, some could
2029        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2030        HTML has been updated to handle .nf/.fi and to complain if it encounters
2031        .br or .in.
2032    
2033     4. Updated comments in configure.ac that get placed in config.h.in and also
2034        arranged for config.h to be included in the distribution, with the name
2035        config.h.generic, for the benefit of those who have to compile without
2036        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2037    
2038     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2039        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2040        makevp.bat for the current PCRE, using the additional files
2041        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2042    
2043     6. A Windows user reported a minor discrepancy with test 2, which turned out
2044        to be caused by a trailing space on an input line that had got lost in his
2045        copy. The trailing space was an accident, so I've just removed it.
2046    
2047     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2048        that is needed.
2049    
2050     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2051        as "const" (a) because they are and (b) because it helps the PHP
2052        maintainers who have recently made a script to detect big data structures
2053        in the php code that should be moved to the .rodata section. I remembered
2054        to update Builducptable as well, so it won't revert if ucptable.h is ever
2055        re-created.
2056    
2057     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2058        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2059        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2060        support is not required. This saves 1.5-2K of code, which is important in
2061        some applications.
2062    
2063        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2064        so as not to refer to the tables, even though these functions will never be
2065        called when UTF-8 support is disabled. Otherwise there are problems with a
2066        shared library.
2067    
2068    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2069    
2070        (a) It was defining its arguments as char * instead of void *.
2071    
2072        (b) It was assuming that all moves were upwards in memory; this was true
2073            a long time ago when I wrote it, but is no longer the case.
2074    
2075        The emulated memove() is provided for those environments that have neither
2076        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2077        is clearly not the case, as these two bugs were recently reported.
2078    
2079    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2080        and Detrail to create the HTML documentation, the .txt form of the man
2081        pages, and it removes trailing spaces from listed files. It also creates
2082        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2083        case, it wraps all the #defines with #ifndefs. This script should be run
2084        before "make dist".
2085    
2086    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2087        with Unicode property support.
2088    
2089        (a) For a maximizing quantifier, if the two different cases of the
2090            character were of different lengths in their UTF-8 codings (there are
2091            some cases like this - I found 11), and the matching function had to
2092            back up over a mixture of the two cases, it incorrectly assumed they
2093            were both the same length.
2094    
2095        (b) When PCRE was configured to use the heap rather than the stack for
2096            recursion during matching, it was not correctly preserving the data for
2097            the other case of a UTF-8 character when checking ahead for a match
2098            while processing a minimizing repeat. If the check also involved
2099            matching a wide character, but failed, corruption could cause an
2100            erroneous result when trying to check for a repeat of the original
2101            character.
2102    
2103    13. Some tidying changes to the testing mechanism:
2104    
2105        (a) The RunTest script now detects the internal link size and whether there
2106            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2107            values substituted by "configure". (The RunGrepTest script already did
2108            this for UTF-8.) The configure.ac script no longer substitutes the
2109            relevant variables.
2110    
2111        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2112            with length and offset values. This means that the output is different
2113            for different internal link sizes. Test 2 is skipped for link sizes
2114            other than 2 because of this, bypassing the problem. Unfortunately,
2115            there was also a test in test 3 (the locale tests) that used /B and
2116            failed for link sizes other than 2. Rather than cut the whole test out,
2117            I have added a new /Z option to pcretest that replaces the length and
2118            offset values with spaces. This is now used to make test 3 independent
2119            of link size. (Test 2 will be tidied up later.)
2120    
2121    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2122        segmentation fault instead of returning the appropriate error message.
2123    
2124    15. In multiline mode when the newline sequence was set to "any", the pattern
2125        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2126        This doesn't seem right; it now treats the CRLF combination as the line
2127        ending, and so does not match in that case. It's only a pattern such as ^$
2128        that would hit this one: something like ^ABC$ would have failed after \r
2129        and then tried again after \r\n.
2130    
2131    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2132        in an attempt to make files that differ only in their line terminators
2133        compare equal. This works on Linux.
2134    
2135    17. Under certain error circumstances pcregrep might try to free random memory
2136        as it exited. This is now fixed, thanks to valgrind.
2137    
2138    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2139        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2140        was because its rules for how to advance for /g after matching an empty
2141        string at the end of a line did not allow for this case. They now check for
2142        it specially.
2143    
2144    20. pcretest is supposed to handle patterns and data of any length, by
2145        extending its buffers when necessary. It was getting this wrong when the
2146        buffer for a data line had to be extended.
2147    
2148    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2149        CRLF as a newline sequence.
2150    
2151    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2152        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2153        I have nevertheless tidied it up.
2154    
2155    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2156    
2157    24. Added a man page for pcre-config.
2158    
2159    
2160    Version 7.0 19-Dec-06
2161    ---------------------
2162    
2163     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2164        moving to gcc 4.1.1.
2165    
2166     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2167        sys/time.h, which is documented as needed for this function. It doesn't
2168        seem to matter on Linux, but it showed up on some releases of OS X.
2169    
2170     3. It seems that there are systems where bytes whose values are greater than
2171        127 match isprint() in the "C" locale. The "C" locale should be the
2172        default when a C program starts up. In most systems, only ASCII printing
2173        characters match isprint(). This difference caused the output from pcretest
2174        to vary, making some of the tests fail. I have changed pcretest so that:
2175    
2176        (a) When it is outputting text in the compiled version of a pattern, bytes
2177            other than 32-126 are always shown as hex escapes.
2178    
2179        (b) When it is outputting text that is a matched part of a subject string,
2180            it does the same, unless a different locale has been set for the match
2181            (using the /L modifier). In this case, it uses isprint() to decide.
2182    
2183     4. Fixed a major bug that caused incorrect computation of the amount of memory
2184        required for a compiled pattern when options that changed within the
2185        pattern affected the logic of the preliminary scan that determines the
2186        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2187        that the computed length was too small. The symptoms of this bug were
2188        either the PCRE error "internal error: code overflow" from pcre_compile(),
2189        or a glibc crash with a message such as "pcretest: free(): invalid next
2190        size (fast)". Examples of patterns that provoked this bug (shown in
2191        pcretest format) are:
2192    
2193          /(?-x: )/x
2194          /(?x)(?-x: \s*#\s*)/
2195          /((?i)[\x{c0}])/8
2196          /(?i:[\x{c0}])/8
2197    
2198        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2199        is now done differently.
2200    
2201     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2202        wrapper classes; (b) implement a new function in the C++ scanner that is
2203        more efficient than the old way of doing things because it avoids levels of
2204        recursion in the regex matching; (c) add a paragraph to the documentation
2205        for the FullMatch() function.
2206    
2207     6. The escape sequence \n was being treated as whatever was defined as
2208        "newline". Not only was this contrary to the documentation, which states
2209        that \n is character 10 (hex 0A), but it also went horribly wrong when
2210        "newline" was defined as CRLF. This has been fixed.
2211    
2212     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2213        was being set to -1 for the "end of line" case (supposedly a value that no
2214        character can have). Though this value is never used (the check for end of
2215        line is "zero bytes in current character"), it caused compiler complaints.
2216        I've changed it to 0xffffffff.
2217    
2218     8. In pcre_version.c, the version string was being built by a sequence of
2219        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2220        string (as it is for production releases) called a macro with an empty
2221        argument. The C standard says the result of this is undefined. The gcc
2222        compiler treats it as an empty string (which was what was wanted) but it is
2223        reported that Visual C gives an error. The source has been hacked around to
2224        avoid this problem.
2225    
2226     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2227        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2228        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2229        of them did).
2230    
2231    10. Originally, pcretest opened its input and output without "b"; then I was
2232        told that "b" was needed in some environments, so it was added for release
2233        5.0 to both the input and output. (It makes no difference on Unix-like
2234        systems.) Later I was told that it is wrong for the input on Windows. I've
2235        now abstracted the modes into two macros, to make it easier to fiddle with
2236        them, and removed "b" from the input mode under Windows.
2237    
2238    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2239    
2240    12. Added -help and --help to pcretest as an official way of being reminded
2241        of the options.
2242    
2243    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2244        and pcrecpp.cc because they annoy compilers at high warning levels.
2245    
2246    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2247    
2248    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2249        scripts are not C programs :-) and which was not noticed because it works
2250        on Linux.
2251    
2252    16. pcretest is supposed to handle any length of pattern and data line (as one
2253        line or as a continued sequence of lines) by extending its input buffer if
2254        necessary. This feature was broken for very long pattern lines, leading to
2255        a string of junk being passed to pcre_compile() if the pattern was longer
2256        than about 50K.
2257    
2258    17. I have done a major re-factoring of the way pcre_compile() computes the
2259        amount of memory needed for a compiled pattern. Previously, there was code
2260        that made a preliminary scan of the pattern in order to do this. That was
2261        OK when PCRE was new, but as the facilities have expanded, it has become
2262        harder and harder to keep it in step with the real compile phase, and there
2263        have been a number of bugs (see for example, 4 above). I have now found a
2264        cunning way of running the real compile function in a "fake" mode that
2265        enables it to compute how much memory it would need, while actually only
2266        ever using a few hundred bytes of working memory and without too many
2267        tests of the mode. This should make future maintenance and development
2268        easier. A side effect of this work is that the limit of 200 on the nesting
2269        depth of parentheses has been removed (though this was never a serious
2270        limitation, I suspect). However, there is a downside: pcre_compile() now
2271        runs more slowly than before (30% or more, depending on the pattern). I
2272        hope this isn't a big issue. There is no effect on runtime performance.
2273    
2274    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2275        newline (only possible for the last line of a file) and it was a
2276        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2277    
2278    19. Added additional timing features to pcretest. (1) The -tm option now times
2279        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2280        separate command line item, by a number that specifies the number of
2281        repeats to use when timing. The default is 50000; this gives better
2282        precision, but takes uncomfortably long for very large patterns.
2283    
2284    20. Extended pcre_study() to be more clever in cases where a branch of a
2285        subpattern has no definite first character. For example, (a*|b*)[cd] would
2286        previously give no result from pcre_study(). Now it recognizes that the
2287        first character must be a, b, c, or d.
2288    
2289    21. There was an incorrect error "recursive call could loop indefinitely" if
2290        a subpattern (or the entire pattern) that was being tested for matching an
2291        empty string contained only one non-empty item after a nested subpattern.
2292        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2293        incorrectly, because the \d was being skipped in the check.
2294    
2295    22. The pcretest program now has a new pattern option /B and a command line
2296        option -b, which is equivalent to adding /B to every pattern. This causes
2297        it to show the compiled bytecode, without the additional information that
2298        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2299        is the same as /B/I).
2300    
2301    23. A new optimization is now able automatically to treat some sequences such
2302        as a*b as a*+b. More specifically, if something simple (such as a character
2303        or a simple class like \d) has an unlimited quantifier, and is followed by
2304        something that cannot possibly match the quantified thing, the quantifier
2305        is automatically "possessified".
2306    
2307    24. A recursive reference to a subpattern whose number was greater than 39
2308        went wrong under certain circumstances in UTF-8 mode. This bug could also
2309        have affected the operation of pcre_study().
2310    
2311    25. Realized that a little bit of performance could be had by replacing
2312        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2313    
2314    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2315    
2316    27. Possessive quantifiers such as a++ were previously implemented by turning
2317        them into atomic groups such as ($>a+). Now they have their own opcodes,
2318        which improves performance. This includes the automatically created ones
2319        from 23 above.
2320    
2321    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2322        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2323        the first matched character to be a colon. This applied both to named and
2324        numbered groups.
2325    
2326    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2327    
2328    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2329        building PCRE on Windows easier, so I have included it in the distribution.
2330    
2331    31. There is now a check in pcretest against a ridiculously large number being
2332        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2333        loop, the loop is abandoned.
2334    
2335    32. Forward references to subpatterns in conditions such as (?(2)...) where
2336        subpattern 2 is defined later cause pcre_compile() to search forwards in
2337        the pattern for the relevant set of parentheses. This search went wrong
2338        when there were unescaped parentheses in a character class, parentheses
2339        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2340    
2341    33. "Subroutine" calls and backreferences were previously restricted to
2342        referencing subpatterns earlier in the regex. This restriction has now
2343        been removed.
2344    
2345    34. Added a number of extra features that are going to be in Perl 5.10. On the
2346        whole, these are just syntactic alternatives for features that PCRE had
2347        previously implemented using the Python syntax or my own invention. The
2348        other formats are all retained for compatibility.
2349    
2350        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2351            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2352            also .NET compatible.
2353    
2354        (b) A recursion or subroutine call to a named group can now be defined as
2355            (?&name) as well as (?P>name).
2356    
2357        (c) A backreference to a named group can now be defined as \k<name> or
2358            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2359            5.10, are also .NET compatible.
2360    
2361        (d) A conditional reference to a named group can now use the syntax
2362            (?(<name>) or (?('name') as well as (?(name).
2363    
2364        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2365            groups (named and numbered) that are never evaluated inline, but can be
2366            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2367            is always false. There may be only one alternative in such a group.
2368    
2369        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2370            as the simple (?(R). The condition is true only if the most recent
2371            recursion is that of the given number or name. It does not search out
2372            through the entire recursion stack.
2373    
2374        (g) The escape \gN or \g{N} has been added, where N is a positive or
2375            negative number, specifying an absolute or relative reference.
2376    
2377    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2378        some "unreachable code" warnings.
2379    
2380    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2381        things, this adds five new scripts.
2382    
2383    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2384        There were also incompatibilities regarding the handling of \Q..\E inside
2385        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2386        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2387    
2388    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2389        matches an empty string, and forcibly breaks the loop. There were bugs in
2390        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2391        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2392        separate and independent bugs (that affected different cases) have been
2393        fixed.
2394    
2395    39. Refactored the code to abolish the use of different opcodes for small
2396        capturing bracket numbers. This is a tidy that I avoided doing when I
2397        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2398        The new approach is not only tidier, it makes it possible to reduce the
2399        memory needed to fix the previous bug (38).
2400    
2401    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2402        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2403        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2404        mode.
2405    
2406    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2407        report.
2408    
2409    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2410        copy construction and assignment in the C++ wrapper.
2411    
2412    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2413        couple of bugs that could have given wrong results in the "--newline=crlf"
2414        case.
2415    
2416    44. Added a number of casts and did some reorganization of signed/unsigned int
2417        variables following suggestions from Dair Grant. Also renamed the variable
2418        "this" as "item" because it is a C++ keyword.
2419    
2420    45. Arranged for dftables to add
2421    
2422          #include "pcre_internal.h"
2423    
2424        to pcre_chartables.c because without it, gcc 4.x may remove the array
2425        definition from the final binary if PCRE is built into a static library and
2426        dead code stripping is activated.
2427    
2428    46. For an unanchored pattern, if a match attempt fails at the start of a
2429        newline sequence, and the newline setting is CRLF or ANY, and the next two
2430        characters are CRLF, advance by two characters instead of one.
2431    
2432    
2433    Version 6.7 04-Jul-06
2434    ---------------------
2435    
2436     1. In order to handle tests when input lines are enormously long, pcretest has
2437        been re-factored so that it automatically extends its buffers when
2438        necessary. The code is crude, but this _is_ just a test program. The
2439        default size has been increased from 32K to 50K.
2440    
2441     2. The code in pcre_study() was using the value of the re argument before
2442        testing it for NULL. (Of course, in any sensible call of the function, it
2443        won't be NULL.)
2444    
2445     3. The memmove() emulation function in pcre_internal.h, which is used on
2446        systems that lack both memmove() and bcopy() - that is, hardly ever -
2447        was missing a "static" storage class specifier.
2448    
2449     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2450        containing an extended class (one that cannot be represented by a bitmap
2451        because it contains high-valued characters or Unicode property items, e.g.
2452        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2453        pattern, but PCRE should not loop if you do not (it no longer does).
2454        [Detail: two cases were found: (a) a repeated subpattern containing an
2455        extended class; (b) a recursive reference to a subpattern that followed a
2456        previous extended class. It wasn't skipping over the extended class
2457        correctly when UTF-8 mode was not set.]
2458    
2459     5. A negated single-character class was not being recognized as fixed-length
2460        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2461        compile error "lookbehind assertion is not fixed length".
2462    
2463     6. The RunPerlTest auxiliary script was showing an unexpected difference
2464        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2465        write a Perl script that can interpret lines of an input file either as
2466        byte characters or as UTF-8, which is what "perltest" was being required to
2467        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2468        can't do is switch easily at run time between having the "use utf8;" pragma
2469        or not. In the end, I fudged it by using the RunPerlTest script to insert
2470        "use utf8;" explicitly for the UTF-8 tests.
2471    
2472     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2473        the end of the subject string, contrary to the documentation and to what
2474        Perl does. This was true of both matching functions. Now it matches only at
2475        the start of the subject and immediately after *internal* newlines.
2476    
2477     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2478        a pointer to an int instead of a pointer to an unsigned long int. This
2479        caused problems on 64-bit systems.
2480    
2481     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2482        instance of the 'standard' template library not being so standard".
2483    
2484    10. There was no check on the number of named subpatterns nor the maximum
2485        length of a subpattern name. The product of these values is used to compute
2486        the size of the memory block for a compiled pattern. By supplying a very
2487        long subpattern name and a large number of named subpatterns, the size
2488        computation could be caused to overflow. This is now prevented by limiting
2489        the length of names to 32 characters, and the number of named subpatterns
2490        to 10,000.
2491    
2492    11. Subpatterns that are repeated with specific counts have to be replicated in
2493        the compiled pattern. The size of memory for this was computed from the
2494        length of the subpattern and the repeat count. The latter is limited to
2495        65535, but there was no limit on the former, meaning that integer overflow
2496        could in principle occur. The compiled length of a repeated subpattern is
2497        now limited to 30,000 bytes in order to prevent this.
2498    
2499    12. Added the optional facility to have named substrings with the same name.
2500    
2501    13. Added the ability to use a named substring as a condition, using the
2502        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2503        are numbers (not recommended). Forward references are permitted.
2504    
2505    14. Added forward references in named backreferences (if you see what I mean).
2506    
2507    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2508        pattern could run off the end of the subject. For example, the pattern
2509        "(?s)(.{1,5})"8 did this with the subject "ab".
2510    
2511    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2512        PCRE_CASELESS was set when matching characters that were quantified with ?
2513        or *.
2514    
2515    17. A character class other than a single negated character that had a minimum
2516        but no maximum quantifier - for example [ab]{6,} - was not handled
2517        correctly by pce_dfa_exec(). It would match only one character.
2518    
2519    18. A valid (though odd) pattern that looked like a POSIX character
2520        class but used an invalid character after [ (for example [[,abc,]]) caused
2521        pcre_compile() to give the error "Failed: internal error: code overflow" or
2522        in some cases to crash with a glibc free() error. This could even happen if
2523        the pattern terminated after [[ but there just happened to be a sequence of
2524        letters, a binary zero, and a closing ] in the memory that followed.
2525    
2526    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2527        over the years. Originally (before any Unicode support), just the bottom 8
2528        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2529        output from "man perlunicode" includes this:
2530    
2531          The regular expression compiler produces polymorphic opcodes.  That
2532          is, the pattern adapts to the data and automatically switches to
2533          the Unicode character scheme when presented with Unicode data--or
2534          instead uses a traditional byte scheme when presented with byte
2535          data.
2536    
2537        Sadly, a wide octal escape does not cause a switch, and in a string with
2538        no other multibyte characters, these octal escapes are treated as before.
2539        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2540        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2541        Unicode string.
2542    
2543        I have not perpetrated such confusion in PCRE. Up till now, it took just
2544        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2545        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2546        translate to the appropriate multibyte character.
2547    
2548    29. Applied some refactoring to reduce the number of warnings from Microsoft
2549        and Borland compilers. This has included removing the fudge introduced
2550        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2551        a warning about an unused variable.
2552    
2553    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2554        characters since release 4.0, because Perl (from release 5.004) does not.
2555        [Or at least, is documented not to: some releases seem to be in conflict
2556        with the documentation.] However, when a pattern was studied with
2557        pcre_study() and all its branches started with \s, PCRE still included VT
2558        as a possible starting character. Of course, this did no harm; it just
2559        caused an unnecessary match attempt.
2560    
2561    22. Removed a now-redundant internal flag bit that recorded the fact that case
2562        dependency changed within the pattern. This was once needed for "required
2563        byte" processing, but is no longer used. This recovers a now-scarce options
2564        bit. Also moved the least significant internal flag bit to the most-
2565        significant bit of the word, which was not previously used (hangover from
2566        the days when it was an int rather than a uint) to free up another bit for
2567        the future.
2568    
2569    23. Added support for CRLF line endings as well as CR and LF. As well as the
2570        default being selectable at build time, it can now be changed at runtime
2571        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2572        specify that it is scanning data with non-default line endings.
2573    
2574    24. Changed the definition of CXXLINK to make it agree with the definition of
2575        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2576    
2577    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2578        recursions. This makes a big different to stack usage for some patterns.
2579    
2580    26. If a subpattern containing a named recursion or subroutine reference such
2581        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2582        the space required for the compiled pattern went wrong and gave too small a
2583        value. Depending on the environment, this could lead to "Failed: internal
2584        error: code overflow at offset 49" or "glibc detected double free or
2585        corruption" errors.
2586    
2587    27. Applied patches from Google (a) to support the new newline modes and (b) to
2588        advance over multibyte UTF-8 characters in GlobalReplace.
2589    
2590    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2591        difference for some implementation of PCRE in some Windows version.
2592    
2593    29. Added some extra testing facilities to pcretest:
2594    
2595        \q<number>   in a data line sets the "match limit" value
2596        \Q<number>   in a data line sets the "match recursion limt" value
2597        -S <number>  sets the stack size, where <number> is in megabytes
2598    
2599        The -S option isn't available for Windows.
2600    
2601    
2602    Version 6.6 06-Feb-06
2603    ---------------------
2604    
2605     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2606        in pcreposix.h. I have copied the definition from pcre.h.
2607    
2608     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2609        because pcre.h is no longer a built file.
2610    
2611     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2612        not normally included in the compiled code.
2613    
2614    
2615    Version 6.5 01-Feb-06
2616    ---------------------
2617    
2618     1. When using the partial match feature with pcre_dfa_exec(), it was not
2619        anchoring the second and subsequent partial matches at the new starting
2620        point. This could lead to incorrect results. For example, with the pattern
2621        /1234/, partially matching against "123" and then "a4" gave a match.
2622    
2623     2. Changes to pcregrep:
2624    
2625        (a) All non-match returns from pcre_exec() were being treated as failures
2626            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2627            error message is output. Some extra information is given for the
2628            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2629            probably the only errors that are likely to be caused by users (by
2630            specifying a regex that has nested indefinite repeats, for instance).
2631            If there are more than 20 of these errors, pcregrep is abandoned.
2632    
2633        (b) A binary zero was treated as data while matching, but terminated the
2634            output line if it was written out. This has been fixed: binary zeroes
2635            are now no different to any other data bytes.
2636    
2637        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2638            used to set a locale for matching. The --locale=xxxx long option has
2639            been added (no short equivalent) to specify a locale explicitly on the
2640            pcregrep command, overriding the environment variables.
2641    
2642        (d) When -B was used with -n, some line numbers in the output were one less
2643            than they should have been.
2644    
2645        (e) Added the -o (--only-matching) option.
2646    
2647        (f) If -A or -C was used with -c (count only), some lines of context were
2648            accidentally printed for the final match.
2649    
2650        (g) Added the -H (--with-filename) option.
2651    
2652        (h) The combination of options -rh failed to suppress file names for files
2653            that were found from directory arguments.
2654    
2655        (i) Added the -D (--devices) and -d (--directories) options.
2656    
2657        (j) Added the -F (--fixed-strings) option.
2658    
2659        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2660    
2661        (l) Added the --colo(u)r option.
2662    
2663        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2664            is not present by default.
2665    
2666     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2667        items such as (?R) or (?1), when the recursion could match a number of
2668        alternatives. If it matched one of the alternatives, but subsequently,
2669        outside the recursion, there was a failure, the code tried to back up into
2670        the recursion. However, because of the way PCRE is implemented, this is not
2671        possible, and the result was an incorrect result from the match.
2672    
2673        In order to prevent this happening, the specification of recursion has
2674        been changed so that all such subpatterns are automatically treated as
2675        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2676    
2677     4. I had overlooked the fact that, in some locales, there are characters for
2678        which isalpha() is true but neither isupper() nor islower() are true. In
2679        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2680        and ordfeminine) are like this. This affected the treatment of \w and \W
2681        when they appeared in character classes, but not when they appeared outside
2682        a character class. The bit map for "word" characters is now created
2683        separately from the results of isalnum() instead of just taking it from the
2684        upper, lower, and digit maps. (Plus the underscore character, of course.)
2685    
2686     5. The above bug also affected the handling of POSIX character classes such as
2687        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2688        permanent tables. Instead, the bit maps for such a class were previously
2689        created as the appropriate unions of the upper, lower, and digit bitmaps.
2690        Now they are created by subtraction from the [[:word:]] class, which has
2691        its own bitmap.
2692    
2693     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2694        It is created by subtracting the vertical space characters (\x09, \x0a,
2695        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2696        subtraction was done in the overall bitmap for a character class, meaning
2697        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2698        be recognized. This bug has been fixed.
2699    
2700     7. Patches from the folks at Google:
2701    
2702          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2703          real life, but is still worth protecting against".
2704    
2705          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2706          regular expressions".
2707    
2708          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2709          have it.
2710    
2711          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2712          "configure" and the latter not, in order to fix a problem somebody had
2713          with compiling the Arg class on HP-UX.
2714    
2715          (e) Improve the error-handling of the C++ wrapper a little bit.
2716    
2717          (f) New tests for checking recursion limiting.
2718    
2719     8. The pcre_memmove() function, which is used only if the environment does not
2720        have a standard memmove() function (and is therefore rarely compiled),
2721        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2722        returning a result (though PCRE never actually uses the result).
2723    
2724     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2725        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2726        returned instead of calling malloc() with an overflowing number that would
2727        most likely cause subsequent chaos.
2728    
2729    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2730    
2731    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2732        with this option is matched, the nmatch and pmatch options of regexec() are
2733        ignored.
2734    
2735    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2736        provided in case anyone wants to the the POSIX interface with UTF-8
2737        strings.
2738    
2739    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2740        C++ linking (needed for some HP-UX environments).
2741    
2742    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2743        (unused parameter) and in the pcre_printint() function (omitted "default"
2744        switch label when the default is to do nothing).
2745    
2746    15. Added some code to make it possible, when PCRE is compiled as a C++
2747        library, to replace subject pointers for pcre_exec() with a smart pointer
2748        class, thus making it possible to process discontinuous strings.
2749    
2750    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2751        much the same function. They were added by different people who were trying
2752        to make PCRE easy to compile on non-Unix systems. It has been suggested
2753        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2754        for compiling on Windows systems. I have therefore replaced it with
2755        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2756        defaults to "extern" for C or "extern C" for C++, which works fine on
2757        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2758        SCOPE with something explicit in config.h. In addition:
2759    
2760        (a) pcreposix.h still had just "extern" instead of either of these macros;
2761            I have replaced it with PCRE_DATA_SCOPE.
2762    
2763        (b) Functions such as _pcre_xclass(), which are internal to the library,
2764            but external in the C sense, all had PCRE_EXPORT in their definitions.
2765            This is apparently wrong for the Windows case, so I have removed it.
2766            (It makes no difference on Unix-like systems.)
2767    
2768    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2769        of recursive calls to match(). This is different to MATCH_LIMIT because
2770        that limits the total number of calls to match(), not all of which increase
2771        the depth of recursion. Limiting the recursion depth limits the amount of
2772        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2773        when PCRE is compiled, and changed at run time. A patch from Google adds
2774        this functionality to the C++ interface.
2775    
2776    18. Changes to the handling of Unicode character properties:
2777    
2778        (a) Updated the table to Unicode 4.1.0.
2779    
2780        (b) Recognize characters that are not in the table as "Cn" (undefined).
2781    
2782        (c) I revised the way the table is implemented to a much improved format
2783            which includes recognition of ranges. It now supports the ranges that
2784            are defined in UnicodeData.txt, and it also amalgamates other
2785            characters into ranges. This has reduced the number of entries in the
2786            table from around 16,000 to around 3,000, thus reducing its size
2787            considerably. I realized I did not need to use a tree structure after
2788            all - a binary chop search is just as efficient. Having reduced the
2789            number of entries, I extended their size from 6 bytes to 8 bytes to
2790            allow for more data.
2791    
2792        (d) Added support for Unicode script names via properties such as \p{Han}.
2793    
2794    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2795        matching that character.
2796    
2797    20. When matching a repeated Unicode property with a minimum greater than zero,
2798        (for example \pL{2,}), PCRE could look past the end of the subject if it
2799        reached it while seeking the minimum number of characters. This could
2800        happen only if some of the characters were more than one byte long, because
2801        there is a check for at least the minimum number of bytes.
2802    
2803    21. Refactored the implementation of \p and \P so as to be more general, to
2804        allow for more different types of property in future. This has changed the
2805        compiled form incompatibly. Anybody with saved compiled patterns that use
2806        \p or \P will have to recompile them.
2807    
2808    22. Added "Any" and "L&" to the supported property types.
2809    
2810    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2811        but give a compile time error if the value is greater than 0xff.
2812    
2813    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2814        accidentally not being installed or uninstalled.
2815    
2816    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2817        made were to insert the current release number. This seemed silly, because
2818        it made things harder for people building PCRE on systems that don't run
2819        "configure". I have turned pcre.h into a distributed file, no longer built
2820        by "configure", with the version identification directly included. There is
2821        no longer a pcre.h.in file.
2822    
2823        However, this change necessitated a change to the pcre-config script as
2824        well. It is built from pcre-config.in, and one of the substitutions was the
2825        release number. I have updated configure.ac so that ./configure now finds
2826        the release number by grepping pcre.h.
2827    
2828    26. Added the ability to run the tests under valgrind.
2829    
2830    
2831    Version 6.4 05-Sep-05
2832    ---------------------
2833    
2834     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2835        "--" to be printed when multiple files were scanned, even when none of the
2836        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2837        consider it to be a bug, and have restored the previous behaviour.
2838    
2839     2. A couple of code tidies to get rid of compiler warnings.
2840    
2841     3. The pcretest program used to cheat by referring to symbols in the library
2842        whose names begin with _pcre_. These are internal symbols that are not
2843        really supposed to be visible externally, and in some environments it is
2844        possible to suppress them. The cheating is now confined to including
2845        certain files from the library's source, which is a bit cleaner.
2846    
2847     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2848        file's purpose clearer.
2849    
2850     5. Reorganized pcre_ucp_findchar().
2851    
2852    
2853    Version 6.3 15-Aug-05
2854    ---------------------
2855    
2856     1. The file libpcre.pc.in did not have general read permission in the tarball.
2857    
2858     2. There were some problems when building without C++ support:
2859    
2860        (a) If C++ support was not built, "make install" and "make test" still
2861            tried to test it.
2862    
2863        (b) There were problems when the value of CXX was explicitly set. Some
2864            changes have been made to try to fix these, and ...
2865    
2866        (c) --disable-cpp can now be used to explicitly disable C++ support.
2867    
2868        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2869            backslash in a target when C++ was disabled. This confuses some
2870            versions of "make", apparently. Using an intermediate variable solves
2871            this. (Same for CPP_LOBJ.)
2872    
2873     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2874        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2875        necessary on certain architectures.
2876    
2877     4. Added a setting of -export-symbols-regex to the link command to remove
2878        those symbols that are exported in the C sense, but actually are local
2879        within the library, and not documented. Their names all begin with
2880        "_pcre_". This is not a perfect job, because (a) we have to except some
2881        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2882        available (and never for static libraries). I have made a note to try to
2883        find a way round (a) in the future.
2884    
2885    
2886    Version 6.2 01-Aug-05
2887    ---------------------
2888    
2889     1. There was no test for integer overflow of quantifier values. A construction
2890        such as {1111111111111111} would give undefined results. What is worse, if
2891        a minimum quantifier for a parenthesized subpattern overflowed and became
2892        negative, the calculation of the memory size went wrong. This could have
2893        led to memory overwriting.
2894    
2895     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2896    
2897     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2898        operating environments where this matters.
2899    
2900     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2901        PCRE options from within the C++ wrapper.
2902    
2903     5. Named capturing subpatterns were not being correctly counted when a pattern
2904        was compiled. This caused two problems: (a) If there were more than 100
2905        such subpatterns, the calculation of the memory needed for the whole
2906        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2907        back references of the form \12, where the number was greater than 9, were
2908        not recognized as back references, even though there were sufficient
2909        previous subpatterns.
2910    
2911     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2912        versions of gcc, e.g. 2.95.4.
2913    
2914    
2915    Version 6.1 21-Jun-05
2916    ---------------------
2917    
2918     1. There was one reference to the variable "posix" in pcretest.c that was not
2919        surrounded by "#if !defined NOPOSIX".
2920    
2921     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2922        the cross-check on the old pcre_info() function, for the benefit of the
2923        cut-down version of PCRE that is currently imported into Exim.
2924    
2925     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2926        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2927        patterns that start (?i) but I don't think that matters. The use of (?i) is
2928        just an example; this all applies to the other options as well.
2929    
2930     4. Since libtool seems to echo the compile commands it is issuing, the output
2931        from "make" can be reduced a bit by putting "@" in front of each libtool
2932        compile command.
2933    
2934     5. Patch from the folks at Google for configure.in to be a bit more thorough
2935        in checking for a suitable C++ installation before trying to compile the
2936        C++ stuff. This should fix a reported problem when a compiler was present,
2937        but no suitable headers.
2938    
2939     6. The man pages all had just "PCRE" as their title. I have changed them to
2940        be the relevant file name. I have also arranged that these names are
2941        retained in the file doc/pcre.txt, which is a concatenation in text format
2942        of all the man pages except the little individual ones for each function.
2943    
2944     7. The NON-UNIX-USE file had not been updated for the different set of source
2945        files that come with release 6. I also added a few comments about the C++
2946        wrapper.
2947    
2948    
2949    Version 6.0 07-Jun-05
2950    ---------------------
2951    
2952     1. Some minor internal re-organization to help with my DFA experiments.
2953    
2954     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2955        didn't matter for the library itself when fully configured, but did matter
2956        when compiling without UCP support, or within Exim, where the ucp files are
2957        not imported.
2958    
2959     3. Refactoring of the library code to split up the various functions into
2960        different source modules. The addition of the new DFA matching code (see
2961        below) to a single monolithic source would have made it really too
2962        unwieldy, quite apart from causing all the code to be include in a
2963        statically linked application, when only some functions are used. This is
2964        relevant even without the DFA addition now that patterns can be compiled in
2965        one application and matched in another.
2966    
2967        The downside of splitting up is that there have to be some external
2968        functions and data tables that are used internally in different modules of
2969        the library but which are not part of the API. These have all had their
2970        names changed to start with "_pcre_" so that they are unlikely to clash
2971        with other external names.
2972    
2973     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2974        a different (DFA) algorithm. Although it is slower than the original
2975        function, it does have some advantages for certain types of matching
2976        problem.
2977    
2978     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2979        including restarting after a partial match.
2980    
2981     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2982        defined when compiling for Windows was sent to me. I have put it into the
2983        code, though I have no means of testing or verifying it.
2984    
2985     7. Added the pcre_refcount() auxiliary function.
2986    
2987     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2988        match before or at the first newline in the subject string. In pcretest,
2989        the /f option on a pattern can be used to set this.
2990    
2991     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2992        would behave wrongly. This has been present in PCRE since release 4.0.
2993    
2994    10. A number of changes to the pcregrep command:
2995    
2996        (a) Refactored how -x works; insert ^(...)$ instead of setting
2997            PCRE_ANCHORED and checking the length, in preparation for adding
2998            something similar for -w.
2999    
3000        (b) Added the -w (match as a word) option.
3001    
3002        (c) Refactored the way lines are read and buffered so as to have more
3003            than one at a time available.
3004    
3005        (d) Implemented a pcregrep test script.
3006    
3007        (e) Added the -M (multiline match) option. This allows patterns to match
3008            over several lines of the subject. The buffering ensures that at least
3009            8K, or the rest of the document (whichever is the shorter) is available
3010            for matching (and similarly the previous 8K for lookbehind assertions).
3011    
3012        (f) Changed the --help output so that it now says
3013    
3014              -w, --word-regex(p)
3015    
3016            instead of two lines, one with "regex" and the other with "regexp"
3017            because that confused at least one person since the short forms are the
3018            same. (This required a bit of code, as the output is generated
3019            automatically from a table. It wasn't just a text change.)
3020    
3021        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3022            option but starts with a hyphen. Could be a pattern or a path name
3023            starting with a hyphen, for instance.
3024    
3025        (h) "-" can be given as a file name to represent stdin.
3026    
3027        (i) When file names are being printed, "(standard input)" is used for
3028            the standard input, for compatibility with GNU grep. Previously
3029            "<stdin>" was used.
3030    
3031        (j) The option --label=xxx can be used to supply a name to be used for
3032            stdin when file names are being printed. There is no short form.
3033    
3034        (k) Re-factored the options decoding logic because we are going to add
3035            two more options that take data. Such options can now be given in four
3036            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3037    
3038        (l) Added the -A, -B, and -C options for requesting that lines of context
3039            around matches be printed.
3040    
3041        (m) Added the -L option to print the names of files that do not contain
3042            any matching lines, that is, the complement of -l.
3043    
3044        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3045            continue to scan other files.
3046    
3047        (o) The -s option was incorrectly implemented. For compatibility with other
3048            greps, it now suppresses the error message for a non-existent or non-
3049            accessible file (but not the return code). There is a new option called
3050            -q that suppresses the output of matching lines, which was what -s was
3051            previously doing.
3052    
3053        (p) Added --include and --exclude options to specify files for inclusion
3054            and exclusion when recursing.
3055    
3056    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3057        Hopefully, it now does.
3058    
3059    12. Missing cast in pcre_study().
3060    
3061    13. Added an "uninstall" target to the makefile.
3062    
3063    14. Replaced "extern" in the function prototypes in Makefile.in with
3064        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3065        world, but is set differently for Windows.
3066    
3067    15. Added a second compiling function called pcre_compile2(). The only
3068        difference is that it has an extra argument, which is a pointer to an
3069        integer error code. When there is a compile-time failure, this is set
3070        non-zero, in addition to the error test pointer being set to point to an
3071        error message. The new argument may be NULL if no error number is required
3072        (but then you may as well call pcre_compile(), which is now just a
3073        wrapper). This facility is provided because some applications need a
3074        numeric error indication, but it has also enabled me to tidy up the way
3075        compile-time errors are handled in the POSIX wrapper.
3076    
3077    16. Added VPATH=.libs to the makefile; this should help when building with one
3078        prefix path and installing with another. (Or so I'm told by someone who
3079        knows more about this stuff than I do.)
3080    
3081    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3082        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3083        match everything, including newlines. This is not POSIX-compatible, but
3084        somebody wanted the feature. From pcretest it can be activated by using
3085        both the P and the s flags.
3086    
3087    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3088    
3089    19. libpcre.pc was being incorrectly installed as executable.
3090    
3091    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3092        it now also looks for '\r' so that it will work unmodified on Windows.
3093    
3094    21. Added Google's contributed C++ wrapper to the distribution.
3095    
3096    22. Added some untidy missing memory free() calls in pcretest, to keep
3097        Electric Fence happy when testing.
3098    
3099    
3100    
3101    Version 5.0 13-Sep-04
3102    ---------------------
3103    
3104     1. Internal change: literal characters are no longer packed up into items
3105        containing multiple characters in a single byte-string. Each character
3106        is now matched using a separate opcode. However, there may be more than one
3107        byte in the character in UTF-8 mode.
3108    
3109     2. The pcre_callout_block structure has two new fields: pattern_position and
3110        next_item_length. These contain the offset in the pattern to the next match
3111        item, and its length, respectively.
3112    
3113     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3114        insertion of callouts before each pattern item. Added the /C option to
3115        pcretest to make use of this.
3116    
3117     4. On the advice of a Windows user, the lines
3118    
3119          #if defined(_WIN32) || defined(WIN32)
3120          _setmode( _fileno( stdout ), 0x8000 );
3121          #endif  /* defined(_WIN32) || defined(WIN32) */
3122    
3123        have been added to the source of pcretest. This apparently does useful
3124        magic in relation to line terminators.
3125    
3126     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3127        for the benefit of those environments where the "b" makes a difference.
3128    
3129     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3130        to know about it. I have put a hack into configure.in that adds in code
3131        to set GCC=yes if CC=icc. This seems to end up at a point in the
3132        generated configure script that is early enough to affect the setting of
3133        compiler options, which is what is needed, but I have no means of testing
3134        whether it really works. (The user who reported this had patched the
3135        generated configure script, which of course I cannot do.)
3136    
3137        LATER: After change 22 below (new libtool files), the configure script
3138        seems to know about icc (and also ecc). Therefore, I have commented out
3139        this hack in configure.in.
3140    
3141     7. Added support for pkg-config (2 patches were sent in).
3142    
3143     8. Negated POSIX character classes that used a combination of internal tables
3144        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3145        [[:^ascii]]. Typically, they would match almost any characters. The other
3146        POSIX classes were not broken in this way.
3147    
3148     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3149        to find the match, as PCRE was deluded into thinking that the match had to
3150        start at the start point or following a newline. The same bug applied to
3151        patterns with negative forward assertions or any backward assertions
3152        preceding ".*" at the start, unless the pattern required a fixed first
3153        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3154    
3155    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3156        starting at the last subject character, bytes beyond the end of the subject
3157        string were read.
3158    
3159    11. Renamed the variable "class" as "classbits" to make life easier for C++
3160        users. (Previously there was a macro definition, but it apparently wasn't
3161        enough.)
3162    
3163    12. Added the new field "tables" to the extra data so that tables can be passed
3164        in at exec time, or the internal tables can be re-selected. This allows
3165        a compiled regex to be saved and re-used at a later time by a different
3166        program that might have everything at different addresses.
3167    
3168    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3169        -R library as well as a -L library.
3170    
3171    14. The debugging options of pcretest (-d on the command line or D on a
3172        pattern) showed incorrect output for anything following an extended class
3173        that contained multibyte characters and which was followed by a quantifier.
3174    
3175    15. Added optional support for general category Unicode character properties
3176        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3177        support. It adds about 90K to the size of the library. The meanings of the
3178        inbuilt class escapes such as \d and \s have NOT been changed.
3179    
3180    16. Updated pcredemo.c to include calls to free() to release the memory for the
3181        compiled pattern.
3182    
3183    17. The generated file chartables.c was being created in the source directory
3184        instead of in the building directory. This caused the build to fail if the
3185        source directory was different from the building directory, and was
3186        read-only.
3187    
3188    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3189        file. No doubt somebody will tell me if they don't make sense... Also added
3190        Dan Mooney's comments about building on OpenVMS.
3191    
3192    19. Added support for partial matching via the PCRE_PARTIAL option for
3193        pcre_exec() and the \P data escape in pcretest.
3194    
3195    20. Extended pcretest with 3 new pattern features:
3196    
3197        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3198              write the compiled pattern to the file whose name is "rest-of-line".
3199              This is a straight binary dump of the data, with the saved pointer to
3200              the character tables forced to be NULL. The study data, if any, is
3201              written too. After writing, pcretest reads a new pattern.
3202    
3203        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3204              compiled pattern from the given file. There must not be any
3205              occurrences of "<" in the file name (pretty unlikely); if there are,
3206              pcretest will instead treat the initial "<" as a pattern delimiter.
3207              After reading in the pattern, pcretest goes on to read data lines as
3208              usual.
3209    
3210        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3211              and 16-bit fields in a compiled pattern, to simulate a pattern that
3212              was compiled on a host of opposite endianness.
3213    
3214    21. The pcre-exec() function can now cope with patterns that were compiled on
3215        hosts of opposite endianness, with this restriction:
3216    
3217          As for any compiled expression that is saved and used later, the tables
3218          pointer field cannot be preserved; the extra_data field in the arguments
3219          to pcre_exec() should be used to pass in a tables address if a value
3220          other than the default internal tables were used at compile time.
3221    
3222    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3223        now diagnosed as an error. Previously, most of the time, a negative number
3224        would have been treated as zero, but if in addition "ovector" was passed as
3225        NULL, a crash could occur.
3226    
3227    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3228        new versions from the libtool 1.5 distribution (the last one is a copy of
3229        a file called libtool.m4). This seems to have fixed the need to patch
3230        "configure" to support Darwin 1.3 (which I used to do). However, I still
3231        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3232        workstation).
3233    
3234    24. Changed the PCRE licence to be the more standard "BSD" licence.
3235    
3236    
3237    Version 4.5 01-Dec-03
3238    ---------------------
3239    
3240     1. There has been some re-arrangement of the code for the match() function so
3241        that it can be compiled in a version that does not call itself recursively.
3242        Instead, it keeps those local variables that need separate instances for
3243        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3244        needs to "recurse". Keeping track of where control must go is done by means
3245        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3246        hide most of the details from the main code, and operates only if
3247        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3248        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3249        operating.
3250    
3251        To make it easier for callers to provide specially tailored get/free
3252        functions for this usage, two new functions, pcre_stack_malloc, and
3253        pcre_stack_free, are used. They are always called in strict stacking order,
3254        and the size of block requested is always the same.
3255    
3256        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3257        PCRE has been compiled to use the stack or the heap for recursion. The
3258        -C option of pcretest uses this to show which version is compiled.
3259    
3260        A new data escape \S, is added to pcretest; it causes the amounts of store
3261        obtained and freed by both kinds of malloc/free at match time to be added
3262        to the output.
3263    
3264     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3265        what's available on my current Linux desktop machine.
3266    
3267     3. When matching a UTF-8 string, the test for a valid string at the start has
3268        been extended. If start_offset is not zero, PCRE now checks that it points
3269        to a byte that is the start of a UTF-8 character. If not, it returns
3270        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3271        this is necessary because there may be backward assertions in the pattern.
3272        When matching the same subject several times, it may save resources to use
3273        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3274    
3275     4. The code for checking the validity of UTF-8 strings has been tightened so
3276        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3277        containing "overlong sequences".
3278    
3279     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3280        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3281        should have been "&", but it just so happened that all the cases this let
3282        through by mistake were picked up later in the function.
3283    
3284     6. I had used a variable called "isblank" - this is a C99 function, causing
3285        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3286    
3287     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3288        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3289        so the version is shown; (c) stop "make test" ech