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