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

Diff of /code/trunk/ChangeLog

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

revision 45 by nigel, Sat Feb 24 21:39:25 2007 UTC revision 941 by zherczeg, Tue Feb 28 11:33:34 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    9.  (*MARK) control verb is now supported by the JIT compiler.
46    
47    10. The command "./RunTest list" lists the available tests without actually
48        running any of them. (Because I keep forgetting what they all are.)
49    
50    11. Add PCRE_INFO_MAXLOOKBEHIND.
51    
52    12. Applied a (slightly modified) user-supplied patch that improves performance
53        when the heap is used for recursion (compiled with --disable-stack-for-
54        recursion). Instead of malloc and free for each heap frame each time a
55        logical recursion happens, frames are retained on a chain and re-used where
56        possible. This sometimes gives as much as 30% improvement.
57    
58    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59        call.
60    
61    14. As documented, (*COMMIT) is now confined to within a positive assertion.
62    
63    15. It is now possible to link pcretest with libedit as an alternative to
64        libreadline.
65    
66    16. (*COMMIT) control verb is now supported by the JIT compiler.
67    
68    
69    Version 8.30 04-February-2012
70    -----------------------------
71    
72    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
73        name is defined in ctype.h.
74    
75    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
76        only in quite long subpatterns.
77    
78    3.  Removed the function pcre_info(), which has been obsolete and deprecated
79        since it was replaced by pcre_fullinfo() in February 2000.
80    
81    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
82        match a (*MARK), and the match failed at the start of the subject, a
83        reference to memory before the start of the subject could occur. This bug
84        was introduced by fix 17 of release 8.21.
85    
86    5.  A reference to an unset group with zero minimum repetition was giving
87        totally wrong answers (in non-JavaScript-compatibility mode). For example,
88        /(another)?(\1?)test/ matched against "hello world test". This bug was
89        introduced in release 8.13.
90    
91    6.  Add support for 16-bit character strings (a large amount of work involving
92        many changes and refactorings).
93    
94    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
95        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
96        from a file.
97    
98    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
99        rounding is not applied in this particular case).
100    
101    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
102        if they appear, or are escaped, in patterns.
103    
104    10. Get rid of a number of -Wunused-but-set-variable warnings.
105    
106    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
107        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
108        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
109        also returns the mark "x". This bug applied to capturing parentheses,
110        non-capturing parentheses, and atomic parentheses. It also applied to some
111        assertions.
112    
113    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
114        information out of configure.ac instead of relying on pcre.h.generic, which
115        is not stored in the repository.
116    
117    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
118        -lreadline.
119    
120    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
121    
122    15. Applied Graycode's patch to put the top-level frame on the stack rather
123        than the heap when not using the stack for recursion. This gives a
124        performance improvement in many cases when recursion is not deep.
125    
126    16. Experimental code added to "pcretest -C" to output the stack frame size.
127    
128    
129    Version 8.21 12-Dec-2011
130    ------------------------
131    
132    1.  Updating the JIT compiler.
133    
134    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
135        are added as well.
136    
137    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
138        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
139        calling _pcre_jit_exec. Some extra comments are added.
140    
141    4.  (*MARK) settings inside atomic groups that do not contain any capturing
142        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
143        was introduced by change 18 for 8.20.
144    
145    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
146        ECMA-262 standard.
147    
148    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
149        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
150        This bug was probably introduced by change 9 of 8.13.
151    
152    7.  While fixing 6 above, I noticed that a number of other items were being
153        incorrectly rejected as "not fixed length". This arose partly because newer
154        opcodes had not been added to the fixed-length checking code. I have (a)
155        corrected the bug and added tests for these items, and (b) arranged for an
156        error to occur if an unknown opcode is encountered while checking for fixed
157        length instead of just assuming "not fixed length". The items that were
158        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
159        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
160        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
161    
162    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
163        being incorrectly compiled and would have given unpredicatble results.
164    
165    9.  A possessively repeated subpattern with minimum repeat count greater than
166        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
167        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
168        the first (A) could occur when it should not.
169    
170    10. Add a cast and remove a redundant test from the code.
171    
172    11. JIT should use pcre_malloc/pcre_free for allocation.
173    
174    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
175        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
176        is anything other than /usr, -L is still shown).
177    
178    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
179    
180    14. Perl does not support \N without a following name in a [] class; PCRE now
181        also gives an error.
182    
183    15. If a forward reference was repeated with an upper limit of around 2000,
184        it caused the error "internal error: overran compiling workspace". The
185        maximum number of forward references (including repeats) was limited by the
186        internal workspace, and dependent on the LINK_SIZE. The code has been
187        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
188        the default depends on LINK_SIZE. There is a new upper limit (for safety)
189        of around 200,000 forward references. While doing this, I also speeded up
190        the filling in of repeated forward references.
191    
192    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
193        incorrectly expecting the subject to contain another "a" after the start.
194    
195    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
196        in the match, the SKIP should be ignored. This was not happening; instead
197        the SKIP was being treated as NOMATCH. For patterns such as
198        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
199        tested.
200    
201    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
202        now much more compatible with Perl, in particular in cases where the result
203        is a non-match for a non-anchored pattern. For example, if
204        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
205        "m", where previously it did not return a name. A side effect of this
206        change is that for partial matches, the last encountered mark name is
207        returned, as for non matches. A number of tests that were previously not
208        Perl-compatible have been moved into the Perl-compatible test files. The
209        refactoring has had the pleasing side effect of removing one argument from
210        the match() function, thus reducing its stack requirements.
211    
212    19. If the /S+ option was used in pcretest to study a pattern using JIT,
213        subsequent uses of /S (without +) incorrectly behaved like /S+.
214    
215    21. Retrieve executable code size support for the JIT compiler and fixing
216        some warnings.
217    
218    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
219        not work when the shorter character appeared right at the end of the
220        subject string.
221    
222    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
223        systems.
224    
225    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
226        output it when the /M option is used in pcretest.
227    
228    25. The CheckMan script was not being included in the distribution. Also, added
229        an explicit "perl" to run Perl scripts from the PrepareRelease script
230        because this is reportedly needed in Windows.
231    
232    26. If study data was being save in a file and studying had not found a set of
233        "starts with" bytes for the pattern, the data written to the file (though
234        never used) was taken from uninitialized memory and so caused valgrind to
235        complain.
236    
237    27. Updated RunTest.bat as provided by Sheri Pierce.
238    
239    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
240    
241    29. Computation of memory usage for the table of capturing group names was
242        giving an unnecessarily large value.
243    
244    
245    Version 8.20 21-Oct-2011
246    ------------------------
247    
248    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
249        a POSIX class. After further experiments with Perl, which convinced me that
250        Perl has bugs and confusions, a closing square bracket is no longer allowed
251        in a POSIX name. This bug also affected patterns with classes that started
252        with full stops.
253    
254    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
255        captured substring, but while checking the failing first alternative,
256        substring 1 is temporarily captured. If the output vector supplied to
257        pcre_exec() was not big enough for this capture, the yield of the function
258        was still zero ("insufficient space for captured substrings"). This cannot
259        be totally fixed without adding another stack variable, which seems a lot
260        of expense for a edge case. However, I have improved the situation in cases
261        such as /(a)(b)x|abc/ matched against "abc", where the return code
262        indicates that fewer than the maximum number of slots in the ovector have
263        been set.
264    
265    3.  Related to (2) above: when there are more back references in a pattern than
266        slots in the output vector, pcre_exec() uses temporary memory during
267        matching, and copies in the captures as far as possible afterwards. It was
268        using the entire output vector, but this conflicts with the specification
269        that only 2/3 is used for passing back captured substrings. Now it uses
270        only the first 2/3, for compatibility. This is, of course, another edge
271        case.
272    
273    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
274        main code base, and can be used by building with --enable-jit. When this is
275        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
276        runtime --no-jit option is given.
277    
278    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
279        ovector, the return from the function was zero, implying that there were
280        other matches that did not fit. The correct "exactly full" value is now
281        returned.
282    
283    6.  If a subpattern that was called recursively or as a subroutine contained
284        (*PRUNE) or any other control that caused it to give a non-standard return,
285        invalid errors such as "Error -26 (nested recursion at the same subject
286        position)" or even infinite loops could occur.
287    
288    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
289        computing the minimum length on reaching *ACCEPT, and so ended up with the
290        wrong value of 1 rather than 0. Further investigation indicates that
291        computing a minimum subject length in the presence of *ACCEPT is difficult
292        (think back references, subroutine calls), and so I have changed the code
293        so that no minimum is registered for a pattern that contains *ACCEPT.
294    
295    8.  If (*THEN) was present in the first (true) branch of a conditional group,
296        it was not handled as intended. [But see 16 below.]
297    
298    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
299        Sheri Pierce.
300    
301    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
302        the first byte in a match must be "a".
303    
304    11. Change 17 for 8.13 increased the recursion depth for patterns like
305        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
306        pattern contains any instances of (*THEN). If it does not, the old
307        optimizations are restored. It would be nice to do this on a per-group
308        basis, but at the moment that is not feasible.
309    
310    12. In some environments, the output of pcretest -C is CRLF terminated. This
311        broke RunTest's code that checks for the link size. A single white space
312        character after the value is now allowed for.
313    
314    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
315        For "fr", it uses the Windows-specific input and output files.
316    
317    14. If (*THEN) appeared in a group that was called recursively or as a
318        subroutine, it did not work as intended. [But see next item.]
319    
320    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
321        pattern fragments (but not containing any | characters). If A and B are
322        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
323        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
324        D. In other words, Perl considers parentheses that do not contain any |
325        characters to be part of a surrounding alternative, whereas PCRE was
326        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
327        differently. PCRE now behaves in the same way as Perl, except in the case
328        of subroutine/recursion calls such as (?1) which have in any case always
329        been different (but PCRE had them first :-).
330    
331    16. Related to 15 above: Perl does not treat the | in a conditional group as
332        creating alternatives. Such a group is treated in the same way as an
333        ordinary group without any | characters when processing (*THEN). PCRE has
334        been changed to match Perl's behaviour.
335    
336    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
337        RunGrepTest script failed.
338    
339    18. Change 22 for version 13 caused atomic groups to use more stack. This is
340        inevitable for groups that contain captures, but it can lead to a lot of
341        stack use in large patterns. The old behaviour has been restored for atomic
342        groups that do not contain any capturing parentheses.
343    
344    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
345        suppress the check for a minimum subject length at run time. (If it was
346        given to pcre_exec() or pcre_dfa_exec() it did work.)
347    
348    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
349        fail to work when decoding hex characters in data strings in EBCDIC
350        environments.
351    
352    21. It appears that in at least one Mac OS environment, the isxdigit() function
353        is implemented as a macro that evaluates to its argument more than once,
354        contravening the C 90 Standard (I haven't checked a later standard). There
355        was an instance in pcretest which caused it to go wrong when processing
356        \x{...} escapes in subject strings. The has been rewritten to avoid using
357        things like p++ in the argument of isxdigit().
358    
359    
360    Version 8.13 16-Aug-2011
361    ------------------------
362    
363    1.  The Unicode data tables have been updated to Unicode 6.0.0.
364    
365    2.  Two minor typos in pcre_internal.h have been fixed.
366    
367    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
368        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
369        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
370    
371    4.  There were a number of related bugs in the code for matching backrefences
372        caselessly in UTF-8 mode when codes for the characters concerned were
373        different numbers of bytes. For example, U+023A and U+2C65 are an upper
374        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
375        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
376        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
377        2-byte code at the end of the subject (it thought there wasn't enough data
378        left).
379    
380    5.  Comprehensive information about what went wrong is now returned by
381        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
382        as the output vector has at least 2 elements. The offset of the start of
383        the failing character and a reason code are placed in the vector.
384    
385    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
386        now returned is for the first byte of the failing character, instead of the
387        last byte inspected. This is an incompatible change, but I hope it is small
388        enough not to be a problem. It makes the returned offset consistent with
389        pcre_exec() and pcre_dfa_exec().
390    
391    7.  pcretest now gives a text phrase as well as the error number when
392        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
393        failure, the offset and reason code are output.
394    
395    8.  When \R was used with a maximizing quantifier it failed to skip backwards
396        over a \r\n pair if the subsequent match failed. Instead, it just skipped
397        back over a single character (\n). This seems wrong (because it treated the
398        two characters as a single entity when going forwards), conflicts with the
399        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
400        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
401        has been changed.
402    
403    9.  Some internal refactoring has changed the processing so that the handling
404        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
405        time (the PCRE_DOTALL option was changed this way some time ago: version
406        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
407        which was always a bit of a fudge. It also means that there is one less
408        argument for the match() function, which reduces its stack requirements
409        slightly. This change also fixes an incompatibility with Perl: the pattern
410        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
411    
412    10. More internal refactoring has drastically reduced the number of recursive
413        calls to match() for possessively repeated groups such as (abc)++ when
414        using pcre_exec().
415    
416    11. While implementing 10, a number of bugs in the handling of groups were
417        discovered and fixed:
418    
419        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
420        (a|)*(?1) gave a compile-time internal error.
421        ((a|)+)+  did not notice that the outer group could match an empty string.
422        (^a|^)+   was not marked as anchored.
423        (.*a|.*)+ was not marked as matching at start or after a newline.
424    
425    12. Yet more internal refactoring has removed another argument from the match()
426        function. Special calls to this function are now indicated by setting a
427        value in a variable in the "match data" data block.
428    
429    13. Be more explicit in pcre_study() instead of relying on "default" for
430        opcodes that mean there is no starting character; this means that when new
431        ones are added and accidentally left out of pcre_study(), testing should
432        pick them up.
433    
434    14. The -s option of pcretest has been documented for ages as being an old
435        synonym of -m (show memory usage). I have changed it to mean "force study
436        for every regex", that is, assume /S for every regex. This is similar to -i
437        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
438        using it. It makes it easier to run collections of tests with and without
439        study enabled, and thereby test pcre_study() more easily. All the standard
440        tests are now run with and without -s (but some patterns can be marked as
441        "never study" - see 20 below).
442    
443    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
444        restoration of the capturing data to the outer values was not happening
445        correctly.
446    
447    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
448        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
449        pattern had matched an empty string, and so incorrectly returned a no
450        match.
451    
452    17. There was optimizing code for the last branch of non-capturing parentheses,
453        and also for the obeyed branch of a conditional subexpression, which used
454        tail recursion to cut down on stack usage. Unfortunately, now that there is
455        the possibility of (*THEN) occurring in these branches, tail recursion is
456        no longer possible because the return has to be checked for (*THEN). These
457        two optimizations have therefore been removed. [But see 8.20/11 above.]
458    
459    18. If a pattern containing \R was studied, it was assumed that \R always
460        matched two bytes, thus causing the minimum subject length to be
461        incorrectly computed because \R can also match just one byte.
462    
463    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
464        was incorrectly computed.
465    
466    20. If /S is present twice on a test pattern in pcretest input, it now
467        *disables* studying, thereby overriding the use of -s on the command line
468        (see 14 above). This is necessary for one or two tests to keep the output
469        identical in both cases.
470    
471    21. When (*ACCEPT) was used in an assertion that matched an empty string and
472        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
473    
474    22. When an atomic group that contained a capturing parenthesis was
475        successfully matched, but the branch in which it appeared failed, the
476        capturing was not being forgotten if a higher numbered group was later
477        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
478        group 1 to "a", when in fact it should be unset. This applied to multi-
479        branched capturing and non-capturing groups, repeated or not, and also to
480        positive assertions (capturing in negative assertions does not happen
481        in PCRE) and also to nested atomic groups.
482    
483    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
484        subject after a captured substring, to make it easier to tell which of a
485        number of identical substrings has been captured.
486    
487    24. The way atomic groups are processed by pcre_exec() has been changed so that
488        if they are repeated, backtracking one repetition now resets captured
489        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
490        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
491        "aaa". Previously, it would have been "a". As part of this code
492        refactoring, the way recursive calls are handled has also been changed.
493    
494    25. If an assertion condition captured any substrings, they were not passed
495        back unless some other capturing happened later. For example, if
496        (?(?=(a))a) was matched against "a", no capturing was returned.
497    
498    26. When studying a pattern that contained subroutine calls or assertions,
499        the code for finding the minimum length of a possible match was handling
500        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
501        group 1 called group 2 while simultaneously a separate group 2 called group
502        1). A stack overflow occurred in this case. I have fixed this by limiting
503        the recursion depth to 10.
504    
505    27. Updated RunTest.bat in the distribution to the version supplied by Tom
506        Fortmann. This supports explicit test numbers on the command line, and has
507        argument validation and error reporting.
508    
509    28. An instance of \X with an unlimited repeat could fail if at any point the
510        first character it looked at was a mark character.
511    
512    29. Some minor code refactoring concerning Unicode properties and scripts
513        should reduce the stack requirement of match() slightly.
514    
515    30. Added the '=' option to pcretest to check the setting of unused capturing
516        slots at the end of the pattern, which are documented as being -1, but are
517        not included in the return count.
518    
519    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
520        compiled something random. Now it gives a compile-time error (as does
521        Perl).
522    
523    32. A *MARK encountered during the processing of a positive assertion is now
524        recorded and passed back (compatible with Perl).
525    
526    33. If --only-matching or --colour was set on a pcregrep call whose pattern
527        had alternative anchored branches, the search for a second match in a line
528        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
529        matched the line "0102" twice. The same bug affected patterns that started
530        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
531        twice.
532    
533    34. Previously, PCRE did not allow quantification of assertions. However, Perl
534        does, and because of capturing effects, quantifying parenthesized
535        assertions may at times be useful. Quantifiers are now allowed for
536        parenthesized assertions.
537    
538    35. A minor code tidy in pcre_compile() when checking options for \R usage.
539    
540    36. \g was being checked for fancy things in a character class, when it should
541        just be a literal "g".
542    
543    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
544        appearance of a nested POSIX class supersedes an apparent external class.
545        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
546        unescaped square brackets may also appear as part of class names. For
547        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
548        more like Perl. (But see 8.20/1 above.)
549    
550    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
551        was because it thought it was \N{name}, which is not supported).
552    
553    39. Add minix to OS list not supporting the -S option in pcretest.
554    
555    40. PCRE tries to detect cases of infinite recursion at compile time, but it
556        cannot analyze patterns in sufficient detail to catch mutual recursions
557        such as ((?1))((?2)). There is now a runtime test that gives an error if a
558        subgroup is called recursively as a subpattern for a second time at the
559        same position in the subject string. In previous releases this might have
560        been caught by the recursion limit, or it might have run out of stack.
561    
562    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
563        happen only once. PCRE was, however incorrectly giving a compile time error
564        "recursive call could loop indefinitely" because it cannot analyze the
565        pattern in sufficient detail. The compile time test no longer happens when
566        PCRE is compiling a conditional subpattern, but actual runaway loops are
567        now caught at runtime (see 40 above).
568    
569    42. It seems that Perl allows any characters other than a closing parenthesis
570        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
571        has been changed to be the same.
572    
573    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
574        as not to get warnings when autogen.sh is called. Also changed
575        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
576    
577    44. To help people who use pcregrep to scan files containing exceedingly long
578        lines, the following changes have been made:
579    
580        (a) The default value of the buffer size parameter has been increased from
581            8K to 20K. (The actual buffer used is three times this size.)
582    
583        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
584            PCRE is built.
585    
586        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
587            to be set at run time.
588    
589        (d) Numerical values in pcregrep options can be followed by K or M, for
590            example --buffer-size=50K.
591    
592        (e) If a line being scanned overflows pcregrep's buffer, an error is now
593            given and the return code is set to 2.
594    
595    45. Add a pointer to the latest mark to the callout data block.
596    
597    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
598        partial match of an empty string instead of no match. This was specific to
599        the use of ".".
600    
601    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
602        complete match instead of a partial match. This bug was dependent on both
603        the PCRE_UTF8 and PCRE_DOTALL options being set.
604    
605    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
606        starting byte set, because \b was not being ignored.
607    
608    
609    Version 8.12 15-Jan-2011
610    ------------------------
611    
612    1.  Fixed some typos in the markup of the man pages, and wrote a script that
613        checks for such things as part of the documentation building process.
614    
615    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
616        --match-limit and --recursion-limit options (added for 8.11). In
617        particular, this made one of the standard tests fail. (The integer value
618        went into the wrong half of a long int.)
619    
620    3.  If the --colour option was given to pcregrep with -v (invert match), it
621        did strange things, either producing crazy output, or crashing. It should,
622        of course, ignore a request for colour when reporting lines that do not
623        match.
624    
625    4.  Another pcregrep bug caused similar problems if --colour was specified with
626        -M (multiline) and the pattern match finished with a line ending.
627    
628    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
629        matched in multiline mode, the following line was shown as part of the
630        match. This seems wrong, so I have changed it.
631    
632    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
633        the check for further matches in the same line (so they could be coloured)
634        to overrun the end of the current line. If another match was found, it was
635        incorrectly shown (and then shown again when found in the next line).
636    
637    7.  If pcregrep was compiled under Windows, there was a reference to the
638        function pcregrep_exit() before it was defined. I am assuming this was
639        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
640        reported by a user. I've moved the definition above the reference.
641    
642    
643    Version 8.11 10-Dec-2010
644    ------------------------
645    
646    1.  (*THEN) was not working properly if there were untried alternatives prior
647        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
648        backtracked to try for "b" instead of moving to the next alternative branch
649        at the same level (in this case, to look for "c"). The Perl documentation
650        is clear that when (*THEN) is backtracked onto, it goes to the "next
651        alternative in the innermost enclosing group".
652    
653    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
654        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
655        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
656        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
657        (*THEN).
658    
659    3.  If \s appeared in a character class, it removed the VT character from
660        the class, even if it had been included by some previous item, for example
661        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
662        of \s, but is part of the POSIX "space" class.)
663    
664    4.  A partial match never returns an empty string (because you can always
665        match an empty string at the end of the subject); however the checking for
666        an empty string was starting at the "start of match" point. This has been
667        changed to the "earliest inspected character" point, because the returned
668        data for a partial match starts at this character. This means that, for
669        example, /(?<=abc)def/ gives a partial match for the subject "abc"
670        (previously it gave "no match").
671    
672    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
673        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
674        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
675        has an implication that the given string is incomplete (because a partial
676        match is preferred over a full match). For this reason, these items now
677        give a partial match in this situation. [Aside: previously, the one case
678        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
679        match rather than a full match, which was wrong by the old rules, but is
680        now correct.]
681    
682    6.  There was a bug in the handling of #-introduced comments, recognized when
683        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
684        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
685        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
686        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
687        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
688        places in pcre_compile().
689    
690    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
691        comments when looking ahead for named forward references to subpatterns,
692        the only newline sequence it recognized was NL. It now handles newlines
693        according to the set newline convention.
694    
695    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
696        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
697        cater for a lack of strerror(). These oversights have been fixed.
698    
699    9.  Added --match-limit and --recursion-limit to pcregrep.
700    
701    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
702    
703    11. When the -o option was used, pcregrep was setting a return code of 1, even
704        when matches were found, and --line-buffered was not being honoured.
705    
706    12. Added an optional parentheses number to the -o and --only-matching options
707        of pcregrep.
708    
709    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
710        can match an empty string. The code to do it in pcretest and pcredemo
711        needed fixing:
712    
713        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
714            only one byte after an empty string match just before CRLF (this case
715            just got forgotten; "any" and "anycrlf" were OK).
716    
717        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
718            mode when an empty string match preceded an ASCII character followed by
719            a non-ASCII character. (The code for advancing by one character rather
720            than one byte was nonsense.)
721    
722        (c) The pcredemo.c sample program did not have any code at all to handle
723            the cases when CRLF is a valid newline sequence.
724    
725    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
726        as a starting offset was within the subject string. There is now a new
727        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
728        negative or greater than the length of the string. In order to test this,
729        pcretest is extended to allow the setting of negative starting offsets.
730    
731    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
732        starting offset points to the beginning of a UTF-8 character was
733        unnecessarily clumsy. I tidied it up.
734    
735    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
736        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
737    
738    17. Nobody had reported that the --include_dir option, which was added in
739        release 7.7 should have been called --include-dir (hyphen, not underscore)
740        for compatibility with GNU grep. I have changed it to --include-dir, but
741        left --include_dir as an undocumented synonym, and the same for
742        --exclude-dir, though that is not available in GNU grep, at least as of
743        release 2.5.4.
744    
745    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
746        characters from a string of bytes) have been redefined so as not to use
747        loops, in order to improve performance in some environments. At the same
748        time, I abstracted some of the common code into auxiliary macros to save
749        repetition (this should not affect the compiled code).
750    
751    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
752        compile-time error is now given if \c is not followed by an ASCII
753        character, that is, a byte less than 128. (In EBCDIC mode, the code is
754        different, and any byte value is allowed.)
755    
756    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
757        START_OPTIMIZE option, which is now allowed at compile time - but just
758        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
759        to pcregrep and other applications that have no direct access to PCRE
760        options. The new /Y option in pcretest sets this option when calling
761        pcre_compile().
762    
763    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
764        back references. Groups containing recursive back references were forced to
765        be atomic by that change, but in the case of named groups, the amount of
766        memory required was incorrectly computed, leading to "Failed: internal
767        error: code overflow". This has been fixed.
768    
769    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
770        pcretest.c, to avoid build problems in some Borland environments.
771    
772    
773    Version 8.10 25-Jun-2010
774    ------------------------
775    
776    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
777        THEN.
778    
779    2.  (*ACCEPT) was not working when inside an atomic group.
780    
781    3.  Inside a character class, \B is treated as a literal by default, but
782        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
783        causes the error). The code is unchanged, but I tidied the documentation.
784    
785    4.  Inside a character class, PCRE always treated \R and \X as literals,
786        whereas Perl faults them if its -w option is set. I have changed PCRE so
787        that it faults them when PCRE_EXTRA is set.
788    
789    5.  Added support for \N, which always matches any character other than
790        newline. (It is the same as "." when PCRE_DOTALL is not set.)
791    
792    6.  When compiling pcregrep with newer versions of gcc which may have
793        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
794        declared with attribute warn_unused_result" were given. Just casting the
795        result to (void) does not stop the warnings; a more elaborate fudge is
796        needed. I've used a macro to implement this.
797    
798    7.  Minor change to pcretest.c to avoid a compiler warning.
799    
800    8.  Added four artifical Unicode properties to help with an option to make
801        \s etc use properties (see next item). The new properties are: Xan
802        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
803    
804    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
805        use Unicode properties. (*UCP) at the start of a pattern can be used to set
806        this option. Modified pcretest to add /W to test this facility. Added
807        REG_UCP to make it available via the POSIX interface.
808    
809    10. Added --line-buffered to pcregrep.
810    
811    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
812        studied, and the match started with a letter with a code point greater than
813        127 whose first byte was different to the first byte of the other case of
814        the letter, the other case of this starting letter was not recognized
815        (#976).
816    
817    12. If a pattern that was studied started with a repeated Unicode property
818        test, for example, \p{Nd}+, there was the theoretical possibility of
819        setting up an incorrect bitmap of starting bytes, but fortunately it could
820        not have actually happened in practice until change 8 above was made (it
821        added property types that matched character-matching opcodes).
822    
823    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
824        possible starting bytes for non-anchored patterns.
825    
826    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
827        \R, and also a number of cases that involve Unicode properties, both
828        explicit and implicit when PCRE_UCP is set.
829    
830    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
831        input, it could crash or give wrong results if characters with values
832        greater than 0xc0 were present in the subject string. (Detail: it assumed
833        UTF-8 input when processing these items.)
834    
835    16. Added a lot of (int) casts to avoid compiler warnings in systems where
836        size_t is 64-bit (#991).
837    
838    17. Added a check for running out of memory when PCRE is compiled with
839        --disable-stack-for-recursion (#990).
840    
841    18. If the last data line in a file for pcretest does not have a newline on
842        the end, a newline was missing in the output.
843    
844    19. The default pcre_chartables.c file recognizes only ASCII characters (values
845        less than 128) in its various bitmaps. However, there is a facility for
846        generating tables according to the current locale when PCRE is compiled. It
847        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
848        characters, are recognized by isspace() and therefore were getting set in
849        these tables, and indeed these tables seem to approximate to ISO 8859. This
850        caused a problem in UTF-8 mode when pcre_study() was used to create a list
851        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
852        which of course cannot start UTF-8 characters. I have changed the code so
853        that only real ASCII characters (less than 128) and the correct starting
854        bytes for UTF-8 encodings are set for characters greater than 127 when in
855        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
856        altogether.)
857    
858    20. Added the /T option to pcretest so as to be able to run tests with non-
859        standard character tables, thus making it possible to include the tests
860        used for 19 above in the standard set of tests.
861    
862    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
863        reference to a subpattern the other side of a comment that contains an
864        opening parenthesis caused either an internal compiling error, or a
865        reference to the wrong subpattern.
866    
867    
868    Version 8.02 19-Mar-2010
869    ------------------------
870    
871    1.  The Unicode data tables have been updated to Unicode 5.2.0.
872    
873    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
874        configured.
875    
876    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
877        original author of that file, following a query about its status.
878    
879    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
880        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
881    
882    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
883        quantifier applied to a forward-referencing subroutine call, could compile
884        incorrect code or give the error "internal error: previously-checked
885        referenced subpattern not found".
886    
887    6.  Both MS Visual Studio and Symbian OS have problems with initializing
888        variables to point to external functions. For these systems, therefore,
889        pcre_malloc etc. are now initialized to local functions that call the
890        relevant global functions.
891    
892    7.  There were two entries missing in the vectors called coptable and poptable
893        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
894        I've fixed the data, and added a kludgy way of testing at compile time that
895        the lengths are correct (equal to the number of opcodes).
896    
897    8.  Following on from 7, I added a similar kludge to check the length of the
898        eint vector in pcreposix.c.
899    
900    9.  Error texts for pcre_compile() are held as one long string to avoid too
901        much relocation at load time. To find a text, the string is searched,
902        counting zeros. There was no check for running off the end of the string,
903        which could happen if a new error number was added without updating the
904        string.
905    
906    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
907    
908    11. \K was not working if it appeared in an atomic group or in a group that
909        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
910        \K is "not well defined" if used in an assertion. PCRE now accepts it if
911        the assertion is positive, but not if it is negative.
912    
913    12. Change 11 fortuitously reduced the size of the stack frame used in the
914        "match()" function of pcre_exec.c by one pointer. Forthcoming
915        implementation of support for (*MARK) will need an extra pointer on the
916        stack; I have reserved it now, so that the stack frame size does not
917        decrease.
918    
919    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
920        item in branch that calls a recursion is a subroutine call - as in the
921        second branch in the above example - was incorrectly given the compile-
922        time error "recursive call could loop indefinitely" because pcre_compile()
923        was not correctly checking the subroutine for matching a non-empty string.
924    
925    14. The checks for overrunning compiling workspace could trigger after an
926        overrun had occurred. This is a "should never occur" error, but it can be
927        triggered by pathological patterns such as hundreds of nested parentheses.
928        The checks now trigger 100 bytes before the end of the workspace.
929    
930    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
931    
932    
933    Version 8.01 19-Jan-2010
934    ------------------------
935    
936    1.  If a pattern contained a conditional subpattern with only one branch (in
937        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
938        computed the wrong minimum data length (which is of course zero for such
939        subpatterns). This could cause incorrect "no match" results.
940    
941    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
942        the pattern is reset in the first branch, pcre_compile() failed with
943        "internal error: code overflow at offset...". This happened only when
944        the reset was to the original external option setting. (An optimization
945        abstracts leading options settings into an external setting, which was the
946        cause of this.)
947    
948    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
949        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
950        assertion pattern did not match (meaning that the assertion was true), it
951        was incorrectly treated as false if the SKIP had been reached during the
952        matching. This also applied to assertions used as conditions.
953    
954    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
955        assertion subpattern, including such a pattern used as a condition,
956        unpredictable results occurred, instead of the error return
957        PCRE_ERROR_DFA_UITEM.
958    
959    5.  The C++ GlobalReplace function was not working like Perl for the special
960        situation when an empty string is matched. It now does the fancy magic
961        stuff that is necessary.
962    
963    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
964        removed. (These were left over from very, very early versions of PCRE.)
965    
966    7.  Some cosmetic changes to the code to make life easier when compiling it
967        as part of something else:
968    
969        (a) Change DEBUG to PCRE_DEBUG.
970    
971        (b) In pcre_compile(), rename the member of the "branch_chain" structure
972            called "current" as "current_branch", to prevent a collision with the
973            Linux macro when compiled as a kernel module.
974    
975        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
976            prevent a collision with the Linux macro when compiled as a kernel
977            module.
978    
979    8.  In pcre_compile() there are some checks for integer overflows that used to
980        cast potentially large values to (double). This has been changed to that
981        when building, a check for int64_t is made, and if it is found, it is used
982        instead, thus avoiding the use of floating point arithmetic. (There is no
983        other use of FP in PCRE.) If int64_t is not found, the fallback is to
984        double.
985    
986    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
987        2005 (difference between two addresses compared to an unsigned value).
988    
989    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
990        custom one, because of the following reported problem in Windows:
991    
992          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
993              under Win32.
994          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
995              therefore missing the function definition.
996          - The compiler thus generates a "C" signature for the test function.
997          - The linker fails to find the "C" function.
998          - PCRE fails to configure if asked to do so against libbz2.
999    
1000    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1001        messages were output:
1002    
1003          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1004          rerunning libtoolize, to keep the correct libtool macros in-tree.
1005          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1006    
1007        I have done both of these things.
1008    
1009    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1010        most of the time, it *can* run out if it is given a pattern that contains a
1011        runaway infinite recursion. I updated the discussion in the pcrestack man
1012        page.
1013    
1014    13. Now that we have gone to the x.xx style of version numbers, the minor
1015        version may start with zero. Using 08 or 09 is a bad idea because users
1016        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1017        interpreted as invalid octal numbers. I've updated the previous comment in
1018        configure.ac, and also added a check that gives an error if 08 or 09 are
1019        used.
1020    
1021    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1022        causing partial matching to fail when the end of the subject matched \W
1023        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1024    
1025    15. There were some discrepancies between the declarations in pcre_internal.h
1026        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1027        their definitions. The declarations used "const uschar *" and the
1028        definitions used USPTR. Even though USPTR is normally defined as "const
1029        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1030        reported that: "This difference in casting confuses some C++ compilers, for
1031        example, SunCC recognizes above declarations as different functions and
1032        generates broken code for hbpcre." I have changed the declarations to use
1033        USPTR.
1034    
1035    16. GNU libtool is named differently on some systems. The autogen.sh script now
1036        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1037        (FreeBSD).
1038    
1039    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1040        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1041        comment: "Figure out how to create a longlong from a string: strtoll and
1042        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1043        instance, but it only takes 2 args instead of 3!"
1044    
1045    18. A subtle bug concerned with back references has been fixed by a change of
1046        specification, with a corresponding code fix. A pattern such as
1047        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1048        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1049        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1050        same bug. Such groups have to be quantified to be useful, or contained
1051        inside another quantified group. (If there's no repetition, the reference
1052        can never match.) The problem arises because, having left the group and
1053        moved on to the rest of the pattern, a later failure that backtracks into
1054        the group uses the captured value from the final iteration of the group
1055        rather than the correct earlier one. I have fixed this in PCRE by forcing
1056        any group that contains a reference to itself to be an atomic group; that
1057        is, there cannot be any backtracking into it once it has completed. This is
1058        similar to recursive and subroutine calls.
1059    
1060    
1061    Version 8.00 19-Oct-09
1062    ----------------------
1063    
1064    1.  The table for translating pcre_compile() error codes into POSIX error codes
1065        was out-of-date, and there was no check on the pcre_compile() error code
1066        being within the table. This could lead to an OK return being given in
1067        error.
1068    
1069    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1070        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1071        in a Windows environment.
1072    
1073    3.  The pcregrep --count option prints the count for each file even when it is
1074        zero, as does GNU grep. However, pcregrep was also printing all files when
1075        --files-with-matches was added. Now, when both options are given, it prints
1076        counts only for those files that have at least one match. (GNU grep just
1077        prints the file name in this circumstance, but including the count seems
1078        more useful - otherwise, why use --count?) Also ensured that the
1079        combination -clh just lists non-zero counts, with no names.
1080    
1081    4.  The long form of the pcregrep -F option was incorrectly implemented as
1082        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1083        but it seems right to fix it, and I didn't think it was worth preserving
1084        the old behaviour.
1085    
1086    5.  The command line items --regex=pattern and --regexp=pattern were not
1087        recognized by pcregrep, which required --regex pattern or --regexp pattern
1088        (with a space rather than an '='). The man page documented the '=' forms,
1089        which are compatible with GNU grep; these now work.
1090    
1091    6.  No libpcreposix.pc file was created for pkg-config; there was just
1092        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1093    
1094    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1095        when UCP support is not needed, by modifying the Python script that
1096        generates it from Unicode data files. This should not matter if the module
1097        is correctly used as a library, but I received one complaint about 50K of
1098        unwanted data. My guess is that the person linked everything into his
1099        program rather than using a library. Anyway, it does no harm.
1100    
1101    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1102        was a minimum greater than 1 for a wide character in a possessive
1103        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1104        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1105        character. Chaos in the form of incorrect output or a compiling loop could
1106        result.
1107    
1108    9.  The restrictions on what a pattern can contain when partial matching is
1109        requested for pcre_exec() have been removed. All patterns can now be
1110        partially matched by this function. In addition, if there are at least two
1111        slots in the offset vector, the offset of the earliest inspected character
1112        for the match and the offset of the end of the subject are set in them when
1113        PCRE_ERROR_PARTIAL is returned.
1114    
1115    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1116        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1117        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1118        and may be more useful for multi-segment matching.
1119    
1120    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1121        used to be given if ever the end of the subject was reached; now it is
1122        given only if matching could not proceed because another character was
1123        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1124        string "Z", which now yields "no match" instead of "partial match". In the
1125        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1126        final character ended with (*FAIL).
1127    
1128    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1129        if the pattern had a "must contain" character that was already found in the
1130        earlier partial match, unless partial matching was again requested. For
1131        example, with the pattern /dog.(body)?/, the "must contain" character is
1132        "g". If the first part-match was for the string "dog", restarting with
1133        "sbody" failed. This bug has been fixed.
1134    
1135    13. The string returned by pcre_dfa_exec() after a partial match has been
1136        changed so that it starts at the first inspected character rather than the
1137        first character of the match. This makes a difference only if the pattern
1138        starts with a lookbehind assertion or \b or \B (\K is not supported by
1139        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1140        matching functions compatible, and I think it's the right thing to do.
1141    
1142    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1143        so that the demonstration program is easily available in environments where
1144        PCRE has not been installed from source.
1145    
1146    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1147        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1148        library.
1149    
1150    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1151        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1152        is not the first non-POSIX option to be added. Clearly some people find
1153        these options useful.
1154    
1155    17. If a caller to the POSIX matching function regexec() passes a non-zero
1156        value for nmatch with a NULL value for pmatch, the value of
1157        nmatch is forced to zero.
1158    
1159    18. RunGrepTest did not have a test for the availability of the -u option of
1160        the diff command, as RunTest does. It now checks in the same way as
1161        RunTest, and also checks for the -b option.
1162    
1163    19. If an odd number of negated classes containing just a single character
1164        interposed, within parentheses, between a forward reference to a named
1165        subpattern and the definition of the subpattern, compilation crashed with
1166        an internal error, complaining that it could not find the referenced
1167        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1168        [The bug was that it was starting one character too far in when skipping
1169        over the character class, thus treating the ] as data rather than
1170        terminating the class. This meant it could skip too much.]
1171    
1172    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1173        /g option in pcretest when the pattern contains \K, which makes it possible
1174        to have an empty string match not at the start, even when the pattern is
1175        anchored. Updated pcretest and pcredemo to use this option.
1176    
1177    21. If the maximum number of capturing subpatterns in a recursion was greater
1178        than the maximum at the outer level, the higher number was returned, but
1179        with unset values at the outer level. The correct (outer level) value is
1180        now given.
1181    
1182    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1183        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1184        make it do so. The string so far is captured, making this feature
1185        compatible with Perl.
1186    
1187    23. The tests have been re-organized, adding tests 11 and 12, to make it
1188        possible to check the Perl 5.10 features against Perl 5.10.
1189    
1190    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1191        pattern matches a fixed length string. PCRE did not allow this; now it
1192        does. Neither allows recursion.
1193    
1194    25. I finally figured out how to implement a request to provide the minimum
1195        length of subject string that was needed in order to match a given pattern.
1196        (It was back references and recursion that I had previously got hung up
1197        on.) This code has now been added to pcre_study(); it finds a lower bound
1198        to the length of subject needed. It is not necessarily the greatest lower
1199        bound, but using it to avoid searching strings that are too short does give
1200        some useful speed-ups. The value is available to calling programs via
1201        pcre_fullinfo().
1202    
1203    26. While implementing 25, I discovered to my embarrassment that pcretest had
1204        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1205        study optimizations had never been tested with that matching function.
1206        Oops. What is worse, even when it was passed study data, there was a bug in
1207        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1208        were also very few tests of studied patterns with pcre_dfa_exec().
1209    
1210    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1211        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1212        on the other side of the coin, they are no longer allowed to have different
1213        names, because these cannot be distinguished in PCRE, and this has caused
1214        confusion. (This is a difference from Perl.)
1215    
1216    28. When duplicate subpattern names are present (necessarily with different
1217        numbers, as required by 27 above), and a test is made by name in a
1218        conditional pattern, either for a subpattern having been matched, or for
1219        recursion in such a pattern, all the associated numbered subpatterns are
1220        tested, and the overall condition is true if the condition is true for any
1221        one of them. This is the way Perl works, and is also more like the way
1222        testing by number works.
1223    
1224    
1225    Version 7.9 11-Apr-09
1226    ---------------------
1227    
1228    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1229        (pcretest), all targets were linked against these libraries. This included
1230        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1231        libraries. This caused unwanted dependencies to be created. This problem
1232        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1233        pcretest is linked with readline.
1234    
1235    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1236        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1237        moved outside it again, because FALSE and TRUE are already defined in AIX,
1238        but BOOL is not.
1239    
1240    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1241        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1242    
1243    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1244        hyphens) following file names and line numbers when outputting matching
1245        lines. This is not true; no spaces are inserted. I have also clarified the
1246        wording for the --colour (or --color) option.
1247    
1248    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1249        was not coloured; this is different to GNU grep, so I have changed it to be
1250        the same.
1251    
1252    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1253        each matching line was coloured. Now it goes on to look for further matches
1254        of any of the test patterns, which is the same behaviour as GNU grep.
1255    
1256    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1257        doesn't make sense to accept an empty string match in pcregrep, so I have
1258        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1259        seems to be how GNU grep behaves.
1260    
1261    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1262        start or after a newline", because the conditional assertion was not being
1263        correctly handled. The rule now is that both the assertion and what follows
1264        in the first alternative must satisfy the test.
1265    
1266    9.  If auto-callout was enabled in a pattern with a conditional group whose
1267        condition was an assertion, PCRE could crash during matching, both with
1268        pcre_exec() and pcre_dfa_exec().
1269    
1270    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1271        used for matching.
1272    
1273    11. Unicode property support in character classes was not working for
1274        characters (bytes) greater than 127 when not in UTF-8 mode.
1275    
1276    12. Added the -M command line option to pcretest.
1277    
1278    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1279    
1280    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1281    
1282    16. Added comments and documentation about mis-use of no_arg in the C++
1283        wrapper.
1284    
1285    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1286        from Martin Jerabek that uses macro names for all relevant character and
1287        string constants.
1288    
1289    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1290        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1291        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1292        these, but not everybody uses configure.
1293    
1294    19. A conditional group that had only one branch was not being correctly
1295        recognized as an item that could match an empty string. This meant that an
1296        enclosing group might also not be so recognized, causing infinite looping
1297        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1298        with the subject "ab", where knowledge that the repeated group can match
1299        nothing is needed in order to break the loop.
1300    
1301    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1302        exec(), but without supplying a callout function, matching went wrong.
1303    
1304    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1305        leak if the size of the offset vector was greater than 30. When the vector
1306        is smaller, the saved offsets during recursion go onto a local stack
1307        vector, but for larger vectors malloc() is used. It was failing to free
1308        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1309        error, in fact).
1310    
1311    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1312        heapframe that is used only when UTF-8 support is enabled. This caused no
1313        problem, but was untidy.
1314    
1315    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1316        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1317        included within another project.
1318    
1319    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1320        slightly modified by me:
1321    
1322          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1323              not building pcregrep.
1324    
1325          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1326              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1327    
1328    25. Forward references, both numeric and by name, in patterns that made use of
1329        duplicate group numbers, could behave incorrectly or give incorrect errors,
1330        because when scanning forward to find the reference group, PCRE was not
1331        taking into account the duplicate group numbers. A pattern such as
1332        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1333    
1334    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1335        the feature of a custom pointer more persuasive (as requested by a user).
1336    
1337    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1338        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1339        pre-defined.
1340    
1341    28. Added support for (*UTF8) at the start of a pattern.
1342    
1343    29. Arrange for flags added by the "release type" setting in CMake to be shown
1344        in the configuration summary.
1345    
1346    
1347    Version 7.8 05-Sep-08
1348    ---------------------
1349    
1350    1.  Replaced UCP searching code with optimized version as implemented for Ad
1351        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1352        stage table and inline lookup instead of a function, giving speed ups of 2
1353        to 5 times on some simple patterns that I tested. Permission was given to
1354        distribute the MultiStage2.py script that generates the tables (it's not in
1355        the tarball, but is in the Subversion repository).
1356    
1357    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1358        scripts.
1359    
1360    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1361        a group with a zero qualifier. The result of the study could be incorrect,
1362        or the function might crash, depending on the pattern.
1363    
1364    4.  Caseless matching was not working for non-ASCII characters in back
1365        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1366        It now works when Unicode Property Support is available.
1367    
1368    5.  In pcretest, an escape such as \x{de} in the data was always generating
1369        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1370        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1371        truncation.
1372    
1373    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1374    
1375    7.  Added two (int) casts to pcregrep when printing the difference of two
1376        pointers, in case they are 64-bit values.
1377    
1378    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1379        test 2 if it fails.
1380    
1381    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1382        and a #define of that name to empty if it is not externally set. This is to
1383        allow users of MSVC to set it if necessary.
1384    
1385    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1386        the convenience functions in the pcre_get.c source file.
1387    
1388    11. An option change at the start of a pattern that had top-level alternatives
1389        could cause overwriting and/or a crash. This command provoked a crash in
1390        some environments:
1391    
1392          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1393    
1394        This potential security problem was recorded as CVE-2008-2371.
1395    
1396    12. For a pattern where the match had to start at the beginning or immediately
1397        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1398        pcre_dfa_exec() could read past the end of the passed subject if there was
1399        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1400        pcretest so that it places the subject at the end of its malloc-ed buffer.
1401    
1402    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1403        exec() might read past the end of the data buffer in UTF-8 mode.
1404    
1405    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1406        the data contained the byte 0x85 as part of a UTF-8 character within its
1407        first line. This applied both to normal and DFA matching.
1408    
1409    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1410        /^[^d]*?$/8 failed to match "abc".
1411    
1412    16. Added a missing copyright notice to pcrecpp_internal.h.
1413    
1414    17. Make it more clear in the documentation that values returned from
1415        pcre_exec() in ovector are byte offsets, not character counts.
1416    
1417    18. Tidied a few places to stop certain compilers from issuing warnings.
1418    
1419    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1420        supplied by Stefan Weber. I made a further small update for 7.8 because
1421        there is a change of source arrangements: the pcre_searchfuncs.c module is
1422        replaced by pcre_ucd.c.
1423    
1424    
1425    Version 7.7 07-May-08
1426    ---------------------
1427    
1428    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1429        a string to a long long, pretend we don't even have a long long." This is
1430        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1431    
1432    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1433        pre-7.6 versions, which defined a global no_arg variable instead of putting
1434        it in the RE class. (See also #8 below.)
1435    
1436    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1437        Lopes.
1438    
1439    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1440    
1441        (1) The include/exclude patterns were being applied to the whole pathnames
1442            of files, instead of just to the final components.
1443    
1444        (2) If there was more than one level of directory, the subdirectories were
1445            skipped unless they satisfied the include/exclude conditions. This is
1446            inconsistent with GNU grep (and could even be seen as contrary to the
1447            pcregrep specification - which I improved to make it absolutely clear).
1448            The action now is always to scan all levels of directory, and just
1449            apply the include/exclude patterns to regular files.
1450    
1451    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1452        --exclude_dir in the tests to avoid scanning .svn directories.
1453    
1454    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1455        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1456        doesn't support NULs in patterns.
1457    
1458    7.  Added some missing "const"s to declarations of static tables in
1459        pcre_compile.c and pcre_dfa_exec.c.
1460    
1461    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1462        caused by fix #2  above. (Subsequently also a second patch to fix the
1463        first patch. And a third patch - this was a messy problem.)
1464    
1465    9.  Applied Craig's patch to remove the use of push_back().
1466    
1467    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1468        matching function regexec().
1469    
1470    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1471        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1472        references. PCRE supports relative numbers with this syntax (I don't think
1473        Oniguruma does).
1474    
1475    12. Previously, a group with a zero repeat such as (...){0} was completely
1476        omitted from the compiled regex. However, this means that if the group
1477        was called as a subroutine from elsewhere in the pattern, things went wrong
1478        (an internal error was given). Such groups are now left in the compiled
1479        pattern, with a new opcode that causes them to be skipped at execution
1480        time.
1481    
1482    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1483        to the way PCRE behaves:
1484    
1485        (a) A lone ] character is dis-allowed (Perl treats it as data).
1486    
1487        (b) A back reference to an unmatched subpattern matches an empty string
1488            (Perl fails the current match path).
1489    
1490        (c) A data ] in a character class must be notated as \] because if the
1491            first data character in a class is ], it defines an empty class. (In
1492            Perl it is not possible to have an empty class.) The empty class []
1493            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1494            The negative empty class [^] matches any one character, independently
1495            of the DOTALL setting.
1496    
1497    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1498        non-existent subpattern following a character class starting with ']' and
1499        containing () gave an internal compiling error instead of "reference to
1500        non-existent subpattern". Fortunately, when the pattern did exist, the
1501        compiled code was correct. (When scanning forwards to check for the
1502        existencd of the subpattern, it was treating the data ']' as terminating
1503        the class, so got the count wrong. When actually compiling, the reference
1504        was subsequently set up correctly.)
1505    
1506    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1507        it was being rejected as not supported by pcre_dfa_exec(), even though
1508        other assertions are supported. I have made pcre_dfa_exec() support
1509        (*FAIL).
1510    
1511    16. The implementation of 13c above involved the invention of a new opcode,
1512        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1513        cannot be changed at match time, I realized I could make a small
1514        improvement to matching performance by compiling OP_ALLANY instead of
1515        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1516        on the OP_ANY path.
1517    
1518    17. Compiling pcretest on Windows with readline support failed without the
1519        following two fixes: (1) Make the unistd.h include conditional on
1520        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1521    
1522    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1523        ncurses library to be included for pcretest when ReadLine support is
1524        requested, but also to allow for it to be overridden. This patch came from
1525        Daniel Bergström.
1526    
1527    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1528        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1529        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1530        spotting this.
1531    
1532    
1533    Version 7.6 28-Jan-08
1534    ---------------------
1535    
1536    1.  A character class containing a very large number of characters with
1537        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1538        overflow.
1539    
1540    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1541        HAVE_LONG_LONG is not defined.
1542    
1543    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1544        bring them up to date and include new features. This patch includes:
1545    
1546        - Fixed PH's badly added libz and libbz2 support.
1547        - Fixed a problem with static linking.
1548        - Added pcredemo. [But later removed - see 7 below.]
1549        - Fixed dftables problem and added an option.
1550        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1551            HAVE_LONG_LONG.
1552        - Added readline support for pcretest.
1553        - Added an listing of the option settings after cmake has run.
1554    
1555    4.  A user submitted a patch to Makefile that makes it easy to create
1556        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1557        Makefile.am that cause it to include this special target, without
1558        affecting anything else. Note that the same mingw target plus all
1559        the other distribution libraries and programs are now supported
1560        when configuring with CMake (see 6 below) instead of with
1561        Configure/Make.
1562    
1563    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1564        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1565        exported in the Windows port". It has not yet been confirmed that the patch
1566        solves the problem, but it does no harm.
1567    
1568    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1569        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1570        with CMake, and also correct the comment about stack recursion.
1571    
1572    7.  Remove the automatic building of pcredemo from the ./configure system and
1573        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1574        of a program that users should build themselves after PCRE is installed, so
1575        building it automatically is not really right. What is more, it gave
1576        trouble in some build environments.
1577    
1578    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1579    
1580    
1581    Version 7.5 10-Jan-08
1582    ---------------------
1583    
1584    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1585        values in parens when parsing an RE using the C++ wrapper."
1586    
1587    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1588        Characters greater than 255 were excluded from the class instead of being
1589        included.
1590    
1591    3.  The same bug as (2) above applied to negated POSIX classes such as
1592        [:^space:].
1593    
1594    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1595        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1596        I have changed it.
1597    
1598    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1599        first named subpattern) and a construct such as (?&a) would reference the
1600        first named subpattern whose name started with "a" (in other words, the
1601        length check was missing). Both these problems are fixed. "Subpattern name
1602        expected" is now given for (?&) (a zero-length name), and this patch also
1603        makes it give the same error for \k'' (previously it complained that that
1604        was a reference to a non-existent subpattern).
1605    
1606    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1607        this is right because (?- can be followed by option settings as well as by
1608        digits. I have, however, made the messages clearer.
1609    
1610    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1611        than the number used in the conditional) now cause a compile-time error.
1612        This is actually not compatible with Perl, which accepts such patterns, but
1613        treats the conditional as always being FALSE (as PCRE used to), but it
1614        seems to me that giving a diagnostic is better.
1615    
1616    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1617        and messages.
1618    
1619    9.  Fix two occurrences of "backslash" in comments that should have been
1620        "backspace".
1621    
1622    10. Remove two redundant lines of code that can never be obeyed (their function
1623        was moved elsewhere).
1624    
1625    11. The program that makes PCRE's Unicode character property table had a bug
1626        which caused it to generate incorrect table entries for sequences of
1627        characters that have the same character type, but are in different scripts.
1628        It amalgamated them into a single range, with the script of the first of
1629        them. In other words, some characters were in the wrong script. There were
1630        thirteen such cases, affecting characters in the following ranges:
1631    
1632          U+002b0 - U+002c1
1633          U+0060c - U+0060d
1634          U+0061e - U+00612
1635          U+0064b - U+0065e
1636          U+0074d - U+0076d
1637          U+01800 - U+01805
1638          U+01d00 - U+01d77
1639          U+01d9b - U+01dbf
1640          U+0200b - U+0200f
1641          U+030fc - U+030fe
1642          U+03260 - U+0327f
1643          U+0fb46 - U+0fbb1
1644          U+10450 - U+1049d
1645    
1646    12. The -o option (show only the matching part of a line) for pcregrep was not
1647        compatible with GNU grep in that, if there was more than one match in a
1648        line, it showed only the first of them. It now behaves in the same way as
1649        GNU grep.
1650    
1651    13. If the -o and -v options were combined for pcregrep, it printed a blank
1652        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1653        does the same. The return code can be used to tell if there were any
1654        non-matching lines.
1655    
1656    14. Added --file-offsets and --line-offsets to pcregrep.
1657    
1658    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1659        infinitely looping recursion. The bug was that positive lookaheads were not
1660        being skipped when checking for a possible empty match (negative lookaheads
1661        and both kinds of lookbehind were skipped).
1662    
1663    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1664        inclusion of <windows.h> to before rather than after the definition of
1665        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1666    
1667    17. Specifying a possessive quantifier with a specific limit for a Unicode
1668        character property caused pcre_compile() to compile bad code, which led at
1669        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1670        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1671        caused the error; without that there was no problem.
1672    
1673    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1674    
1675    19. Added --enable-pcretest-libreadline.
1676    
1677    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1678        RE::GlobalReplace(). As a result, the number of replacements returned was
1679        double what it should be. I removed one of the increments, but Craig sent a
1680        later patch that removed the other one (the right fix) and added unit tests
1681        that check the return values (which was not done before).
1682    
1683    21. Several CMake things:
1684    
1685        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1686            the names libpcre and libpcreposix, not just pcre and pcreposix.
1687    
1688        (2) The above change means that pcretest and pcregrep are now correctly
1689            linked with the newly-built libraries, not previously installed ones.
1690    
1691        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1692    
1693    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1694        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1695        UTF-8 newline character). The key issue is that the pattern starts .*;
1696        this means that the match must be either at the beginning, or after a
1697        newline. The bug was in the code for advancing after a failed match and
1698        checking that the new position followed a newline. It was not taking
1699        account of UTF-8 characters correctly.
1700    
1701    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1702        character classes. PCRE was not treating the sequence [:...:] as a
1703        character class unless the ... were all letters. Perl, however, seems to
1704        allow any characters between [: and :], though of course it rejects as
1705        unknown any "names" that contain non-letters, because all the known class
1706        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1707        for example, whereas PCRE did not - it did not recognize a POSIX character
1708        class. This seemed a bit dangerous, so the code has been changed to be
1709        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1710        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1711        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1712        Perl does, and where it didn't before.
1713    
1714    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1715        Windows environments %n is disabled by default.
1716    
1717    
1718    Version 7.4 21-Sep-07
1719    ---------------------
1720    
1721    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1722        means that a class such as [\s] counted as "explicit reference to CR or
1723        LF". That isn't really right - the whole point of the change was to try to
1724        help when there was an actual mention of one of the two characters. So now
1725        the change happens only if \r or \n (or a literal CR or LF) character is
1726        encountered.
1727    
1728    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1729        of both had grown to the point where there were only 3 bits left.
1730        Fortunately, there was spare space in the data structure, and so I have
1731        moved the internal flags into a new 16-bit field to free up more option
1732        bits.
1733    
1734    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1735        but did not set the internal JCHANGED flag - either of these is enough to
1736        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1737        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1738        start sets both bits.
1739    
1740    4.  Added options (at build time, compile time, exec time) to change \R from
1741        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1742    
1743    5.  doc/pcresyntax.html was missing from the distribution.
1744    
1745    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1746        compatibility, even though it is no longer used.
1747    
1748    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1749        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1750        windows.h file is present (where different names are used). [This was
1751        reversed later after testing - see 16 below.]
1752    
1753    8.  Changed all #include <config.h> to #include "config.h". There were also
1754        some further <pcre.h> cases that I changed to "pcre.h".
1755    
1756    9.  When pcregrep was used with the --colour option, it missed the line ending
1757        sequence off the lines that it output.
1758    
1759    10. It was pointed out to me that arrays of string pointers cause lots of
1760        relocations when a shared library is dynamically loaded. A technique of
1761        using a single long string with a table of offsets can drastically reduce
1762        these. I have refactored PCRE in four places to do this. The result is
1763        dramatic:
1764    
1765          Originally:                          290
1766          After changing UCP table:            187
1767          After changing error message table:   43
1768          After changing table of "verbs"       36
1769          After changing table of Posix names   22
1770    
1771        Thanks to the folks working on Gregex for glib for this insight.
1772    
1773    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1774        unicode-properties was also set.
1775    
1776    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1777    
1778    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1779        checked only for CRLF.
1780    
1781    14. Added casts to pcretest.c to avoid compiler warnings.
1782    
1783    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1784    
1785    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1786        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1787        entirely. This removes changes made in 7 above.
1788    
1789    17. The CMake files have been updated, and there is now more information about
1790        building with CMake in the NON-UNIX-USE document.
1791    
1792    
1793    Version 7.3 28-Aug-07
1794    ---------------------
1795    
1796     1. In the rejigging of the build system that eventually resulted in 7.1, the
1797        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1798        brackets there is not right, since it causes compilers to look for an
1799        installed pcre.h, not the version that is in the source that is being
1800        compiled (which of course may be different). I have changed it back to:
1801    
1802          #include "pcre.h"
1803    
1804        I have a vague recollection that the change was concerned with compiling in
1805        different directories, but in the new build system, that is taken care of
1806        by the VPATH setting the Makefile.
1807    
1808     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1809        when the subject happened to end in the byte 0x85 (e.g. if the last
1810        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1811        characters but of course it shouldn't be taken as a newline when it is part
1812        of another character. The bug was that, for an unlimited repeat of . in
1813        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1814        characters when looking for a newline.
1815    
1816     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1817    
1818     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1819        in debug output.
1820    
1821     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1822        long printing in the pcrecpp unittest when running under MinGW.
1823    
1824     6. ESC_K was left out of the EBCDIC table.
1825    
1826     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1827        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1828        limit also applies to "virtual nesting" when a pattern is recursive, and in
1829        this case 1000 isn't so big. I have been able to remove this limit at the
1830        expense of backing off one optimization in certain circumstances. Normally,
1831        when pcre_exec() would call its internal match() function recursively and
1832        immediately return the result unconditionally, it uses a "tail recursion"
1833        feature to save stack. However, when a subpattern that can match an empty
1834        string has an unlimited repetition quantifier, it no longer makes this
1835        optimization. That gives it a stack frame in which to save the data for
1836        checking that an empty string has been matched. Previously this was taken
1837        from the 1000-entry workspace that had been reserved. So now there is no
1838        explicit limit, but more stack is used.
1839    
1840     8. Applied Daniel's patches to solve problems with the import/export magic
1841        syntax that is required for Windows, and which was going wrong for the
1842        pcreposix and pcrecpp parts of the library. These were overlooked when this
1843        problem was solved for the main library.
1844    
1845     9. There were some crude static tests to avoid integer overflow when computing
1846        the size of patterns that contain repeated groups with explicit upper
1847        limits. As the maximum quantifier is 65535, the maximum group length was
1848        set at 30,000 so that the product of these two numbers did not overflow a
1849        32-bit integer. However, it turns out that people want to use groups that
1850        are longer than 30,000 bytes (though not repeat them that many times).
1851        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1852        made it possible to implement the integer overflow checks in a much more
1853        dynamic way, which I have now done. The artificial limitation on group
1854        length has been removed - we now have only the limit on the total length of
1855        the compiled pattern, which depends on the LINK_SIZE setting.
1856    
1857    10. Fixed a bug in the documentation for get/copy named substring when
1858        duplicate names are permitted. If none of the named substrings are set, the
1859        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1860        empty string.
1861    
1862    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1863        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1864        because the ] is interpreted as the first data character and the
1865        terminating ] is not found. PCRE has been made compatible with Perl in this
1866        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1867        cause memory overwriting.
1868    
1869    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1870        string has been matched (to stop an infinite loop). It was not recognizing
1871        a conditional subpattern that could match an empty string if that
1872        subpattern was within another subpattern. For example, it looped when
1873        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1874        condition was not nested. This bug has been fixed.
1875    
1876    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1877        past the start of the subject in the presence of bytes with the top bit
1878        set, for example "\x8aBCD".
1879    
1880    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1881        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1882    
1883    14. Optimized (?!) to (*FAIL).
1884    
1885    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1886        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1887        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1888        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1889        does: it's just the validity check that is more restrictive.
1890    
1891    16. Inserted checks for integer overflows during escape sequence (backslash)
1892        processing, and also fixed erroneous offset values for syntax errors during
1893        backslash processing.
1894    
1895    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1896        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1897    
1898    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1899        caused an overrun.
1900    
1901    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1902        something other than just ASCII characters) inside a group that had an
1903        unlimited repeat caused a loop at compile time (while checking to see
1904        whether the group could match an empty string).
1905    
1906    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1907        [\P{Any}] did so. (Error in the code for printing property names.)
1908    
1909    21. An orphan \E inside a character class could cause a crash.
1910    
1911    22. A repeated capturing bracket such as (A)? could cause a wild memory
1912        reference during compilation.
1913    
1914    23. There are several functions in pcre_compile() that scan along a compiled
1915        expression for various reasons (e.g. to see if it's fixed length for look
1916        behind). There were bugs in these functions when a repeated \p or \P was
1917        present in the pattern. These operators have additional parameters compared
1918        with \d, etc, and these were not being taken into account when moving along
1919        the compiled data. Specifically:
1920    
1921        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1922            length.
1923    
1924        (b) An item such as \pL+ within a repeated group could cause crashes or
1925            loops.
1926    
1927        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1928            "reference to non-existent subpattern" error.
1929    
1930        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1931    
1932    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1933        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1934    
1935    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1936    
1937    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1938        character were causing crashes (broken optimization).
1939    
1940    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1941        \p or \P) caused a compile-time loop.
1942    
1943    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1944        break. For example, the unstudied pattern [\r\n]A does not match the string
1945        "\r\nA" because change 7.0/46 below moves the current point on by two
1946        characters after failing to match at the start. However, the pattern \nA
1947        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1948        the same is true. There doesn't seem any very clean way out of this, but
1949        what I have chosen to do makes the common cases work: PCRE now takes note
1950        of whether there can be an explicit match for \r or \n anywhere in the
1951        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1952        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1953        pattern has explicit CR or LF references.
1954    
1955    29. Added (*CR) etc for changing newline setting at start of pattern.
1956    
1957    
1958    Version 7.2 19-Jun-07
1959    ---------------------
1960    
1961     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1962        which is apparently normally available under Windows.
1963    
1964     2. Re-jig the pcregrep tests with different newline settings in an attempt
1965        to make them independent of the local environment's newline setting.
1966    
1967     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1968    
1969     4. Some of the "internals" tests were previously cut out when the link size
1970        was not 2, because the output contained actual offsets. The recent new
1971        "Z" feature of pcretest means that these can be cut out, making the tests
1972        usable with all link sizes.
1973    
1974     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1975        stack recursion. This gives a massive performance boost under BSD, but just
1976        a small improvement under Linux. However, it saves one field in the frame
1977        in all cases.
1978    
1979     6. Added more features from the forthcoming Perl 5.10:
1980    
1981        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1982            recursion call. It refers to the nth most recently opened parentheses.
1983    
1984        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1985            to be opened parentheses.
1986    
1987        (c) Conditions that refer to capturing parentheses can be specified
1988            relatively, for example, (?(-2)... or (?(+3)...
1989    
1990        (d) \K resets the start of the current match so that everything before
1991            is not part of it.
1992    
1993        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1994    
1995        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1996            reference syntax.
1997    
1998        (g) (?| introduces a group in which the numbering of parentheses in each
1999            alternative starts with the same number.
2000    
2001        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2002    
2003     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2004        PCRE_INFO_JCHANGED.
2005    
2006     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2007        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2008        for detecting groups that can match an empty string.
2009    
2010     9. A pattern with a very large number of alternatives (more than several
2011        hundred) was running out of internal workspace during the pre-compile
2012        phase, where pcre_compile() figures out how much memory will be needed. A
2013        bit of new cunning has reduced the workspace needed for groups with
2014        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2015        workspace instead of running out of the 4096 that are available.
2016    
2017    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2018    
2019    11. Applied patch from Google to remove an optimization that didn't quite work.
2020        The report of the bug said:
2021    
2022          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2023          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2024          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2025    
2026    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2027        it matched the wrong number of bytes.
2028    
2029    
2030    Version 7.1 24-Apr-07
2031    ---------------------
2032    
2033     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2034        that is more "standard", making use of automake and other Autotools. There
2035        is some re-arrangement of the files and adjustment of comments consequent
2036        on this.
2037    
2038     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2039        for recursive directory scanning broke on some systems because the files
2040        are not scanned in any specific order and on different systems the order
2041        was different. A call to "sort" has been inserted into RunGrepTest for the
2042        approprate test as a short-term fix. In the longer term there may be an
2043        alternative.
2044    
2045     3. I had an email from Eric Raymond about problems translating some of PCRE's
2046        man pages to HTML (despite the fact that I distribute HTML pages, some
2047        people do their own conversions for various reasons). The problems
2048        concerned the use of low-level troff macros .br and .in. I have therefore
2049        removed all such uses from the man pages (some were redundant, some could
2050        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2051        HTML has been updated to handle .nf/.fi and to complain if it encounters
2052        .br or .in.
2053    
2054     4. Updated comments in configure.ac that get placed in config.h.in and also
2055        arranged for config.h to be included in the distribution, with the name
2056        config.h.generic, for the benefit of those who have to compile without
2057        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2058    
2059     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2060        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2061        makevp.bat for the current PCRE, using the additional files
2062        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2063    
2064     6. A Windows user reported a minor discrepancy with test 2, which turned out
2065        to be caused by a trailing space on an input line that had got lost in his
2066        copy. The trailing space was an accident, so I've just removed it.
2067    
2068     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2069        that is needed.
2070    
2071     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2072        as "const" (a) because they are and (b) because it helps the PHP
2073        maintainers who have recently made a script to detect big data structures
2074        in the php code that should be moved to the .rodata section. I remembered
2075        to update Builducptable as well, so it won't revert if ucptable.h is ever
2076        re-created.
2077    
2078     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2079        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2080        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2081        support is not required. This saves 1.5-2K of code, which is important in
2082        some applications.
2083    
2084        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2085        so as not to refer to the tables, even though these functions will never be
2086        called when UTF-8 support is disabled. Otherwise there are problems with a
2087        shared library.
2088    
2089    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2090    
2091        (a) It was defining its arguments as char * instead of void *.
2092    
2093        (b) It was assuming that all moves were upwards in memory; this was true
2094            a long time ago when I wrote it, but is no longer the case.
2095    
2096        The emulated memove() is provided for those environments that have neither
2097        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2098        is clearly not the case, as these two bugs were recently reported.
2099    
2100    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2101        and Detrail to create the HTML documentation, the .txt form of the man
2102        pages, and it removes trailing spaces from listed files. It also creates
2103        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2104        case, it wraps all the #defines with #ifndefs. This script should be run
2105        before "make dist".
2106    
2107    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2108        with Unicode property support.
2109    
2110        (a) For a maximizing quantifier, if the two different cases of the
2111            character were of different lengths in their UTF-8 codings (there are
2112            some cases like this - I found 11), and the matching function had to
2113            back up over a mixture of the two cases, it incorrectly assumed they
2114            were both the same length.
2115    
2116        (b) When PCRE was configured to use the heap rather than the stack for
2117            recursion during matching, it was not correctly preserving the data for
2118            the other case of a UTF-8 character when checking ahead for a match
2119            while processing a minimizing repeat. If the check also involved
2120            matching a wide character, but failed, corruption could cause an
2121            erroneous result when trying to check for a repeat of the original
2122            character.
2123    
2124    13. Some tidying changes to the testing mechanism:
2125    
2126        (a) The RunTest script now detects the internal link size and whether there
2127            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2128            values substituted by "configure". (The RunGrepTest script already did
2129            this for UTF-8.) The configure.ac script no longer substitutes the
2130            relevant variables.
2131    
2132        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2133            with length and offset values. This means that the output is different
2134            for different internal link sizes. Test 2 is skipped for link sizes
2135            other than 2 because of this, bypassing the problem. Unfortunately,
2136            there was also a test in test 3 (the locale tests) that used /B and
2137            failed for link sizes other than 2. Rather than cut the whole test out,
2138            I have added a new /Z option to pcretest that replaces the length and
2139            offset values with spaces. This is now used to make test 3 independent
2140            of link size. (Test 2 will be tidied up later.)
2141    
2142    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2143        segmentation fault instead of returning the appropriate error message.
2144    
2145    15. In multiline mode when the newline sequence was set to "any", the pattern
2146        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2147        This doesn't seem right; it now treats the CRLF combination as the line
2148        ending, and so does not match in that case. It's only a pattern such as ^$
2149        that would hit this one: something like ^ABC$ would have failed after \r
2150        and then tried again after \r\n.
2151    
2152    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2153        in an attempt to make files that differ only in their line terminators
2154        compare equal. This works on Linux.
2155    
2156    17. Under certain error circumstances pcregrep might try to free random memory
2157        as it exited. This is now fixed, thanks to valgrind.
2158    
2159    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2160        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2161        was because its rules for how to advance for /g after matching an empty
2162        string at the end of a line did not allow for this case. They now check for
2163        it specially.
2164    
2165    20. pcretest is supposed to handle patterns and data of any length, by
2166        extending its buffers when necessary. It was getting this wrong when the
2167        buffer for a data line had to be extended.
2168    
2169    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2170        CRLF as a newline sequence.
2171    
2172    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2173        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2174        I have nevertheless tidied it up.
2175    
2176    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2177    
2178    24. Added a man page for pcre-config.
2179    
2180    
2181    Version 7.0 19-Dec-06
2182    ---------------------
2183    
2184     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2185        moving to gcc 4.1.1.
2186    
2187     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2188        sys/time.h, which is documented as needed for this function. It doesn't
2189        seem to matter on Linux, but it showed up on some releases of OS X.
2190    
2191     3. It seems that there are systems where bytes whose values are greater than
2192        127 match isprint() in the "C" locale. The "C" locale should be the
2193        default when a C program starts up. In most systems, only ASCII printing
2194        characters match isprint(). This difference caused the output from pcretest
2195        to vary, making some of the tests fail. I have changed pcretest so that:
2196    
2197        (a) When it is outputting text in the compiled version of a pattern, bytes
2198            other than 32-126 are always shown as hex escapes.
2199    
2200        (b) When it is outputting text that is a matched part of a subject string,
2201            it does the same, unless a different locale has been set for the match
2202            (using the /L modifier). In this case, it uses isprint() to decide.
2203    
2204     4. Fixed a major bug that caused incorrect computation of the amount of memory
2205        required for a compiled pattern when options that changed within the
2206        pattern affected the logic of the preliminary scan that determines the
2207        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2208        that the computed length was too small. The symptoms of this bug were
2209        either the PCRE error "internal error: code overflow" from pcre_compile(),
2210        or a glibc crash with a message such as "pcretest: free(): invalid next
2211        size (fast)". Examples of patterns that provoked this bug (shown in
2212        pcretest format) are:
2213    
2214          /(?-x: )/x
2215          /(?x)(?-x: \s*#\s*)/
2216          /((?i)[\x{c0}])/8
2217          /(?i:[\x{c0}])/8
2218    
2219        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2220        is now done differently.
2221    
2222     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2223        wrapper classes; (b) implement a new function in the C++ scanner that is
2224        more efficient than the old way of doing things because it avoids levels of
2225        recursion in the regex matching; (c) add a paragraph to the documentation
2226        for the FullMatch() function.
2227    
2228     6. The escape sequence \n was being treated as whatever was defined as
2229        "newline". Not only was this contrary to the documentation, which states
2230        that \n is character 10 (hex 0A), but it also went horribly wrong when
2231        "newline" was defined as CRLF. This has been fixed.
2232    
2233     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2234        was being set to -1 for the "end of line" case (supposedly a value that no
2235        character can have). Though this value is never used (the check for end of
2236        line is "zero bytes in current character"), it caused compiler complaints.
2237        I've changed it to 0xffffffff.
2238    
2239     8. In pcre_version.c, the version string was being built by a sequence of
2240        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2241        string (as it is for production releases) called a macro with an empty
2242        argument. The C standard says the result of this is undefined. The gcc
2243        compiler treats it as an empty string (which was what was wanted) but it is
2244        reported that Visual C gives an error. The source has been hacked around to
2245        avoid this problem.
2246    
2247     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2248        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2249        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2250        of them did).
2251    
2252    10. Originally, pcretest opened its input and output without "b"; then I was
2253        told that "b" was needed in some environments, so it was added for release
2254        5.0 to both the input and output. (It makes no difference on Unix-like
2255        systems.) Later I was told that it is wrong for the input on Windows. I've
2256        now abstracted the modes into two macros, to make it easier to fiddle with
2257        them, and removed "b" from the input mode under Windows.
2258    
2259    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2260    
2261    12. Added -help and --help to pcretest as an official way of being reminded
2262        of the options.
2263    
2264    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2265        and pcrecpp.cc because they annoy compilers at high warning levels.
2266    
2267    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2268    
2269    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2270        scripts are not C programs :-) and which was not noticed because it works
2271        on Linux.
2272    
2273    16. pcretest is supposed to handle any length of pattern and data line (as one
2274        line or as a continued sequence of lines) by extending its input buffer if
2275        necessary. This feature was broken for very long pattern lines, leading to
2276        a string of junk being passed to pcre_compile() if the pattern was longer
2277        than about 50K.
2278    
2279    17. I have done a major re-factoring of the way pcre_compile() computes the
2280        amount of memory needed for a compiled pattern. Previously, there was code
2281        that made a preliminary scan of the pattern in order to do this. That was
2282        OK when PCRE was new, but as the facilities have expanded, it has become
2283        harder and harder to keep it in step with the real compile phase, and there
2284        have been a number of bugs (see for example, 4 above). I have now found a
2285        cunning way of running the real compile function in a "fake" mode that
2286        enables it to compute how much memory it would need, while actually only
2287        ever using a few hundred bytes of working memory and without too many
2288        tests of the mode. This should make future maintenance and development
2289        easier. A side effect of this work is that the limit of 200 on the nesting
2290        depth of parentheses has been removed (though this was never a serious
2291        limitation, I suspect). However, there is a downside: pcre_compile() now
2292        runs more slowly than before (30% or more, depending on the pattern). I
2293        hope this isn't a big issue. There is no effect on runtime performance.
2294    
2295    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2296        newline (only possible for the last line of a file) and it was a
2297        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2298    
2299    19. Added additional timing features to pcretest. (1) The -tm option now times
2300        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2301        separate command line item, by a number that specifies the number of
2302        repeats to use when timing. The default is 50000; this gives better
2303        precision, but takes uncomfortably long for very large patterns.
2304    
2305    20. Extended pcre_study() to be more clever in cases where a branch of a
2306        subpattern has no definite first character. For example, (a*|b*)[cd] would
2307        previously give no result from pcre_study(). Now it recognizes that the
2308        first character must be a, b, c, or d.
2309    
2310    21. There was an incorrect error "recursive call could loop indefinitely" if
2311        a subpattern (or the entire pattern) that was being tested for matching an
2312        empty string contained only one non-empty item after a nested subpattern.
2313        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2314        incorrectly, because the \d was being skipped in the check.
2315    
2316    22. The pcretest program now has a new pattern option /B and a command line
2317        option -b, which is equivalent to adding /B to every pattern. This causes
2318        it to show the compiled bytecode, without the additional information that
2319        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2320        is the same as /B/I).
2321    
2322    23. A new optimization is now able automatically to treat some sequences such
2323        as a*b as a*+b. More specifically, if something simple (such as a character
2324        or a simple class like \d) has an unlimited quantifier, and is followed by
2325        something that cannot possibly match the quantified thing, the quantifier
2326        is automatically "possessified".
2327    
2328    24. A recursive reference to a subpattern whose number was greater than 39
2329        went wrong under certain circumstances in UTF-8 mode. This bug could also
2330        have affected the operation of pcre_study().
2331    
2332    25. Realized that a little bit of performance could be had by replacing
2333        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2334    
2335    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2336    
2337    27. Possessive quantifiers such as a++ were previously implemented by turning
2338        them into atomic groups such as ($>a+). Now they have their own opcodes,
2339        which improves performance. This includes the automatically created ones
2340        from 23 above.
2341    
2342    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2343        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2344        the first matched character to be a colon. This applied both to named and
2345        numbered groups.
2346    
2347    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2348    
2349    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2350        building PCRE on Windows easier, so I have included it in the distribution.
2351    
2352    31. There is now a check in pcretest against a ridiculously large number being
2353        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2354        loop, the loop is abandoned.
2355    
2356    32. Forward references to subpatterns in conditions such as (?(2)...) where
2357        subpattern 2 is defined later cause pcre_compile() to search forwards in
2358        the pattern for the relevant set of parentheses. This search went wrong
2359        when there were unescaped parentheses in a character class, parentheses
2360        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2361    
2362    33. "Subroutine" calls and backreferences were previously restricted to
2363        referencing subpatterns earlier in the regex. This restriction has now
2364        been removed.
2365    
2366    34. Added a number of extra features that are going to be in Perl 5.10. On the
2367        whole, these are just syntactic alternatives for features that PCRE had
2368        previously implemented using the Python syntax or my own invention. The
2369        other formats are all retained for compatibility.
2370    
2371        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2372            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2373            also .NET compatible.
2374    
2375        (b) A recursion or subroutine call to a named group can now be defined as
2376            (?&name) as well as (?P>name).
2377    
2378        (c) A backreference to a named group can now be defined as \k<name> or
2379            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2380            5.10, are also .NET compatible.
2381    
2382        (d) A conditional reference to a named group can now use the syntax
2383            (?(<name>) or (?('name') as well as (?(name).
2384    
2385        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2386            groups (named and numbered) that are never evaluated inline, but can be
2387            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2388            is always false. There may be only one alternative in such a group.
2389    
2390        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2391            as the simple (?(R). The condition is true only if the most recent
2392            recursion is that of the given number or name. It does not search out
2393            through the entire recursion stack.
2394    
2395        (g) The escape \gN or \g{N} has been added, where N is a positive or
2396            negative number, specifying an absolute or relative reference.
2397    
2398    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2399        some "unreachable code" warnings.
2400    
2401    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2402        things, this adds five new scripts.
2403    
2404    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2405        There were also incompatibilities regarding the handling of \Q..\E inside
2406        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2407        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2408    
2409    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2410        matches an empty string, and forcibly breaks the loop. There were bugs in
2411        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2412        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2413        separate and independent bugs (that affected different cases) have been
2414        fixed.
2415    
2416    39. Refactored the code to abolish the use of different opcodes for small
2417        capturing bracket numbers. This is a tidy that I avoided doing when I
2418        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2419        The new approach is not only tidier, it makes it possible to reduce the
2420        memory needed to fix the previous bug (38).
2421    
2422    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2423        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2424        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2425        mode.
2426    
2427    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2428        report.
2429    
2430    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2431        copy construction and assignment in the C++ wrapper.
2432    
2433    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2434        couple of bugs that could have given wrong results in the "--newline=crlf"
2435        case.
2436    
2437    44. Added a number of casts and did some reorganization of signed/unsigned int
2438        variables following suggestions from Dair Grant. Also renamed the variable
2439        "this" as "item" because it is a C++ keyword.
2440    
2441    45. Arranged for dftables to add
2442    
2443          #include "pcre_internal.h"
2444    
2445        to pcre_chartables.c because without it, gcc 4.x may remove the array
2446        definition from the final binary if PCRE is built into a static library and
2447        dead code stripping is activated.
2448    
2449    46. For an unanchored pattern, if a match attempt fails at the start of a
2450        newline sequence, and the newline setting is CRLF or ANY, and the next two
2451        characters are CRLF, advance by two characters instead of one.
2452    
2453    
2454    Version 6.7 04-Jul-06
2455    ---------------------
2456    
2457     1. In order to handle tests when input lines are enormously long, pcretest has
2458        been re-factored so that it automatically extends its buffers when
2459        necessary. The code is crude, but this _is_ just a test program. The
2460        default size has been increased from 32K to 50K.
2461    
2462     2. The code in pcre_study() was using the value of the re argument before
2463        testing it for NULL. (Of course, in any sensible call of the function, it
2464        won't be NULL.)
2465    
2466     3. The memmove() emulation function in pcre_internal.h, which is used on
2467        systems that lack both memmove() and bcopy() - that is, hardly ever -
2468        was missing a "static" storage class specifier.
2469    
2470     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2471        containing an extended class (one that cannot be represented by a bitmap
2472        because it contains high-valued characters or Unicode property items, e.g.
2473        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2474        pattern, but PCRE should not loop if you do not (it no longer does).
2475        [Detail: two cases were found: (a) a repeated subpattern containing an
2476        extended class; (b) a recursive reference to a subpattern that followed a
2477        previous extended class. It wasn't skipping over the extended class
2478        correctly when UTF-8 mode was not set.]
2479    
2480     5. A negated single-character class was not being recognized as fixed-length
2481        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2482        compile error "lookbehind assertion is not fixed length".
2483    
2484     6. The RunPerlTest auxiliary script was showing an unexpected difference
2485        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2486        write a Perl script that can interpret lines of an input file either as
2487        byte characters or as UTF-8, which is what "perltest" was being required to
2488        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2489        can't do is switch easily at run time between having the "use utf8;" pragma
2490        or not. In the end, I fudged it by using the RunPerlTest script to insert
2491        "use utf8;" explicitly for the UTF-8 tests.
2492    
2493     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2494        the end of the subject string, contrary to the documentation and to what
2495        Perl does. This was true of both matching functions. Now it matches only at
2496        the start of the subject and immediately after *internal* newlines.
2497    
2498     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2499        a pointer to an int instead of a pointer to an unsigned long int. This
2500        caused problems on 64-bit systems.
2501    
2502     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2503        instance of the 'standard' template library not being so standard".
2504    
2505    10. There was no check on the number of named subpatterns nor the maximum
2506        length of a subpattern name. The product of these values is used to compute
2507        the size of the memory block for a compiled pattern. By supplying a very
2508        long subpattern name and a large number of named subpatterns, the size
2509        computation could be caused to overflow. This is now prevented by limiting
2510        the length of names to 32 characters, and the number of named subpatterns
2511        to 10,000.
2512    
2513    11. Subpatterns that are repeated with specific counts have to be replicated in
2514        the compiled pattern. The size of memory for this was computed from the
2515        length of the subpattern and the repeat count. The latter is limited to
2516        65535, but there was no limit on the former, meaning that integer overflow
2517        could in principle occur. The compiled length of a repeated subpattern is
2518        now limited to 30,000 bytes in order to prevent this.
2519    
2520    12. Added the optional facility to have named substrings with the same name.
2521    
2522    13. Added the ability to use a named substring as a condition, using the
2523        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2524        are numbers (not recommended). Forward references are permitted.
2525    
2526    14. Added forward references in named backreferences (if you see what I mean).
2527    
2528    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2529        pattern could run off the end of the subject. For example, the pattern
2530        "(?s)(.{1,5})"8 did this with the subject "ab".
2531    
2532    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2533        PCRE_CASELESS was set when matching characters that were quantified with ?
2534        or *.
2535    
2536    17. A character class other than a single negated character that had a minimum
2537        but no maximum quantifier - for example [ab]{6,} - was not handled
2538        correctly by pce_dfa_exec(). It would match only one character.
2539    
2540    18. A valid (though odd) pattern that looked like a POSIX character
2541        class but used an invalid character after [ (for example [[,abc,]]) caused
2542        pcre_compile() to give the error "Failed: internal error: code overflow" or
2543        in some cases to crash with a glibc free() error. This could even happen if
2544        the pattern terminated after [[ but there just happened to be a sequence of
2545        letters, a binary zero, and a closing ] in the memory that followed.
2546    
2547    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2548        over the years. Originally (before any Unicode support), just the bottom 8
2549        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2550        output from "man perlunicode" includes this:
2551    
2552          The regular expression compiler produces polymorphic opcodes.  That
2553          is, the pattern adapts to the data and automatically switches to
2554          the Unicode character scheme when presented with Unicode data--or
2555          instead uses a traditional byte scheme when presented with byte
2556          data.
2557    
2558        Sadly, a wide octal escape does not cause a switch, and in a string with
2559        no other multibyte characters, these octal escapes are treated as before.
2560        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2561        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2562        Unicode string.
2563    
2564        I have not perpetrated such confusion in PCRE. Up till now, it took just
2565        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2566        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2567        translate to the appropriate multibyte character.
2568    
2569    29. Applied some refactoring to reduce the number of warnings from Microsoft
2570        and Borland compilers. This has included removing the fudge introduced
2571        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2572        a warning about an unused variable.
2573    
2574    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2575        characters since release 4.0, because Perl (from release 5.004) does not.
2576        [Or at least, is documented not to: some releases seem to be in conflict
2577        with the documentation.] However, when a pattern was studied with
2578        pcre_study() and all its branches started with \s, PCRE still included VT
2579        as a possible starting character. Of course, this did no harm; it just
2580        caused an unnecessary match attempt.
2581    
2582    22. Removed a now-redundant internal flag bit that recorded the fact that case
2583        dependency changed within the pattern. This was once needed for "required
2584        byte" processing, but is no longer used. This recovers a now-scarce options
2585        bit. Also moved the least significant internal flag bit to the most-
2586        significant bit of the word, which was not previously used (hangover from
2587        the days when it was an int rather than a uint) to free up another bit for
2588        the future.
2589    
2590    23. Added support for CRLF line endings as well as CR and LF. As well as the
2591        default being selectable at build time, it can now be changed at runtime
2592        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2593        specify that it is scanning data with non-default line endings.
2594    
2595    24. Changed the definition of CXXLINK to make it agree with the definition of
2596        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2597    
2598    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2599        recursions. This makes a big different to stack usage for some patterns.
2600    
2601    26. If a subpattern containing a named recursion or subroutine reference such
2602        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2603        the space required for the compiled pattern went wrong and gave too small a
2604        value. Depending on the environment, this could lead to "Failed: internal
2605        error: code overflow at offset 49" or "glibc detected double free or
2606        corruption" errors.
2607    
2608    27. Applied patches from Google (a) to support the new newline modes and (b) to
2609        advance over multibyte UTF-8 characters in GlobalReplace.
2610    
2611    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2612        difference for some implementation of PCRE in some Windows version.
2613    
2614    29. Added some extra testing facilities to pcretest:
2615    
2616        \q<number>   in a data line sets the "match limit" value
2617        \Q<number>   in a data line sets the "match recursion limt" value
2618        -S <number>  sets the stack size, where <number> is in megabytes
2619    
2620        The -S option isn't available for Windows.
2621    
2622    
2623    Version 6.6 06-Feb-06
2624    ---------------------
2625    
2626     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2627        in pcreposix.h. I have copied the definition from pcre.h.
2628    
2629     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2630        because pcre.h is no longer a built file.
2631    
2632     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2633        not normally included in the compiled code.
2634    
2635    
2636    Version 6.5 01-Feb-06
2637    ---------------------
2638    
2639     1. When using the partial match feature with pcre_dfa_exec(), it was not
2640        anchoring the second and subsequent partial matches at the new starting
2641        point. This could lead to incorrect results. For example, with the pattern
2642        /1234/, partially matching against "123" and then "a4" gave a match.
2643    
2644     2. Changes to pcregrep:
2645    
2646        (a) All non-match returns from pcre_exec() were being treated as failures
2647            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2648            error message is output. Some extra information is given for the
2649            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2650            probably the only errors that are likely to be caused by users (by
2651            specifying a regex that has nested indefinite repeats, for instance).
2652            If there are more than 20 of these errors, pcregrep is abandoned.
2653    
2654        (b) A binary zero was treated as data while matching, but terminated the
2655            output line if it was written out. This has been fixed: binary zeroes
2656            are now no different to any other data bytes.
2657    
2658        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2659            used to set a locale for matching. The --locale=xxxx long option has
2660            been added (no short equivalent) to specify a locale explicitly on the
2661            pcregrep command, overriding the environment variables.
2662    
2663        (d) When -B was used with -n, some line numbers in the output were one less
2664            than they should have been.
2665    
2666        (e) Added the -o (--only-matching) option.
2667    
2668        (f) If -A or -C was used with -c (count only), some lines of context were
2669            accidentally printed for the final match.
2670    
2671        (g) Added the -H (--with-filename) option.
2672    
2673        (h) The combination of options -rh failed to suppress file names for files
2674            that were found from directory arguments.
2675    
2676        (i) Added the -D (--devices) and -d (--directories) options.
2677    
2678        (j) Added the -F (--fixed-strings) option.
2679    
2680        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2681    
2682        (l) Added the --colo(u)r option.
2683    
2684        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2685            is not present by default.
2686    
2687     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2688        items such as (?R) or (?1), when the recursion could match a number of
2689        alternatives. If it matched one of the alternatives, but subsequently,
2690        outside the recursion, there was a failure, the code tried to back up into
2691        the recursion. However, because of the way PCRE is implemented, this is not
2692        possible, and the result was an incorrect result from the match.
2693    
2694        In order to prevent this happening, the specification of recursion has
2695        been changed so that all such subpatterns are automatically treated as
2696        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2697    
2698     4. I had overlooked the fact that, in some locales, there are characters for
2699        which isalpha() is true but neither isupper() nor islower() are true. In
2700        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2701        and ordfeminine) are like this. This affected the treatment of \w and \W
2702        when they appeared in character classes, but not when they appeared outside
2703        a character class. The bit map for "word" characters is now created
2704        separately from the results of isalnum() instead of just taking it from the
2705        upper, lower, and digit maps. (Plus the underscore character, of course.)
2706    
2707     5. The above bug also affected the handling of POSIX character classes such as
2708        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2709        permanent tables. Instead, the bit maps for such a class were previously
2710        created as the appropriate unions of the upper, lower, and digit bitmaps.
2711        Now they are created by subtraction from the [[:word:]] class, which has
2712        its own bitmap.
2713    
2714     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2715        It is created by subtracting the vertical space characters (\x09, \x0a,
2716        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2717        subtraction was done in the overall bitmap for a character class, meaning
2718        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2719        be recognized. This bug has been fixed.
2720    
2721     7. Patches from the folks at Google:
2722    
2723          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2724          real life, but is still worth protecting against".
2725    
2726          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2727          regular expressions".
2728    
2729          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2730          have it.
2731    
2732          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2733          "configure" and the latter not, in order to fix a problem somebody had
2734          with compiling the Arg class on HP-UX.
2735    
2736          (e) Improve the error-handling of the C++ wrapper a little bit.
2737    
2738          (f) New tests for checking recursion limiting.
2739    
2740     8. The pcre_memmove() function, which is used only if the environment does not
2741        have a standard memmove() function (and is therefore rarely compiled),
2742        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2743        returning a result (though PCRE never actually uses the result).
2744    
2745     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2746        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2747        returned instead of calling malloc() with an overflowing number that would
2748        most likely cause subsequent chaos.
2749    
2750    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2751    
2752    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2753        with this option is matched, the nmatch and pmatch options of regexec() are
2754        ignored.
2755    
2756    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2757        provided in case anyone wants to the the POSIX interface with UTF-8
2758        strings.
2759    
2760    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2761        C++ linking (needed for some HP-UX environments).
2762    
2763    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2764        (unused parameter) and in the pcre_printint() function (omitted "default"
2765        switch label when the default is to do nothing).
2766    
2767    15. Added some code to make it possible, when PCRE is compiled as a C++
2768        library, to replace subject pointers for pcre_exec() with a smart pointer
2769        class, thus making it possible to process discontinuous strings.
2770    
2771    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2772        much the same function. They were added by different people who were trying
2773        to make PCRE easy to compile on non-Unix systems. It has been suggested
2774        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2775        for compiling on Windows systems. I have therefore replaced it with
2776        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2777        defaults to "extern" for C or "extern C" for C++, which works fine on
2778        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2779        SCOPE with something explicit in config.h. In addition:
2780    
2781        (a) pcreposix.h still had just "extern" instead of either of these macros;
2782            I have replaced it with PCRE_DATA_SCOPE.
2783    
2784        (b) Functions such as _pcre_xclass(), which are internal to the library,
2785            but external in the C sense, all had PCRE_EXPORT in their definitions.
2786            This is apparently wrong for the Windows case, so I have removed it.
2787            (It makes no difference on Unix-like systems.)
2788    
2789    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2790        of recursive calls to match(). This is different to MATCH_LIMIT because
2791        that limits the total number of calls to match(), not all of which increase
2792        the depth of recursion. Limiting the recursion depth limits the amount of
2793        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2794        when PCRE is compiled, and changed at run time. A patch from Google adds
2795        this functionality to the C++ interface.
2796    
2797    18. Changes to the handling of Unicode character properties:
2798    
2799        (a) Updated the table to Unicode 4.1.0.
2800    
2801        (b) Recognize characters that are not in the table as "Cn" (undefined).
2802    
2803        (c) I revised the way the table is implemented to a much improved format
2804            which includes recognition of ranges. It now supports the ranges that
2805            are defined in UnicodeData.txt, and it also amalgamates other
2806            characters into ranges. This has reduced the number of entries in the
2807            table from around 16,000 to around 3,000, thus reducing its size
2808            considerably. I realized I did not need to use a tree structure after
2809            all - a binary chop search is just as efficient. Having reduced the
2810            number of entries, I extended their size from 6 bytes to 8 bytes to
2811            allow for more data.
2812    
2813        (d) Added support for Unicode script names via properties such as \p{Han}.
2814    
2815    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2816        matching that character.
2817    
2818    20. When matching a repeated Unicode property with a minimum greater than zero,
2819        (for example \pL{2,}), PCRE could look past the end of the subject if it
2820        reached it while seeking the minimum number of characters. This could
2821        happen only if some of the characters were more than one byte long, because
2822        there is a check for at least the minimum number of bytes.
2823    
2824    21. Refactored the implementation of \p and \P so as to be more general, to
2825        allow for more different types of property in future. This has changed the
2826        compiled form incompatibly. Anybody with saved compiled patterns that use
2827        \p or \P will have to recompile them.
2828    
2829    22. Added "Any" and "L&" to the supported property types.
2830    
2831    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2832        but give a compile time error if the value is greater than 0xff.
2833    
2834    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2835        accidentally not being installed or uninstalled.
2836    
2837    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2838        made were to insert the current release number. This seemed silly, because
2839        it made things harder for people building PCRE on systems that don't run
2840        "configure". I have turned pcre.h into a distributed file, no longer built
2841        by "configure", with the version identification directly included. There is
2842        no longer a pcre.h.in file.
2843    
2844        However, this change necessitated a change to the pcre-config script as
2845        well. It is built from pcre-config.in, and one of the substitutions was the
2846        release number. I have updated configure.ac so that ./configure now finds
2847        the release number by grepping pcre.h.
2848    
2849    26. Added the ability to run the tests under valgrind.
2850    
2851    
2852    Version 6.4 05-Sep-05
2853    ---------------------
2854    
2855     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2856        "--" to be printed when multiple files were scanned, even when none of the
2857        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2858        consider it to be a bug, and have restored the previous behaviour.
2859    
2860     2. A couple of code tidies to get rid of compiler warnings.
2861    
2862     3. The pcretest program used to cheat by referring to symbols in the library
2863        whose names begin with _pcre_. These are internal symbols that are not
2864        really supposed to be visible externally, and in some environments it is
2865        possible to suppress them. The cheating is now confined to including
2866        certain files from the library's source, which is a bit cleaner.
2867    
2868     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2869        file's purpose clearer.
2870    
2871     5. Reorganized pcre_ucp_findchar().
2872    
2873    
2874    Version 6.3 15-Aug-05
2875    ---------------------
2876    
2877     1. The file libpcre.pc.in did not have general read permission in the tarball.
2878    
2879     2. There were some problems when building without C++ support:
2880    
2881        (a) If C++ support was not built, "make install" and "make test" still
2882            tried to test it.
2883    
2884        (b) There were problems when the value of CXX was explicitly set. Some
2885            changes have been made to try to fix these, and ...
2886    
2887        (c) --disable-cpp can now be used to explicitly disable C++ support.
2888    
2889        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2890            backslash in a target when C++ was disabled. This confuses some
2891            versions of "make", apparently. Using an intermediate variable solves
2892            this. (Same for CPP_LOBJ.)
2893    
2894     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2895        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2896        necessary on certain architectures.
2897    
2898     4. Added a setting of -export-symbols-regex to the link command to remove
2899        those symbols that are exported in the C sense, but actually are local
2900        within the library, and not documented. Their names all begin with
2901        "_pcre_". This is not a perfect job, because (a) we have to except some
2902        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2903        available (and never for static libraries). I have made a note to try to
2904        find a way round (a) in the future.
2905    
2906    
2907    Version 6.2 01-Aug-05
2908    ---------------------
2909    
2910     1. There was no test for integer overflow of quantifier values. A construction
2911        such as {1111111111111111} would give undefined results. What is worse, if
2912        a minimum quantifier for a parenthesized subpattern overflowed and became
2913        negative, the calculation of the memory size went wrong. This could have
2914        led to memory overwriting.
2915    
2916     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2917    
2918     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2919        operating environments where this matters.
2920    
2921     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2922        PCRE options from within the C++ wrapper.
2923    
2924     5. Named capturing subpatterns were not being correctly counted when a pattern
2925        was compiled. This caused two problems: (a) If there were more than 100
2926        such subpatterns, the calculation of the memory needed for the whole
2927        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2928        back references of the form \12, where the number was greater than 9, were
2929        not recognized as back references, even though there were sufficient
2930        previous subpatterns.
2931    
2932     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2933        versions of gcc, e.g. 2.95.4.
2934    
2935    
2936    Version 6.1 21-Jun-05
2937    ---------------------
2938    
2939     1. There was one reference to the variable "posix" in pcretest.c that was not
2940        surrounded by "#if !defined NOPOSIX".
2941    
2942     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2943        the cross-check on the old pcre_info() function, for the benefit of the
2944        cut-down version of PCRE that is currently imported into Exim.
2945    
2946     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2947        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2948        patterns that start (?i) but I don't think that matters. The use of (?i) is
2949        just an example; this all applies to the other options as well.
2950    
2951     4. Since libtool seems to echo the compile commands it is issuing, the output
2952        from "make" can be reduced a bit by putting "@" in front of each libtool
2953        compile command.
2954    
2955     5. Patch from the folks at Google for configure.in to be a bit more thorough
2956        in checking for a suitable C++ installation before trying to compile the
2957        C++ stuff. This should fix a reported problem when a compiler was present,
2958        but no suitable headers.
2959    
2960     6. The man pages all had just "PCRE" as their title. I have changed them to
2961        be the relevant file name. I have also arranged that these names are
2962        retained in the file doc/pcre.txt, which is a concatenation in text format
2963        of all the man pages except the little individual ones for each function.
2964    
2965     7. The NON-UNIX-USE file had not been updated for the different set of source
2966        files that come with release 6. I also added a few comments about the C++
2967        wrapper.
2968    
2969    
2970    Version 6.0 07-Jun-05
2971    ---------------------
2972    
2973     1. Some minor internal re-organization to help with my DFA experiments.
2974    
2975     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2976        didn't matter for the library itself when fully configured, but did matter
2977        when compiling without UCP support, or within Exim, where the ucp files are
2978        not imported.
2979    
2980     3. Refactoring of the library code to split up the various functions into
2981        different source modules. The addition of the new DFA matching code (see
2982        below) to a single monolithic source would have made it really too
2983        unwieldy, quite apart from causing all the code to be include in a
2984        statically linked application, when only some functions are used. This is
2985        relevant even without the DFA addition now that patterns can be compiled in
2986        one application and matched in another.
2987    
2988        The downside of splitting up is that there have to be some external
2989        functions and data tables that are used internally in different modules of
2990        the library but which are not part of the API. These have all had their
2991        names changed to start with "_pcre_" so that they are unlikely to clash
2992        with other external names.
2993    
2994     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2995        a different (DFA) algorithm. Although it is slower than the original
2996        function, it does have some advantages for certain types of matching
2997        problem.
2998    
2999     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3000        including restarting after a partial match.
3001    
3002     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3003        defined when compiling for Windows was sent to me. I have put it into the
3004        code, though I have no means of testing or verifying it.
3005    
3006     7. Added the pcre_refcount() auxiliary function.
3007    
3008     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3009        match before or at the first newline in the subject string. In pcretest,
3010        the /f option on a pattern can be used to set this.
3011    
3012     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3013        would behave wrongly. This has been present in PCRE since release 4.0.
3014    
3015    10. A number of changes to the pcregrep command:
3016    
3017        (a) Refactored how -x works; insert ^(...)$ instead of setting
3018            PCRE_ANCHORED and checking the length, in preparation for adding
3019            something similar for -w.
3020    
3021        (b) Added the -w (match as a word) option.
3022    
3023        (c) Refactored the way lines are read and buffered so as to have more
3024            than one at a time available.
3025    
3026        (d) Implemented a pcregrep test script.
3027    
3028        (e) Added the -M (multiline match) option. This allows patterns to match
3029            over several lines of the subject. The buffering ensures that at least
3030            8K, or the rest of the document (whichever is the shorter) is available
3031            for matching (and similarly the previous 8K for lookbehind assertions).
3032    
3033        (f) Changed the --help output so that it now says
3034    
3035              -w, --word-regex(p)
3036    
3037            instead of two lines, one with "regex" and the other with "regexp"
3038            because that confused at least one person since the short forms are the
3039            same. (This required a bit of code, as the output is generated
3040            automatically from a table. It wasn't just a text change.)
3041    
3042        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3043            option but starts with a hyphen. Could be a pattern or a path name
3044            starting with a hyphen, for instance.
3045    
3046        (h) "-" can be given as a file name to represent stdin.
3047    
3048        (i) When file names are being printed, "(standard input)" is used for
3049            the standard input, for compatibility with GNU grep. Previously
3050            "<stdin>" was used.
3051    
3052        (j) The option --label=xxx can be used to supply a name to be used for
3053            stdin when file names are being printed. There is no short form.
3054    
3055        (k) Re-factored the options decoding logic because we are going to add
3056            two more options that take data. Such options can now be given in four
3057            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3058    
3059        (l) Added the -A, -B, and -C options for requesting that lines of context
3060            around matches be printed.
3061    
3062        (m) Added the -L option to print the names of files that do not contain
3063            any matching lines, that is, the complement of -l.
3064    
3065        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3066            continue to scan other files.
3067    
3068        (o) The -s option was incorrectly implemented. For compatibility with other
3069            greps, it now suppresses the error message for a non-existent or non-
3070            accessible file (but not the return code). There is a new option called
3071            -q that suppresses the output of matching lines, which was what -s was
3072            previously doing.
3073    
3074        (p) Added --include and --exclude options to specify files for inclusion
3075            and exclusion when recursing.
3076    
3077    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3078        Hopefully, it now does.
3079    
3080    12. Missing cast in pcre_study().
3081    
3082    13. Added an "uninstall" target to the makefile.
3083    
3084    14. Replaced "extern" in the function prototypes in Makefile.in with
3085        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3086        world, but is set differently for Windows.
3087    
3088    15. Added a second compiling function called pcre_compile2(). The only
3089        difference is that it has an extra argument, which is a pointer to an
3090        integer error code. When there is a compile-time failure, this is set
3091        non-zero, in addition to the error test pointer being set to point to an
3092        error message. The new argument may be NULL if no error number is required
3093        (but then you may as well call pcre_compile(), which is now just a
3094        wrapper). This facility is provided because some applications need a
3095        numeric error indication, but it has also enabled me to tidy up the way
3096        compile-time errors are handled in the POSIX wrapper.
3097    
3098    16. Added VPATH=.libs to the makefile; this should help when building with one
3099        prefix path and installing with another. (Or so I'm told by someone who
3100        knows more about this stuff than I do.)
3101    
3102    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3103        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3104        match everything, including newlines. This is not POSIX-compatible, but
3105        somebody wanted the feature. From pcretest it can be activated by using
3106        both the P and the s flags.
3107    
3108    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3109    
3110    19. libpcre.pc was being incorrectly installed as executable.
3111    
3112    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3113        it now also looks for '\r' so that it will work unmodified on Windows.
3114    
3115    21. Added Google's contributed C++ wrapper to the distribution.
3116    
3117    22. Added some untidy missing memory free() calls in pcretest, to keep
3118        Electric Fence happy when testing.
3119    
3120    
3121    
3122    Version 5.0 13-Sep-04
3123    ---------------------
3124    
3125     1. Internal change: literal characters are no longer packed up into items
3126        containing multiple characters in a single byte-string. Each character
3127        is now matched using a separate opcode. However, there may be more than one
3128        byte in the character in UTF-8 mode.
3129    
3130     2. The pcre_callout_block structure has two new fields: pattern_position and
3131        next_item_length. These contain the offset in the pattern to the next match
3132        item, and its length, respectively.
3133    
3134     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3135        insertion of callouts before each pattern item. Added the /C option to
3136        pcretest to make use of this.
3137    
3138     4. On the advice of a Windows user, the lines
3139    
3140          #if defined(_WIN32) || defined(WIN32)
3141          _setmode( _fileno( stdout ), 0x8000 );
3142          #endif  /* defined(_WIN32) || defined(WIN32) */
3143    
3144        have been added to the source of pcretest. This apparently does useful
3145        magic in relation to line terminators.
3146    
3147     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3148        for the benefit of those environments where the "b" makes a difference.
3149    
3150     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3151        to know about it. I have put a hack into configure.in that adds in code
3152        to set GCC=yes if CC=icc. This seems to end up at a point in the
3153        generated configure script that is early enough to affect the setting of
3154        compiler options, which is what is needed, but I have no means of testing
3155        whether it really works. (The user who reported this had patched the
3156        generated configure script, which of course I cannot do.)
3157    
3158        LATER: After change 22 below (new libtool files), the configure script
3159        seems to know about icc (and also ecc). Therefore, I have commented out
3160        this hack in configure.in.
3161    
3162     7. Added support for pkg-config (2 patches were sent in).
3163    
3164     8. Negated POSIX character classes that used a combination of internal tables
3165        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3166        [[:^ascii]]. Typically, they would match almost any characters. The other
3167        POSIX classes were not broken in this way.
3168    
3169     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3170        to find the match, as PCRE was deluded into thinking that the match had to
3171        start at the start point or following a newline. The same bug applied to
3172        patterns with negative forward assertions or any backward assertions
3173        preceding ".*" at the start, unless the pattern required a fixed first
3174        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3175    
3176    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3177        starting at the last subject character, bytes beyond the end of the subject
3178        string were read.
3179    
3180    11. Renamed the variable "class" as "classbits" to make life easier for C++
3181        users. (Previously there was a macro definition, but it apparently wasn't
3182        enough.)
3183    
3184    12. Added the new field "tables" to the extra data so that tables can be passed
3185        in at exec time, or the internal tables can be re-selected. This allows
3186        a compiled regex to be saved and re-used at a later time by a different
3187        program that might have everything at different addresses.
3188    
3189    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3190        -R library as well as a -L library.
3191    
3192    14. The debugging options of pcretest (-d on the command line or D on a
3193        pattern) showed incorrect output for anything following an extended class
3194        that contained multibyte characters and which was followed by a quantifier.
3195    
3196    15. Added optional support for general category Unicode character properties
3197        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3198        support. It adds about 90K to the size of the library. The meanings of the
3199        inbuilt class escapes such as \d and \s have NOT been changed.
3200    
3201    16. Updated pcredemo.c to include calls to free() to release the memory for the
3202        compiled pattern.
3203    
3204    17. The generated file chartables.c was being created in the source directory
3205        instead of in the building directory. This caused the build to fail if the
3206        source directory was different from the building directory, and was
3207        read-only.
3208    
3209    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3210        file. No doubt somebody will tell me if they don't make sense... Also added
3211        Dan Mooney's comments about building on OpenVMS.
3212    
3213    19. Added support for partial matching via the PCRE_PARTIAL option for
3214        pcre_exec() and the \P data escape in pcretest.
3215    
3216    20. Extended pcretest with 3 new pattern features:
3217    
3218        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3219              write the compiled pattern to the file whose name is "rest-of-line".
3220              This is a straight binary dump of the data, with the saved pointer to
3221              the character tables forced to be NULL. The study data, if any, is
3222              written too. After writing, pcretest reads a new pattern.
3223    
3224        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3225              compiled pattern from the given file. There must not be any
3226              occurrences of "<" in the file name (pretty unlikely); if there are,
3227              pcretest will instead treat the initial "<" as a pattern delimiter.
3228              After reading in the pattern, pcretest goes on to read data lines as
3229              usual.
3230    
3231        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3232              and 16-bit fields in a compiled pattern, to simulate a pattern that
3233              was compiled on a host of opposite endianness.
3234    
3235    21. The pcre-exec() function can now cope with patterns that were compiled on
3236        hosts of opposite endianness, with this restriction:
3237    
3238          As for any compiled expression that is saved and used later, the tables
3239          pointer field cannot be preserved; the extra_data field in the arguments
3240          to pcre_exec() should be used to pass in a tables address if a value
3241          other than the default internal tables were used at compile time.
3242    
3243    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3244        now diagnosed as an error. Previously, most of the time, a negative number
3245        would have been treated as zero, but if in addition "ovector" was passed as
3246        NULL, a crash could occur.
3247    
3248    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3249        new versions from the libtool 1.5 distribution (the last one is a copy of
3250        a file called libtool.m4). This seems to have fixed the need to patch
3251        "configure" to support Darwin 1.3 (which I used to do). However, I still
3252        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3253        workstation).
3254    
3255    24. Changed the PCRE licence to be the more standard "BSD" licence.
3256    
3257    
3258    Version 4.5 01-Dec-03
3259    ---------------------
3260    
3261     1. There has been some re-arrangement of the code for the match() function so
3262        that it can be compiled in a version that does not call itself recursively.
3263        Instead, it keeps those local variables that need separate instances for
3264        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3265        needs to "recurse". Keeping track of where control must go is done by means
3266        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3267        hide most of the details from the main code, and operates only if
3268        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3269        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3270        operating.
3271    
3272        To make it easier for callers to provide specially tailored get/free
3273        functions for this usage, two new functions, pcre_stack_malloc, and
3274        pcre_stack_free, are used. They are always called in strict stacking order,
3275        and the size of block requested is always the same.
3276    
3277        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3278        PCRE has been compiled to use the stack or the heap for recursion. The
3279        -C option of pcretest uses this to show which version is compiled.
3280    
3281        A new data escape \S, is added to pcretest; it causes the amounts of store
3282        obtained and freed by both kinds of malloc/free at match time to be added
3283        to the output.
3284    
3285     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3286        what's available on my current Linux desktop machine.
3287    
3288     3. When matching a UTF-8 string, the test for a valid string at the start has
3289        been extended. If start_offset is not zero, PCRE now checks that it points
3290        to a byte that is the start of a UTF-8 character. If not, it returns
3291        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3292        this is necessary because there may be backward assertions in the pattern.
3293        When matching the same subject several times, it may save resources to use
3294        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3295    
3296     4. The code for checking the validity of UTF-8 strings has been tightened so
3297        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3298        containing "overlong sequences".
3299    
3300     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3301        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3302        should have been "&", but it just so happened that all the cases this let
3303        through by mistake were picked up later in the function.
3304    
3305     6. I had used a variable called "isblank" - this is a C99 function, causing
3306        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3307    
3308     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3309        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3310        so the version is shown; (c) stop "make test" echoing "./RunTest".
3311    
3312     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3313    
3314     9. The prototype for memmove() for systems that don't have it was using
3315        size_t, but the inclusion of the header that defines size_t was later. I've
3316        moved the #includes for the C headers earlier to avoid this.
3317    
3318    10. Added some adjustments to the code to make it easier to compiler on certain
3319        special systems:
3320    
3321          (a) Some "const" qualifiers were missing.
3322          (b) Added the macro EXPORT before all exported functions; by default this
3323              is defined to be empty.
3324          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3325              that it reads its output file name as an argument instead of writing
3326              to the standard output and assuming this can be redirected.
3327    
3328    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3329        class containing characters with values greater than 255, PCRE compilation
3330        went into a loop.
3331    
3332    12. A recursive reference to a subpattern that was within another subpattern
3333        that had a minimum quantifier of zero caused PCRE to crash. For example,
3334        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3335        recursion. If the recursively-called subpattern itself had a zero repeat,
3336        that was OK.
3337    
3338    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3339        buffer into which it was copied (for escape processing) was still set at
3340        1024, so long lines caused crashes.
3341    
3342    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3343        "internal error: code overflow...". This applied to any character class
3344        that was followed by a possessive quantifier.
3345    
3346    15. Modified the Makefile to add libpcre.la as a prerequisite for
3347        libpcreposix.la because I was told this is needed for a parallel build to
3348        work.
3349    
3350    16. If a pattern that contained .* following optional items at the start was
3351        studied, the wrong optimizing data was generated, leading to matching
3352        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3353        matching string must start with a or b or c. The correct conclusion for
3354        this pattern is that a match can start with any character.
3355    
3356    
3357    Version 4.4 13-Aug-03
3358    ---------------------
3359    
3360     1. In UTF-8 mode, a character class containing characters with values between
3361        127 and 255 was not handled correctly if the compiled pattern was studied.
3362        In fixing this, I have also improved the studying algorithm for such
3363        classes (slightly).
3364    
3365     2. Three internal functions had redundant arguments passed to them. Removal
3366        might give a very teeny performance improvement.
3367    
3368     3. Documentation bug: the value of the capture_top field in a callout is *one
3369        more than* the number of the hightest numbered captured substring.
3370    
3371     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3372        in incorrectly linking with a previously installed version. They now link
3373        explicitly with libpcre.la.
3374    
3375     5. configure.in no longer needs to recognize Cygwin specially.
3376    
3377     6. A problem in pcre.in for Windows platforms is fixed.
3378    
3379     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3380        pcretest, it used to include the size of the study block as part of its
3381        output. Unfortunately, the structure contains a field that has a different
3382        size on different hardware architectures. This meant that the tests that
3383        showed this size failed. As the block is currently always of a fixed size,
3384        this information isn't actually particularly useful in pcretest output, so
3385        I have just removed it.
3386    
3387     8. Three pre-processor statements accidentally did not start in column 1.
3388        Sadly, there are *still* compilers around that complain, even though
3389        standard C has not required this for well over a decade. Sigh.
3390    
3391     9. In pcretest, the code for checking callouts passed small integers in the
3392        callout_data field, which is a void * field. However, some picky compilers
3393        complained about the casts involved for this on 64-bit systems. Now
3394        pcretest passes the address of the small integer instead, which should get
3395        rid of the warnings.
3396    
3397    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3398        both compile and run time, and gives an error if an invalid UTF-8 sequence
3399        is found. There is a option for disabling this check in cases where the
3400        string is known to be correct and/or the maximum performance is wanted.
3401    
3402    11. In response to a bug report, I changed one line in Makefile.in from
3403    
3404            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3405        to
3406            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3407    
3408        to look similar to other lines, but I have no way of telling whether this
3409        is the right thing to do, as I do not use Windows. No doubt I'll get told
3410        if it's wrong...
3411    
3412    
3413    Version 4.3 21-May-03
3414    ---------------------
3415    
3416    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3417       Makefile.
3418    
3419    2. Some refactoring to improve the quality of the code:
3420    
3421       (i)   The utf8_table... variables are now declared "const".
3422    
3423       (ii)  The code for \cx, which used the "case flipping" table to upper case
3424             lower case letters, now just substracts 32. This is ASCII-specific,
3425             but the whole concept of \cx is ASCII-specific, so it seems
3426             reasonable.
3427    
3428       (iii) PCRE was using its character types table to recognize decimal and
3429             hexadecimal digits in the pattern. This is silly, because it handles
3430             only 0-9, a-f, and A-F, but the character types table is locale-
3431             specific, which means strange things might happen. A private
3432             table is now used for this - though it costs 256 bytes, a table is
3433             much faster than multiple explicit tests. Of course, the standard
3434             character types table is still used for matching digits in subject
3435             strings against \d.
3436    
3437       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3438             ending in _t are). So I've renamed it as ESC_tee.
3439    
3440    3. The first argument for regexec() in the POSIX wrapper should have been
3441       defined as "const".
3442    
3443    4. Changed pcretest to use malloc() for its buffers so that they can be
3444       Electric Fenced for debugging.
3445    
3446    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3447       to read one or more bytes before the start of the subject string. Often this
3448       had no effect on PCRE's behaviour, but in some circumstances it could
3449       provoke a segmentation fault.
3450    
3451    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3452       to try to read one or more bytes before the start of the subject string.
3453    
3454    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3455       UTF-8 support could misbehave in various ways if the subject string
3456       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3457       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3458       back over UTF-8 characters.)
3459    
3460    
3461    Version 4.2 14-Apr-03
3462    ---------------------
3463    
3464    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3465    
3466    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3467         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3468         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3469         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3470         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3471           and BUILD_EXEEXT
3472         Note: automatic setting of the BUILD variables is not yet working
3473         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3474           compile-time but not at link-time
3475         [LINK]: use for linking executables only
3476         make different versions for Windows and non-Windows
3477         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3478           libraries
3479         [LINK_FOR_BUILD]: new variable
3480         [OBJEXT]: use throughout
3481         [EXEEXT]: use throughout
3482         <winshared>: new target
3483         <wininstall>: new target
3484         <dftables.o>: use native compiler
3485         <dftables>: use native linker
3486         <install>: handle Windows platform correctly
3487         <clean>: ditto
3488         <check>: ditto
3489         copy DLL to top builddir before testing
3490    
3491       As part of these changes, -no-undefined was removed again. This was reported
3492       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3493       in any case.
3494    
3495    3. Some tidies to get rid of compiler warnings:
3496    
3497       . In the match_data structure, match_limit was an unsigned long int, whereas
3498         match_call_count was an int. I've made them both unsigned long ints.
3499    
3500       . In pcretest the fact that a const uschar * doesn't automatically cast to
3501         a void * provoked a warning.
3502    
3503       . Turning on some more compiler warnings threw up some "shadow" variables
3504         and a few more missing casts.
3505    
3506    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3507       option, a class that contained a single character with a value between 128
3508       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3509    
3510    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3511       option, a class that contained several characters, but with at least one
3512       whose value was between 128 and 255 caused PCRE to crash.
3513    
3514    
3515    Version 4.1 12-Mar-03
3516    ---------------------
3517    
3518    1. Compiling with gcc -pedantic found a couple of places where casts were
3519    needed, and a string in dftables.c that was longer than standard compilers are
3520    required to support.
3521    
3522    2. Compiling with Sun's compiler found a few more places where the code could
3523    be tidied up in order to avoid warnings.
3524    
3525    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3526    first of these names is deprecated in the latest Autoconf in favour of the name
3527    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3528    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3529    analogy I have changed it to CFLAGS_FOR_BUILD.
3530    
3531    4. Added -no-undefined to the linking command in the Makefile, because this is
3532    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3533    linking step for the pcreposix library.
3534    
3535    5. PCRE was failing to diagnose the case of two named groups with the same
3536    name.
3537    
3538    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3539    literal character that is needed in the subject for a match, and scans along to
3540    ensure that it is present before embarking on the full matching process. This
3541    saves time in cases of nested unlimited repeats that are never going to match.
3542    Problem: the scan can take a lot of time if the subject is very long (e.g.
3543    megabytes), thus penalizing straightforward matches. It is now done only if the
3544    amount of subject to be scanned is less than 1000 bytes.
3545    
3546    7. A lesser problem with the same optimization is that it was recording the
3547    first character of an anchored pattern as "needed", thus provoking a search
3548    right along the subject, even when the first match of the pattern was going to
3549    fail. The "needed" character is now not set for anchored patterns, unless it
3550    follows something in the pattern that is of non-fixed length. Thus, it still
3551    fulfils its original purpose of finding quick non-matches in cases of nested
3552    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3553    
3554    
3555    Version 4.0 17-Feb-03
3556    ---------------------
3557    
3558    1. If a comment in an extended regex that started immediately after a meta-item
3559    extended to the end of string, PCRE compiled incorrect data. This could lead to
3560    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3561    
3562    2. Moved to autoconf 2.53 and libtool 1.4.2.
3563    
3564    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3565    the special perltest8 script is no longer needed - all the tests can be run
3566    from a single perltest script.
3567    
3568    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3569    by \s. It has now been removed in PCRE. This means it isn't recognized as
3570    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3571    class [:space:] *does* include VT, thereby creating a mess.
3572    
3573    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3574    space and tab.
3575    
3576    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3577    its new features into the main test script, reducing the number of scripts.
3578    
3579    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3580    were backward compatible, and made the (?i) apply to the whole pattern, as if
3581    /i were given. Now it behaves more logically, and applies the option setting
3582    only to what follows. PCRE has been changed to follow suit. However, if it
3583    finds options settings right at the start of the pattern, it extracts them into
3584    the global options, as before. Thus, they show up in the info data.
3585    
3586    8. Added support for the \Q...\E escape sequence. Characters in between are
3587    treated as literals. This is slightly different from Perl in that $ and @ are
3588    also handled as literals inside the quotes. In Perl, they will cause variable