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

Diff of /code/trunk/ChangeLog

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

revision 11 by nigel, Sat Feb 24 21:38:17 2007 UTC revision 912 by ph10, Sun Feb 12 16:53:39 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    
12    Version 8.30 04-February-2012
13    -----------------------------
14    
15    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
16        name is defined in ctype.h.
17    
18    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
19        only in quite long subpatterns.
20    
21    3.  Removed the function pcre_info(), which has been obsolete and deprecated
22        since it was replaced by pcre_fullinfo() in February 2000.
23    
24    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
25        match a (*MARK), and the match failed at the start of the subject, a
26        reference to memory before the start of the subject could occur. This bug
27        was introduced by fix 17 of release 8.21.
28    
29    5.  A reference to an unset group with zero minimum repetition was giving
30        totally wrong answers (in non-JavaScript-compatibility mode). For example,
31        /(another)?(\1?)test/ matched against "hello world test". This bug was
32        introduced in release 8.13.
33    
34    6.  Add support for 16-bit character strings (a large amount of work involving
35        many changes and refactorings).
36    
37    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
38        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
39        from a file.
40    
41    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
42        rounding is not applied in this particular case).
43    
44    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
45        if they appear, or are escaped, in patterns.
46    
47    10. Get rid of a number of -Wunused-but-set-variable warnings.
48    
49    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
50        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
51        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
52        also returns the mark "x". This bug applied to capturing parentheses,
53        non-capturing parentheses, and atomic parentheses. It also applied to some
54        assertions.
55    
56    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
57        information out of configure.ac instead of relying on pcre.h.generic, which
58        is not stored in the repository.
59    
60    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
61        -lreadline.
62    
63    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
64    
65    15. Applied Graycode's patch to put the top-level frame on the stack rather
66        than the heap when not using the stack for recursion. This gives a
67        performance improvement in many cases when recursion is not deep.
68    
69    16. Experimental code added to "pcretest -C" to output the stack frame size.
70    
71    
72    Version 8.21 12-Dec-2011
73    ------------------------
74    
75    1.  Updating the JIT compiler.
76    
77    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
78        are added as well.
79    
80    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
81        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
82        calling _pcre_jit_exec. Some extra comments are added.
83    
84    4.  (*MARK) settings inside atomic groups that do not contain any capturing
85        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
86        was introduced by change 18 for 8.20.
87    
88    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
89        ECMA-262 standard.
90    
91    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
92        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
93        This bug was probably introduced by change 9 of 8.13.
94    
95    7.  While fixing 6 above, I noticed that a number of other items were being
96        incorrectly rejected as "not fixed length". This arose partly because newer
97        opcodes had not been added to the fixed-length checking code. I have (a)
98        corrected the bug and added tests for these items, and (b) arranged for an
99        error to occur if an unknown opcode is encountered while checking for fixed
100        length instead of just assuming "not fixed length". The items that were
101        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
102        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
103        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
104    
105    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
106        being incorrectly compiled and would have given unpredicatble results.
107    
108    9.  A possessively repeated subpattern with minimum repeat count greater than
109        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
110        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
111        the first (A) could occur when it should not.
112    
113    10. Add a cast and remove a redundant test from the code.
114    
115    11. JIT should use pcre_malloc/pcre_free for allocation.
116    
117    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
118        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
119        is anything other than /usr, -L is still shown).
120    
121    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
122    
123    14. Perl does not support \N without a following name in a [] class; PCRE now
124        also gives an error.
125    
126    15. If a forward reference was repeated with an upper limit of around 2000,
127        it caused the error "internal error: overran compiling workspace". The
128        maximum number of forward references (including repeats) was limited by the
129        internal workspace, and dependent on the LINK_SIZE. The code has been
130        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
131        the default depends on LINK_SIZE. There is a new upper limit (for safety)
132        of around 200,000 forward references. While doing this, I also speeded up
133        the filling in of repeated forward references.
134    
135    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
136        incorrectly expecting the subject to contain another "a" after the start.
137    
138    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
139        in the match, the SKIP should be ignored. This was not happening; instead
140        the SKIP was being treated as NOMATCH. For patterns such as
141        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
142        tested.
143    
144    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
145        now much more compatible with Perl, in particular in cases where the result
146        is a non-match for a non-anchored pattern. For example, if
147        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
148        "m", where previously it did not return a name. A side effect of this
149        change is that for partial matches, the last encountered mark name is
150        returned, as for non matches. A number of tests that were previously not
151        Perl-compatible have been moved into the Perl-compatible test files. The
152        refactoring has had the pleasing side effect of removing one argument from
153        the match() function, thus reducing its stack requirements.
154    
155    19. If the /S+ option was used in pcretest to study a pattern using JIT,
156        subsequent uses of /S (without +) incorrectly behaved like /S+.
157    
158    21. Retrieve executable code size support for the JIT compiler and fixing
159        some warnings.
160    
161    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
162        not work when the shorter character appeared right at the end of the
163        subject string.
164    
165    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
166        systems.
167    
168    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
169        output it when the /M option is used in pcretest.
170    
171    25. The CheckMan script was not being included in the distribution. Also, added
172        an explicit "perl" to run Perl scripts from the PrepareRelease script
173        because this is reportedly needed in Windows.
174    
175    26. If study data was being save in a file and studying had not found a set of
176        "starts with" bytes for the pattern, the data written to the file (though
177        never used) was taken from uninitialized memory and so caused valgrind to
178        complain.
179    
180    27. Updated RunTest.bat as provided by Sheri Pierce.
181    
182    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
183    
184    29. Computation of memory usage for the table of capturing group names was
185        giving an unnecessarily large value.
186    
187    
188    Version 8.20 21-Oct-2011
189    ------------------------
190    
191    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
192        a POSIX class. After further experiments with Perl, which convinced me that
193        Perl has bugs and confusions, a closing square bracket is no longer allowed
194        in a POSIX name. This bug also affected patterns with classes that started
195        with full stops.
196    
197    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
198        captured substring, but while checking the failing first alternative,
199        substring 1 is temporarily captured. If the output vector supplied to
200        pcre_exec() was not big enough for this capture, the yield of the function
201        was still zero ("insufficient space for captured substrings"). This cannot
202        be totally fixed without adding another stack variable, which seems a lot
203        of expense for a edge case. However, I have improved the situation in cases
204        such as /(a)(b)x|abc/ matched against "abc", where the return code
205        indicates that fewer than the maximum number of slots in the ovector have
206        been set.
207    
208    3.  Related to (2) above: when there are more back references in a pattern than
209        slots in the output vector, pcre_exec() uses temporary memory during
210        matching, and copies in the captures as far as possible afterwards. It was
211        using the entire output vector, but this conflicts with the specification
212        that only 2/3 is used for passing back captured substrings. Now it uses
213        only the first 2/3, for compatibility. This is, of course, another edge
214        case.
215    
216    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
217        main code base, and can be used by building with --enable-jit. When this is
218        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
219        runtime --no-jit option is given.
220    
221    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
222        ovector, the return from the function was zero, implying that there were
223        other matches that did not fit. The correct "exactly full" value is now
224        returned.
225    
226    6.  If a subpattern that was called recursively or as a subroutine contained
227        (*PRUNE) or any other control that caused it to give a non-standard return,
228        invalid errors such as "Error -26 (nested recursion at the same subject
229        position)" or even infinite loops could occur.
230    
231    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
232        computing the minimum length on reaching *ACCEPT, and so ended up with the
233        wrong value of 1 rather than 0. Further investigation indicates that
234        computing a minimum subject length in the presence of *ACCEPT is difficult
235        (think back references, subroutine calls), and so I have changed the code
236        so that no minimum is registered for a pattern that contains *ACCEPT.
237    
238    8.  If (*THEN) was present in the first (true) branch of a conditional group,
239        it was not handled as intended. [But see 16 below.]
240    
241    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
242        Sheri Pierce.
243    
244    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
245        the first byte in a match must be "a".
246    
247    11. Change 17 for 8.13 increased the recursion depth for patterns like
248        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
249        pattern contains any instances of (*THEN). If it does not, the old
250        optimizations are restored. It would be nice to do this on a per-group
251        basis, but at the moment that is not feasible.
252    
253    12. In some environments, the output of pcretest -C is CRLF terminated. This
254        broke RunTest's code that checks for the link size. A single white space
255        character after the value is now allowed for.
256    
257    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
258        For "fr", it uses the Windows-specific input and output files.
259    
260    14. If (*THEN) appeared in a group that was called recursively or as a
261        subroutine, it did not work as intended. [But see next item.]
262    
263    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
264        pattern fragments (but not containing any | characters). If A and B are
265        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
266        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
267        D. In other words, Perl considers parentheses that do not contain any |
268        characters to be part of a surrounding alternative, whereas PCRE was
269        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
270        differently. PCRE now behaves in the same way as Perl, except in the case
271        of subroutine/recursion calls such as (?1) which have in any case always
272        been different (but PCRE had them first :-).
273    
274    16. Related to 15 above: Perl does not treat the | in a conditional group as
275        creating alternatives. Such a group is treated in the same way as an
276        ordinary group without any | characters when processing (*THEN). PCRE has
277        been changed to match Perl's behaviour.
278    
279    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
280        RunGrepTest script failed.
281    
282    18. Change 22 for version 13 caused atomic groups to use more stack. This is
283        inevitable for groups that contain captures, but it can lead to a lot of
284        stack use in large patterns. The old behaviour has been restored for atomic
285        groups that do not contain any capturing parentheses.
286    
287    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
288        suppress the check for a minimum subject length at run time. (If it was
289        given to pcre_exec() or pcre_dfa_exec() it did work.)
290    
291    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
292        fail to work when decoding hex characters in data strings in EBCDIC
293        environments.
294    
295    21. It appears that in at least one Mac OS environment, the isxdigit() function
296        is implemented as a macro that evaluates to its argument more than once,
297        contravening the C 90 Standard (I haven't checked a later standard). There
298        was an instance in pcretest which caused it to go wrong when processing
299        \x{...} escapes in subject strings. The has been rewritten to avoid using
300        things like p++ in the argument of isxdigit().
301    
302    
303    Version 8.13 16-Aug-2011
304    ------------------------
305    
306    1.  The Unicode data tables have been updated to Unicode 6.0.0.
307    
308    2.  Two minor typos in pcre_internal.h have been fixed.
309    
310    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
311        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
312        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
313    
314    4.  There were a number of related bugs in the code for matching backrefences
315        caselessly in UTF-8 mode when codes for the characters concerned were
316        different numbers of bytes. For example, U+023A and U+2C65 are an upper
317        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
318        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
319        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
320        2-byte code at the end of the subject (it thought there wasn't enough data
321        left).
322    
323    5.  Comprehensive information about what went wrong is now returned by
324        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
325        as the output vector has at least 2 elements. The offset of the start of
326        the failing character and a reason code are placed in the vector.
327    
328    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
329        now returned is for the first byte of the failing character, instead of the
330        last byte inspected. This is an incompatible change, but I hope it is small
331        enough not to be a problem. It makes the returned offset consistent with
332        pcre_exec() and pcre_dfa_exec().
333    
334    7.  pcretest now gives a text phrase as well as the error number when
335        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
336        failure, the offset and reason code are output.
337    
338    8.  When \R was used with a maximizing quantifier it failed to skip backwards
339        over a \r\n pair if the subsequent match failed. Instead, it just skipped
340        back over a single character (\n). This seems wrong (because it treated the
341        two characters as a single entity when going forwards), conflicts with the
342        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
343        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
344        has been changed.
345    
346    9.  Some internal refactoring has changed the processing so that the handling
347        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
348        time (the PCRE_DOTALL option was changed this way some time ago: version
349        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
350        which was always a bit of a fudge. It also means that there is one less
351        argument for the match() function, which reduces its stack requirements
352        slightly. This change also fixes an incompatibility with Perl: the pattern
353        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
354    
355    10. More internal refactoring has drastically reduced the number of recursive
356        calls to match() for possessively repeated groups such as (abc)++ when
357        using pcre_exec().
358    
359    11. While implementing 10, a number of bugs in the handling of groups were
360        discovered and fixed:
361    
362        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
363        (a|)*(?1) gave a compile-time internal error.
364        ((a|)+)+  did not notice that the outer group could match an empty string.
365        (^a|^)+   was not marked as anchored.
366        (.*a|.*)+ was not marked as matching at start or after a newline.
367    
368    12. Yet more internal refactoring has removed another argument from the match()
369        function. Special calls to this function are now indicated by setting a
370        value in a variable in the "match data" data block.
371    
372    13. Be more explicit in pcre_study() instead of relying on "default" for
373        opcodes that mean there is no starting character; this means that when new
374        ones are added and accidentally left out of pcre_study(), testing should
375        pick them up.
376    
377    14. The -s option of pcretest has been documented for ages as being an old
378        synonym of -m (show memory usage). I have changed it to mean "force study
379        for every regex", that is, assume /S for every regex. This is similar to -i
380        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
381        using it. It makes it easier to run collections of tests with and without
382        study enabled, and thereby test pcre_study() more easily. All the standard
383        tests are now run with and without -s (but some patterns can be marked as
384        "never study" - see 20 below).
385    
386    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
387        restoration of the capturing data to the outer values was not happening
388        correctly.
389    
390    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
391        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
392        pattern had matched an empty string, and so incorrectly returned a no
393        match.
394    
395    17. There was optimizing code for the last branch of non-capturing parentheses,
396        and also for the obeyed branch of a conditional subexpression, which used
397        tail recursion to cut down on stack usage. Unfortunately, now that there is
398        the possibility of (*THEN) occurring in these branches, tail recursion is
399        no longer possible because the return has to be checked for (*THEN). These
400        two optimizations have therefore been removed. [But see 8.20/11 above.]
401    
402    18. If a pattern containing \R was studied, it was assumed that \R always
403        matched two bytes, thus causing the minimum subject length to be
404        incorrectly computed because \R can also match just one byte.
405    
406    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
407        was incorrectly computed.
408    
409    20. If /S is present twice on a test pattern in pcretest input, it now
410        *disables* studying, thereby overriding the use of -s on the command line
411        (see 14 above). This is necessary for one or two tests to keep the output
412        identical in both cases.
413    
414    21. When (*ACCEPT) was used in an assertion that matched an empty string and
415        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
416    
417    22. When an atomic group that contained a capturing parenthesis was
418        successfully matched, but the branch in which it appeared failed, the
419        capturing was not being forgotten if a higher numbered group was later
420        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
421        group 1 to "a", when in fact it should be unset. This applied to multi-
422        branched capturing and non-capturing groups, repeated or not, and also to
423        positive assertions (capturing in negative assertions does not happen
424        in PCRE) and also to nested atomic groups.
425    
426    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
427        subject after a captured substring, to make it easier to tell which of a
428        number of identical substrings has been captured.
429    
430    24. The way atomic groups are processed by pcre_exec() has been changed so that
431        if they are repeated, backtracking one repetition now resets captured
432        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
433        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
434        "aaa". Previously, it would have been "a". As part of this code
435        refactoring, the way recursive calls are handled has also been changed.
436    
437    25. If an assertion condition captured any substrings, they were not passed
438        back unless some other capturing happened later. For example, if
439        (?(?=(a))a) was matched against "a", no capturing was returned.
440    
441    26. When studying a pattern that contained subroutine calls or assertions,
442        the code for finding the minimum length of a possible match was handling
443        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
444        group 1 called group 2 while simultaneously a separate group 2 called group
445        1). A stack overflow occurred in this case. I have fixed this by limiting
446        the recursion depth to 10.
447    
448    27. Updated RunTest.bat in the distribution to the version supplied by Tom
449        Fortmann. This supports explicit test numbers on the command line, and has
450        argument validation and error reporting.
451    
452    28. An instance of \X with an unlimited repeat could fail if at any point the
453        first character it looked at was a mark character.
454    
455    29. Some minor code refactoring concerning Unicode properties and scripts
456        should reduce the stack requirement of match() slightly.
457    
458    30. Added the '=' option to pcretest to check the setting of unused capturing
459        slots at the end of the pattern, which are documented as being -1, but are
460        not included in the return count.
461    
462    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
463        compiled something random. Now it gives a compile-time error (as does
464        Perl).
465    
466    32. A *MARK encountered during the processing of a positive assertion is now
467        recorded and passed back (compatible with Perl).
468    
469    33. If --only-matching or --colour was set on a pcregrep call whose pattern
470        had alternative anchored branches, the search for a second match in a line
471        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
472        matched the line "0102" twice. The same bug affected patterns that started
473        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
474        twice.
475    
476    34. Previously, PCRE did not allow quantification of assertions. However, Perl
477        does, and because of capturing effects, quantifying parenthesized
478        assertions may at times be useful. Quantifiers are now allowed for
479        parenthesized assertions.
480    
481    35. A minor code tidy in pcre_compile() when checking options for \R usage.
482    
483    36. \g was being checked for fancy things in a character class, when it should
484        just be a literal "g".
485    
486    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
487        appearance of a nested POSIX class supersedes an apparent external class.
488        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
489        unescaped square brackets may also appear as part of class names. For
490        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
491        more like Perl. (But see 8.20/1 above.)
492    
493    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
494        was because it thought it was \N{name}, which is not supported).
495    
496    39. Add minix to OS list not supporting the -S option in pcretest.
497    
498    40. PCRE tries to detect cases of infinite recursion at compile time, but it
499        cannot analyze patterns in sufficient detail to catch mutual recursions
500        such as ((?1))((?2)). There is now a runtime test that gives an error if a
501        subgroup is called recursively as a subpattern for a second time at the
502        same position in the subject string. In previous releases this might have
503        been caught by the recursion limit, or it might have run out of stack.
504    
505    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
506        happen only once. PCRE was, however incorrectly giving a compile time error
507        "recursive call could loop indefinitely" because it cannot analyze the
508        pattern in sufficient detail. The compile time test no longer happens when
509        PCRE is compiling a conditional subpattern, but actual runaway loops are
510        now caught at runtime (see 40 above).
511    
512    42. It seems that Perl allows any characters other than a closing parenthesis
513        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
514        has been changed to be the same.
515    
516    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
517        as not to get warnings when autogen.sh is called. Also changed
518        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
519    
520    44. To help people who use pcregrep to scan files containing exceedingly long
521        lines, the following changes have been made:
522    
523        (a) The default value of the buffer size parameter has been increased from
524            8K to 20K. (The actual buffer used is three times this size.)
525    
526        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
527            PCRE is built.
528    
529        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
530            to be set at run time.
531    
532        (d) Numerical values in pcregrep options can be followed by K or M, for
533            example --buffer-size=50K.
534    
535        (e) If a line being scanned overflows pcregrep's buffer, an error is now
536            given and the return code is set to 2.
537    
538    45. Add a pointer to the latest mark to the callout data block.
539    
540    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
541        partial match of an empty string instead of no match. This was specific to
542        the use of ".".
543    
544    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
545        complete match instead of a partial match. This bug was dependent on both
546        the PCRE_UTF8 and PCRE_DOTALL options being set.
547    
548    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
549        starting byte set, because \b was not being ignored.
550    
551    
552    Version 8.12 15-Jan-2011
553    ------------------------
554    
555    1.  Fixed some typos in the markup of the man pages, and wrote a script that
556        checks for such things as part of the documentation building process.
557    
558    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
559        --match-limit and --recursion-limit options (added for 8.11). In
560        particular, this made one of the standard tests fail. (The integer value
561        went into the wrong half of a long int.)
562    
563    3.  If the --colour option was given to pcregrep with -v (invert match), it
564        did strange things, either producing crazy output, or crashing. It should,
565        of course, ignore a request for colour when reporting lines that do not
566        match.
567    
568    4.  Another pcregrep bug caused similar problems if --colour was specified with
569        -M (multiline) and the pattern match finished with a line ending.
570    
571    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
572        matched in multiline mode, the following line was shown as part of the
573        match. This seems wrong, so I have changed it.
574    
575    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
576        the check for further matches in the same line (so they could be coloured)
577        to overrun the end of the current line. If another match was found, it was
578        incorrectly shown (and then shown again when found in the next line).
579    
580    7.  If pcregrep was compiled under Windows, there was a reference to the
581        function pcregrep_exit() before it was defined. I am assuming this was
582        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
583        reported by a user. I've moved the definition above the reference.
584    
585    
586    Version 8.11 10-Dec-2010
587    ------------------------
588    
589    1.  (*THEN) was not working properly if there were untried alternatives prior
590        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
591        backtracked to try for "b" instead of moving to the next alternative branch
592        at the same level (in this case, to look for "c"). The Perl documentation
593        is clear that when (*THEN) is backtracked onto, it goes to the "next
594        alternative in the innermost enclosing group".
595    
596    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
597        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
598        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
599        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
600        (*THEN).
601    
602    3.  If \s appeared in a character class, it removed the VT character from
603        the class, even if it had been included by some previous item, for example
604        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
605        of \s, but is part of the POSIX "space" class.)
606    
607    4.  A partial match never returns an empty string (because you can always
608        match an empty string at the end of the subject); however the checking for
609        an empty string was starting at the "start of match" point. This has been
610        changed to the "earliest inspected character" point, because the returned
611        data for a partial match starts at this character. This means that, for
612        example, /(?<=abc)def/ gives a partial match for the subject "abc"
613        (previously it gave "no match").
614    
615    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
616        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
617        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
618        has an implication that the given string is incomplete (because a partial
619        match is preferred over a full match). For this reason, these items now
620        give a partial match in this situation. [Aside: previously, the one case
621        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
622        match rather than a full match, which was wrong by the old rules, but is
623        now correct.]
624    
625    6.  There was a bug in the handling of #-introduced comments, recognized when
626        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
627        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
628        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
629        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
630        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
631        places in pcre_compile().
632    
633    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
634        comments when looking ahead for named forward references to subpatterns,
635        the only newline sequence it recognized was NL. It now handles newlines
636        according to the set newline convention.
637    
638    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
639        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
640        cater for a lack of strerror(). These oversights have been fixed.
641    
642    9.  Added --match-limit and --recursion-limit to pcregrep.
643    
644    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
645    
646    11. When the -o option was used, pcregrep was setting a return code of 1, even
647        when matches were found, and --line-buffered was not being honoured.
648    
649    12. Added an optional parentheses number to the -o and --only-matching options
650        of pcregrep.
651    
652    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
653        can match an empty string. The code to do it in pcretest and pcredemo
654        needed fixing:
655    
656        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
657            only one byte after an empty string match just before CRLF (this case
658            just got forgotten; "any" and "anycrlf" were OK).
659    
660        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
661            mode when an empty string match preceded an ASCII character followed by
662            a non-ASCII character. (The code for advancing by one character rather
663            than one byte was nonsense.)
664    
665        (c) The pcredemo.c sample program did not have any code at all to handle
666            the cases when CRLF is a valid newline sequence.
667    
668    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
669        as a starting offset was within the subject string. There is now a new
670        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
671        negative or greater than the length of the string. In order to test this,
672        pcretest is extended to allow the setting of negative starting offsets.
673    
674    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
675        starting offset points to the beginning of a UTF-8 character was
676        unnecessarily clumsy. I tidied it up.
677    
678    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
679        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
680    
681    17. Nobody had reported that the --include_dir option, which was added in
682        release 7.7 should have been called --include-dir (hyphen, not underscore)
683        for compatibility with GNU grep. I have changed it to --include-dir, but
684        left --include_dir as an undocumented synonym, and the same for
685        --exclude-dir, though that is not available in GNU grep, at least as of
686        release 2.5.4.
687    
688    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
689        characters from a string of bytes) have been redefined so as not to use
690        loops, in order to improve performance in some environments. At the same
691        time, I abstracted some of the common code into auxiliary macros to save
692        repetition (this should not affect the compiled code).
693    
694    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
695        compile-time error is now given if \c is not followed by an ASCII
696        character, that is, a byte less than 128. (In EBCDIC mode, the code is
697        different, and any byte value is allowed.)
698    
699    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
700        START_OPTIMIZE option, which is now allowed at compile time - but just
701        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
702        to pcregrep and other applications that have no direct access to PCRE
703        options. The new /Y option in pcretest sets this option when calling
704        pcre_compile().
705    
706    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
707        back references. Groups containing recursive back references were forced to
708        be atomic by that change, but in the case of named groups, the amount of
709        memory required was incorrectly computed, leading to "Failed: internal
710        error: code overflow". This has been fixed.
711    
712    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
713        pcretest.c, to avoid build problems in some Borland environments.
714    
715    
716    Version 8.10 25-Jun-2010
717    ------------------------
718    
719    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
720        THEN.
721    
722    2.  (*ACCEPT) was not working when inside an atomic group.
723    
724    3.  Inside a character class, \B is treated as a literal by default, but
725        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
726        causes the error). The code is unchanged, but I tidied the documentation.
727    
728    4.  Inside a character class, PCRE always treated \R and \X as literals,
729        whereas Perl faults them if its -w option is set. I have changed PCRE so
730        that it faults them when PCRE_EXTRA is set.
731    
732    5.  Added support for \N, which always matches any character other than
733        newline. (It is the same as "." when PCRE_DOTALL is not set.)
734    
735    6.  When compiling pcregrep with newer versions of gcc which may have
736        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
737        declared with attribute warn_unused_result" were given. Just casting the
738        result to (void) does not stop the warnings; a more elaborate fudge is
739        needed. I've used a macro to implement this.
740    
741    7.  Minor change to pcretest.c to avoid a compiler warning.
742    
743    8.  Added four artifical Unicode properties to help with an option to make
744        \s etc use properties (see next item). The new properties are: Xan
745        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
746    
747    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
748        use Unicode properties. (*UCP) at the start of a pattern can be used to set
749        this option. Modified pcretest to add /W to test this facility. Added
750        REG_UCP to make it available via the POSIX interface.
751    
752    10. Added --line-buffered to pcregrep.
753    
754    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
755        studied, and the match started with a letter with a code point greater than
756        127 whose first byte was different to the first byte of the other case of
757        the letter, the other case of this starting letter was not recognized
758        (#976).
759    
760    12. If a pattern that was studied started with a repeated Unicode property
761        test, for example, \p{Nd}+, there was the theoretical possibility of
762        setting up an incorrect bitmap of starting bytes, but fortunately it could
763        not have actually happened in practice until change 8 above was made (it
764        added property types that matched character-matching opcodes).
765    
766    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
767        possible starting bytes for non-anchored patterns.
768    
769    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
770        \R, and also a number of cases that involve Unicode properties, both
771        explicit and implicit when PCRE_UCP is set.
772    
773    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
774        input, it could crash or give wrong results if characters with values
775        greater than 0xc0 were present in the subject string. (Detail: it assumed
776        UTF-8 input when processing these items.)
777    
778    16. Added a lot of (int) casts to avoid compiler warnings in systems where
779        size_t is 64-bit (#991).
780    
781    17. Added a check for running out of memory when PCRE is compiled with
782        --disable-stack-for-recursion (#990).
783    
784    18. If the last data line in a file for pcretest does not have a newline on
785        the end, a newline was missing in the output.
786    
787    19. The default pcre_chartables.c file recognizes only ASCII characters (values
788        less than 128) in its various bitmaps. However, there is a facility for
789        generating tables according to the current locale when PCRE is compiled. It
790        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
791        characters, are recognized by isspace() and therefore were getting set in
792        these tables, and indeed these tables seem to approximate to ISO 8859. This
793        caused a problem in UTF-8 mode when pcre_study() was used to create a list
794        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
795        which of course cannot start UTF-8 characters. I have changed the code so
796        that only real ASCII characters (less than 128) and the correct starting
797        bytes for UTF-8 encodings are set for characters greater than 127 when in
798        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
799        altogether.)
800    
801    20. Added the /T option to pcretest so as to be able to run tests with non-
802        standard character tables, thus making it possible to include the tests
803        used for 19 above in the standard set of tests.
804    
805    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
806        reference to a subpattern the other side of a comment that contains an
807        opening parenthesis caused either an internal compiling error, or a
808        reference to the wrong subpattern.
809    
810    
811    Version 8.02 19-Mar-2010
812    ------------------------
813    
814    1.  The Unicode data tables have been updated to Unicode 5.2.0.
815    
816    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
817        configured.
818    
819    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
820        original author of that file, following a query about its status.
821    
822    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
823        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
824    
825    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
826        quantifier applied to a forward-referencing subroutine call, could compile
827        incorrect code or give the error "internal error: previously-checked
828        referenced subpattern not found".
829    
830    6.  Both MS Visual Studio and Symbian OS have problems with initializing
831        variables to point to external functions. For these systems, therefore,
832        pcre_malloc etc. are now initialized to local functions that call the
833        relevant global functions.
834    
835    7.  There were two entries missing in the vectors called coptable and poptable
836        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
837        I've fixed the data, and added a kludgy way of testing at compile time that
838        the lengths are correct (equal to the number of opcodes).
839    
840    8.  Following on from 7, I added a similar kludge to check the length of the
841        eint vector in pcreposix.c.
842    
843    9.  Error texts for pcre_compile() are held as one long string to avoid too
844        much relocation at load time. To find a text, the string is searched,
845        counting zeros. There was no check for running off the end of the string,
846        which could happen if a new error number was added without updating the
847        string.
848    
849    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
850    
851    11. \K was not working if it appeared in an atomic group or in a group that
852        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
853        \K is "not well defined" if used in an assertion. PCRE now accepts it if
854        the assertion is positive, but not if it is negative.
855    
856    12. Change 11 fortuitously reduced the size of the stack frame used in the
857        "match()" function of pcre_exec.c by one pointer. Forthcoming
858        implementation of support for (*MARK) will need an extra pointer on the
859        stack; I have reserved it now, so that the stack frame size does not
860        decrease.
861    
862    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
863        item in branch that calls a recursion is a subroutine call - as in the
864        second branch in the above example - was incorrectly given the compile-
865        time error "recursive call could loop indefinitely" because pcre_compile()
866        was not correctly checking the subroutine for matching a non-empty string.
867    
868    14. The checks for overrunning compiling workspace could trigger after an
869        overrun had occurred. This is a "should never occur" error, but it can be
870        triggered by pathological patterns such as hundreds of nested parentheses.
871        The checks now trigger 100 bytes before the end of the workspace.
872    
873    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
874    
875    
876    Version 8.01 19-Jan-2010
877    ------------------------
878    
879    1.  If a pattern contained a conditional subpattern with only one branch (in
880        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
881        computed the wrong minimum data length (which is of course zero for such
882        subpatterns). This could cause incorrect "no match" results.
883    
884    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
885        the pattern is reset in the first branch, pcre_compile() failed with
886        "internal error: code overflow at offset...". This happened only when
887        the reset was to the original external option setting. (An optimization
888        abstracts leading options settings into an external setting, which was the
889        cause of this.)
890    
891    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
892        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
893        assertion pattern did not match (meaning that the assertion was true), it
894        was incorrectly treated as false if the SKIP had been reached during the
895        matching. This also applied to assertions used as conditions.
896    
897    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
898        assertion subpattern, including such a pattern used as a condition,
899        unpredictable results occurred, instead of the error return
900        PCRE_ERROR_DFA_UITEM.
901    
902    5.  The C++ GlobalReplace function was not working like Perl for the special
903        situation when an empty string is matched. It now does the fancy magic
904        stuff that is necessary.
905    
906    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
907        removed. (These were left over from very, very early versions of PCRE.)
908    
909    7.  Some cosmetic changes to the code to make life easier when compiling it
910        as part of something else:
911    
912        (a) Change DEBUG to PCRE_DEBUG.
913    
914        (b) In pcre_compile(), rename the member of the "branch_chain" structure
915            called "current" as "current_branch", to prevent a collision with the
916            Linux macro when compiled as a kernel module.
917    
918        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
919            prevent a collision with the Linux macro when compiled as a kernel
920            module.
921    
922    8.  In pcre_compile() there are some checks for integer overflows that used to
923        cast potentially large values to (double). This has been changed to that
924        when building, a check for int64_t is made, and if it is found, it is used
925        instead, thus avoiding the use of floating point arithmetic. (There is no
926        other use of FP in PCRE.) If int64_t is not found, the fallback is to
927        double.
928    
929    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
930        2005 (difference between two addresses compared to an unsigned value).
931    
932    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
933        custom one, because of the following reported problem in Windows:
934    
935          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
936              under Win32.
937          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
938              therefore missing the function definition.
939          - The compiler thus generates a "C" signature for the test function.
940          - The linker fails to find the "C" function.
941          - PCRE fails to configure if asked to do so against libbz2.
942    
943    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
944        messages were output:
945    
946          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
947          rerunning libtoolize, to keep the correct libtool macros in-tree.
948          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
949    
950        I have done both of these things.
951    
952    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
953        most of the time, it *can* run out if it is given a pattern that contains a
954        runaway infinite recursion. I updated the discussion in the pcrestack man
955        page.
956    
957    13. Now that we have gone to the x.xx style of version numbers, the minor
958        version may start with zero. Using 08 or 09 is a bad idea because users
959        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
960        interpreted as invalid octal numbers. I've updated the previous comment in
961        configure.ac, and also added a check that gives an error if 08 or 09 are
962        used.
963    
964    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
965        causing partial matching to fail when the end of the subject matched \W
966        in a UTF-8 pattern where \W was quantified with a minimum of 3.
967    
968    15. There were some discrepancies between the declarations in pcre_internal.h
969        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
970        their definitions. The declarations used "const uschar *" and the
971        definitions used USPTR. Even though USPTR is normally defined as "const
972        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
973        reported that: "This difference in casting confuses some C++ compilers, for
974        example, SunCC recognizes above declarations as different functions and
975        generates broken code for hbpcre." I have changed the declarations to use
976        USPTR.
977    
978    16. GNU libtool is named differently on some systems. The autogen.sh script now
979        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
980        (FreeBSD).
981    
982    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
983        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
984        comment: "Figure out how to create a longlong from a string: strtoll and
985        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
986        instance, but it only takes 2 args instead of 3!"
987    
988    18. A subtle bug concerned with back references has been fixed by a change of
989        specification, with a corresponding code fix. A pattern such as
990        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
991        refers, was giving matches when it shouldn't. For example, xa=xaaa would
992        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
993        same bug. Such groups have to be quantified to be useful, or contained
994        inside another quantified group. (If there's no repetition, the reference
995        can never match.) The problem arises because, having left the group and
996        moved on to the rest of the pattern, a later failure that backtracks into
997        the group uses the captured value from the final iteration of the group
998        rather than the correct earlier one. I have fixed this in PCRE by forcing
999        any group that contains a reference to itself to be an atomic group; that
1000        is, there cannot be any backtracking into it once it has completed. This is
1001        similar to recursive and subroutine calls.
1002    
1003    
1004    Version 8.00 19-Oct-09
1005    ----------------------
1006    
1007    1.  The table for translating pcre_compile() error codes into POSIX error codes
1008        was out-of-date, and there was no check on the pcre_compile() error code
1009        being within the table. This could lead to an OK return being given in
1010        error.
1011    
1012    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1013        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1014        in a Windows environment.
1015    
1016    3.  The pcregrep --count option prints the count for each file even when it is
1017        zero, as does GNU grep. However, pcregrep was also printing all files when
1018        --files-with-matches was added. Now, when both options are given, it prints
1019        counts only for those files that have at least one match. (GNU grep just
1020        prints the file name in this circumstance, but including the count seems
1021        more useful - otherwise, why use --count?) Also ensured that the
1022        combination -clh just lists non-zero counts, with no names.
1023    
1024    4.  The long form of the pcregrep -F option was incorrectly implemented as
1025        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1026        but it seems right to fix it, and I didn't think it was worth preserving
1027        the old behaviour.
1028    
1029    5.  The command line items --regex=pattern and --regexp=pattern were not
1030        recognized by pcregrep, which required --regex pattern or --regexp pattern
1031        (with a space rather than an '='). The man page documented the '=' forms,
1032        which are compatible with GNU grep; these now work.
1033    
1034    6.  No libpcreposix.pc file was created for pkg-config; there was just
1035        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1036    
1037    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1038        when UCP support is not needed, by modifying the Python script that
1039        generates it from Unicode data files. This should not matter if the module
1040        is correctly used as a library, but I received one complaint about 50K of
1041        unwanted data. My guess is that the person linked everything into his
1042        program rather than using a library. Anyway, it does no harm.
1043    
1044    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1045        was a minimum greater than 1 for a wide character in a possessive
1046        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1047        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1048        character. Chaos in the form of incorrect output or a compiling loop could
1049        result.
1050    
1051    9.  The restrictions on what a pattern can contain when partial matching is
1052        requested for pcre_exec() have been removed. All patterns can now be
1053        partially matched by this function. In addition, if there are at least two
1054        slots in the offset vector, the offset of the earliest inspected character
1055        for the match and the offset of the end of the subject are set in them when
1056        PCRE_ERROR_PARTIAL is returned.
1057    
1058    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1059        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1060        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1061        and may be more useful for multi-segment matching.
1062    
1063    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1064        used to be given if ever the end of the subject was reached; now it is
1065        given only if matching could not proceed because another character was
1066        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1067        string "Z", which now yields "no match" instead of "partial match". In the
1068        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1069        final character ended with (*FAIL).
1070    
1071    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1072        if the pattern had a "must contain" character that was already found in the
1073        earlier partial match, unless partial matching was again requested. For
1074        example, with the pattern /dog.(body)?/, the "must contain" character is
1075        "g". If the first part-match was for the string "dog", restarting with
1076        "sbody" failed. This bug has been fixed.
1077    
1078    13. The string returned by pcre_dfa_exec() after a partial match has been
1079        changed so that it starts at the first inspected character rather than the
1080        first character of the match. This makes a difference only if the pattern
1081        starts with a lookbehind assertion or \b or \B (\K is not supported by
1082        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1083        matching functions compatible, and I think it's the right thing to do.
1084    
1085    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1086        so that the demonstration program is easily available in environments where
1087        PCRE has not been installed from source.
1088    
1089    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1090        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1091        library.
1092    
1093    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1094        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1095        is not the first non-POSIX option to be added. Clearly some people find
1096        these options useful.
1097    
1098    17. If a caller to the POSIX matching function regexec() passes a non-zero
1099        value for nmatch with a NULL value for pmatch, the value of
1100        nmatch is forced to zero.
1101    
1102    18. RunGrepTest did not have a test for the availability of the -u option of
1103        the diff command, as RunTest does. It now checks in the same way as
1104        RunTest, and also checks for the -b option.
1105    
1106    19. If an odd number of negated classes containing just a single character
1107        interposed, within parentheses, between a forward reference to a named
1108        subpattern and the definition of the subpattern, compilation crashed with
1109        an internal error, complaining that it could not find the referenced
1110        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1111        [The bug was that it was starting one character too far in when skipping
1112        over the character class, thus treating the ] as data rather than
1113        terminating the class. This meant it could skip too much.]
1114    
1115    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1116        /g option in pcretest when the pattern contains \K, which makes it possible
1117        to have an empty string match not at the start, even when the pattern is
1118        anchored. Updated pcretest and pcredemo to use this option.
1119    
1120    21. If the maximum number of capturing subpatterns in a recursion was greater
1121        than the maximum at the outer level, the higher number was returned, but
1122        with unset values at the outer level. The correct (outer level) value is
1123        now given.
1124    
1125    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1126        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1127        make it do so. The string so far is captured, making this feature
1128        compatible with Perl.
1129    
1130    23. The tests have been re-organized, adding tests 11 and 12, to make it
1131        possible to check the Perl 5.10 features against Perl 5.10.
1132    
1133    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1134        pattern matches a fixed length string. PCRE did not allow this; now it
1135        does. Neither allows recursion.
1136    
1137    25. I finally figured out how to implement a request to provide the minimum
1138        length of subject string that was needed in order to match a given pattern.
1139        (It was back references and recursion that I had previously got hung up
1140        on.) This code has now been added to pcre_study(); it finds a lower bound
1141        to the length of subject needed. It is not necessarily the greatest lower
1142        bound, but using it to avoid searching strings that are too short does give
1143        some useful speed-ups. The value is available to calling programs via
1144        pcre_fullinfo().
1145    
1146    26. While implementing 25, I discovered to my embarrassment that pcretest had
1147        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1148        study optimizations had never been tested with that matching function.
1149        Oops. What is worse, even when it was passed study data, there was a bug in
1150        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1151        were also very few tests of studied patterns with pcre_dfa_exec().
1152    
1153    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1154        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1155        on the other side of the coin, they are no longer allowed to have different
1156        names, because these cannot be distinguished in PCRE, and this has caused
1157        confusion. (This is a difference from Perl.)
1158    
1159    28. When duplicate subpattern names are present (necessarily with different
1160        numbers, as required by 27 above), and a test is made by name in a
1161        conditional pattern, either for a subpattern having been matched, or for
1162        recursion in such a pattern, all the associated numbered subpatterns are
1163        tested, and the overall condition is true if the condition is true for any
1164        one of them. This is the way Perl works, and is also more like the way
1165        testing by number works.
1166    
1167    
1168    Version 7.9 11-Apr-09
1169    ---------------------
1170    
1171    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1172        (pcretest), all targets were linked against these libraries. This included
1173        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1174        libraries. This caused unwanted dependencies to be created. This problem
1175        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1176        pcretest is linked with readline.
1177    
1178    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1179        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1180        moved outside it again, because FALSE and TRUE are already defined in AIX,
1181        but BOOL is not.
1182    
1183    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1184        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1185    
1186    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1187        hyphens) following file names and line numbers when outputting matching
1188        lines. This is not true; no spaces are inserted. I have also clarified the
1189        wording for the --colour (or --color) option.
1190    
1191    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1192        was not coloured; this is different to GNU grep, so I have changed it to be
1193        the same.
1194    
1195    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1196        each matching line was coloured. Now it goes on to look for further matches
1197        of any of the test patterns, which is the same behaviour as GNU grep.
1198    
1199    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1200        doesn't make sense to accept an empty string match in pcregrep, so I have
1201        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1202        seems to be how GNU grep behaves.
1203    
1204    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1205        start or after a newline", because the conditional assertion was not being
1206        correctly handled. The rule now is that both the assertion and what follows
1207        in the first alternative must satisfy the test.
1208    
1209    9.  If auto-callout was enabled in a pattern with a conditional group whose
1210        condition was an assertion, PCRE could crash during matching, both with
1211        pcre_exec() and pcre_dfa_exec().
1212    
1213    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1214        used for matching.
1215    
1216    11. Unicode property support in character classes was not working for
1217        characters (bytes) greater than 127 when not in UTF-8 mode.
1218    
1219    12. Added the -M command line option to pcretest.
1220    
1221    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1222    
1223    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1224    
1225    16. Added comments and documentation about mis-use of no_arg in the C++
1226        wrapper.
1227    
1228    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1229        from Martin Jerabek that uses macro names for all relevant character and
1230        string constants.
1231    
1232    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1233        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1234        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1235        these, but not everybody uses configure.
1236    
1237    19. A conditional group that had only one branch was not being correctly
1238        recognized as an item that could match an empty string. This meant that an
1239        enclosing group might also not be so recognized, causing infinite looping
1240        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1241        with the subject "ab", where knowledge that the repeated group can match
1242        nothing is needed in order to break the loop.
1243    
1244    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1245        exec(), but without supplying a callout function, matching went wrong.
1246    
1247    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1248        leak if the size of the offset vector was greater than 30. When the vector
1249        is smaller, the saved offsets during recursion go onto a local stack
1250        vector, but for larger vectors malloc() is used. It was failing to free
1251        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1252        error, in fact).
1253    
1254    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1255        heapframe that is used only when UTF-8 support is enabled. This caused no
1256        problem, but was untidy.
1257    
1258    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1259        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1260        included within another project.
1261    
1262    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1263        slightly modified by me:
1264    
1265          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1266              not building pcregrep.
1267    
1268          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1269              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1270    
1271    25. Forward references, both numeric and by name, in patterns that made use of
1272        duplicate group numbers, could behave incorrectly or give incorrect errors,
1273        because when scanning forward to find the reference group, PCRE was not
1274        taking into account the duplicate group numbers. A pattern such as
1275        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1276    
1277    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1278        the feature of a custom pointer more persuasive (as requested by a user).
1279    
1280    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1281        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1282        pre-defined.
1283    
1284    28. Added support for (*UTF8) at the start of a pattern.
1285    
1286    29. Arrange for flags added by the "release type" setting in CMake to be shown
1287        in the configuration summary.
1288    
1289    
1290    Version 7.8 05-Sep-08
1291    ---------------------
1292    
1293    1.  Replaced UCP searching code with optimized version as implemented for Ad
1294        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1295        stage table and inline lookup instead of a function, giving speed ups of 2
1296        to 5 times on some simple patterns that I tested. Permission was given to
1297        distribute the MultiStage2.py script that generates the tables (it's not in
1298        the tarball, but is in the Subversion repository).
1299    
1300    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1301        scripts.
1302    
1303    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1304        a group with a zero qualifier. The result of the study could be incorrect,
1305        or the function might crash, depending on the pattern.
1306    
1307    4.  Caseless matching was not working for non-ASCII characters in back
1308        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1309        It now works when Unicode Property Support is available.
1310    
1311    5.  In pcretest, an escape such as \x{de} in the data was always generating
1312        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1313        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1314        truncation.
1315    
1316    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1317    
1318    7.  Added two (int) casts to pcregrep when printing the difference of two
1319        pointers, in case they are 64-bit values.
1320    
1321    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1322        test 2 if it fails.
1323    
1324    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1325        and a #define of that name to empty if it is not externally set. This is to
1326        allow users of MSVC to set it if necessary.
1327    
1328    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1329        the convenience functions in the pcre_get.c source file.
1330    
1331    11. An option change at the start of a pattern that had top-level alternatives
1332        could cause overwriting and/or a crash. This command provoked a crash in
1333        some environments:
1334    
1335          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1336    
1337        This potential security problem was recorded as CVE-2008-2371.
1338    
1339    12. For a pattern where the match had to start at the beginning or immediately
1340        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1341        pcre_dfa_exec() could read past the end of the passed subject if there was
1342        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1343        pcretest so that it places the subject at the end of its malloc-ed buffer.
1344    
1345    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1346        exec() might read past the end of the data buffer in UTF-8 mode.
1347    
1348    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1349        the data contained the byte 0x85 as part of a UTF-8 character within its
1350        first line. This applied both to normal and DFA matching.
1351    
1352    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1353        /^[^d]*?$/8 failed to match "abc".
1354    
1355    16. Added a missing copyright notice to pcrecpp_internal.h.
1356    
1357    17. Make it more clear in the documentation that values returned from
1358        pcre_exec() in ovector are byte offsets, not character counts.
1359    
1360    18. Tidied a few places to stop certain compilers from issuing warnings.
1361    
1362    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1363        supplied by Stefan Weber. I made a further small update for 7.8 because
1364        there is a change of source arrangements: the pcre_searchfuncs.c module is
1365        replaced by pcre_ucd.c.
1366    
1367    
1368    Version 7.7 07-May-08
1369    ---------------------
1370    
1371    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1372        a string to a long long, pretend we don't even have a long long." This is
1373        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1374    
1375    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1376        pre-7.6 versions, which defined a global no_arg variable instead of putting
1377        it in the RE class. (See also #8 below.)
1378    
1379    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1380        Lopes.
1381    
1382    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1383    
1384        (1) The include/exclude patterns were being applied to the whole pathnames
1385            of files, instead of just to the final components.
1386    
1387        (2) If there was more than one level of directory, the subdirectories were
1388            skipped unless they satisfied the include/exclude conditions. This is
1389            inconsistent with GNU grep (and could even be seen as contrary to the
1390            pcregrep specification - which I improved to make it absolutely clear).
1391            The action now is always to scan all levels of directory, and just
1392            apply the include/exclude patterns to regular files.
1393    
1394    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1395        --exclude_dir in the tests to avoid scanning .svn directories.
1396    
1397    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1398        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1399        doesn't support NULs in patterns.
1400    
1401    7.  Added some missing "const"s to declarations of static tables in
1402        pcre_compile.c and pcre_dfa_exec.c.
1403    
1404    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1405        caused by fix #2  above. (Subsequently also a second patch to fix the
1406        first patch. And a third patch - this was a messy problem.)
1407    
1408    9.  Applied Craig's patch to remove the use of push_back().
1409    
1410    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1411        matching function regexec().
1412    
1413    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1414        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1415        references. PCRE supports relative numbers with this syntax (I don't think
1416        Oniguruma does).
1417    
1418    12. Previously, a group with a zero repeat such as (...){0} was completely
1419        omitted from the compiled regex. However, this means that if the group
1420        was called as a subroutine from elsewhere in the pattern, things went wrong
1421        (an internal error was given). Such groups are now left in the compiled
1422        pattern, with a new opcode that causes them to be skipped at execution
1423        time.
1424    
1425    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1426        to the way PCRE behaves:
1427    
1428        (a) A lone ] character is dis-allowed (Perl treats it as data).
1429    
1430        (b) A back reference to an unmatched subpattern matches an empty string
1431            (Perl fails the current match path).
1432    
1433        (c) A data ] in a character class must be notated as \] because if the
1434            first data character in a class is ], it defines an empty class. (In
1435            Perl it is not possible to have an empty class.) The empty class []
1436            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1437            The negative empty class [^] matches any one character, independently
1438            of the DOTALL setting.
1439    
1440    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1441        non-existent subpattern following a character class starting with ']' and
1442        containing () gave an internal compiling error instead of "reference to
1443        non-existent subpattern". Fortunately, when the pattern did exist, the
1444        compiled code was correct. (When scanning forwards to check for the
1445        existencd of the subpattern, it was treating the data ']' as terminating
1446        the class, so got the count wrong. When actually compiling, the reference
1447        was subsequently set up correctly.)
1448    
1449    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1450        it was being rejected as not supported by pcre_dfa_exec(), even though
1451        other assertions are supported. I have made pcre_dfa_exec() support
1452        (*FAIL).
1453    
1454    16. The implementation of 13c above involved the invention of a new opcode,
1455        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1456        cannot be changed at match time, I realized I could make a small
1457        improvement to matching performance by compiling OP_ALLANY instead of
1458        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1459        on the OP_ANY path.
1460    
1461    17. Compiling pcretest on Windows with readline support failed without the
1462        following two fixes: (1) Make the unistd.h include conditional on
1463        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1464    
1465    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1466        ncurses library to be included for pcretest when ReadLine support is
1467        requested, but also to allow for it to be overridden. This patch came from
1468        Daniel Bergström.
1469    
1470    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1471        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1472        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1473        spotting this.
1474    
1475    
1476    Version 7.6 28-Jan-08
1477    ---------------------
1478    
1479    1.  A character class containing a very large number of characters with
1480        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1481        overflow.
1482    
1483    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1484        HAVE_LONG_LONG is not defined.
1485    
1486    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1487        bring them up to date and include new features. This patch includes:
1488    
1489        - Fixed PH's badly added libz and libbz2 support.
1490        - Fixed a problem with static linking.
1491        - Added pcredemo. [But later removed - see 7 below.]
1492        - Fixed dftables problem and added an option.
1493        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1494            HAVE_LONG_LONG.
1495        - Added readline support for pcretest.
1496        - Added an listing of the option settings after cmake has run.
1497    
1498    4.  A user submitted a patch to Makefile that makes it easy to create
1499        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1500        Makefile.am that cause it to include this special target, without
1501        affecting anything else. Note that the same mingw target plus all
1502        the other distribution libraries and programs are now supported
1503        when configuring with CMake (see 6 below) instead of with
1504        Configure/Make.
1505    
1506    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1507        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1508        exported in the Windows port". It has not yet been confirmed that the patch
1509        solves the problem, but it does no harm.
1510    
1511    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1512        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1513        with CMake, and also correct the comment about stack recursion.
1514    
1515    7.  Remove the automatic building of pcredemo from the ./configure system and
1516        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1517        of a program that users should build themselves after PCRE is installed, so
1518        building it automatically is not really right. What is more, it gave
1519        trouble in some build environments.
1520    
1521    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1522    
1523    
1524    Version 7.5 10-Jan-08
1525    ---------------------
1526    
1527    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1528        values in parens when parsing an RE using the C++ wrapper."
1529    
1530    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1531        Characters greater than 255 were excluded from the class instead of being
1532        included.
1533    
1534    3.  The same bug as (2) above applied to negated POSIX classes such as
1535        [:^space:].
1536    
1537    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1538        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1539        I have changed it.
1540    
1541    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1542        first named subpattern) and a construct such as (?&a) would reference the
1543        first named subpattern whose name started with "a" (in other words, the
1544        length check was missing). Both these problems are fixed. "Subpattern name
1545        expected" is now given for (?&) (a zero-length name), and this patch also
1546        makes it give the same error for \k'' (previously it complained that that
1547        was a reference to a non-existent subpattern).
1548    
1549    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1550        this is right because (?- can be followed by option settings as well as by
1551        digits. I have, however, made the messages clearer.
1552    
1553    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1554        than the number used in the conditional) now cause a compile-time error.
1555        This is actually not compatible with Perl, which accepts such patterns, but
1556        treats the conditional as always being FALSE (as PCRE used to), but it
1557        seems to me that giving a diagnostic is better.
1558    
1559    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1560        and messages.
1561    
1562    9.  Fix two occurrences of "backslash" in comments that should have been
1563        "backspace".
1564    
1565    10. Remove two redundant lines of code that can never be obeyed (their function
1566        was moved elsewhere).
1567    
1568    11. The program that makes PCRE's Unicode character property table had a bug
1569        which caused it to generate incorrect table entries for sequences of
1570        characters that have the same character type, but are in different scripts.
1571        It amalgamated them into a single range, with the script of the first of
1572        them. In other words, some characters were in the wrong script. There were
1573        thirteen such cases, affecting characters in the following ranges:
1574    
1575          U+002b0 - U+002c1
1576          U+0060c - U+0060d
1577          U+0061e - U+00612
1578          U+0064b - U+0065e
1579          U+0074d - U+0076d
1580          U+01800 - U+01805
1581          U+01d00 - U+01d77
1582          U+01d9b - U+01dbf
1583          U+0200b - U+0200f
1584          U+030fc - U+030fe
1585          U+03260 - U+0327f
1586          U+0fb46 - U+0fbb1
1587          U+10450 - U+1049d
1588    
1589    12. The -o option (show only the matching part of a line) for pcregrep was not
1590        compatible with GNU grep in that, if there was more than one match in a
1591        line, it showed only the first of them. It now behaves in the same way as
1592        GNU grep.
1593    
1594    13. If the -o and -v options were combined for pcregrep, it printed a blank
1595        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1596        does the same. The return code can be used to tell if there were any
1597        non-matching lines.
1598    
1599    14. Added --file-offsets and --line-offsets to pcregrep.
1600    
1601    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1602        infinitely looping recursion. The bug was that positive lookaheads were not
1603        being skipped when checking for a possible empty match (negative lookaheads
1604        and both kinds of lookbehind were skipped).
1605    
1606    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1607        inclusion of <windows.h> to before rather than after the definition of
1608        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1609    
1610    17. Specifying a possessive quantifier with a specific limit for a Unicode
1611        character property caused pcre_compile() to compile bad code, which led at
1612        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1613        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1614        caused the error; without that there was no problem.
1615    
1616    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1617    
1618    19. Added --enable-pcretest-libreadline.
1619    
1620    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1621        RE::GlobalReplace(). As a result, the number of replacements returned was
1622        double what it should be. I removed one of the increments, but Craig sent a
1623        later patch that removed the other one (the right fix) and added unit tests
1624        that check the return values (which was not done before).
1625    
1626    21. Several CMake things:
1627    
1628        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1629            the names libpcre and libpcreposix, not just pcre and pcreposix.
1630    
1631        (2) The above change means that pcretest and pcregrep are now correctly
1632            linked with the newly-built libraries, not previously installed ones.
1633    
1634        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1635    
1636    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1637        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1638        UTF-8 newline character). The key issue is that the pattern starts .*;
1639        this means that the match must be either at the beginning, or after a
1640        newline. The bug was in the code for advancing after a failed match and
1641        checking that the new position followed a newline. It was not taking
1642        account of UTF-8 characters correctly.
1643    
1644    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1645        character classes. PCRE was not treating the sequence [:...:] as a
1646        character class unless the ... were all letters. Perl, however, seems to
1647        allow any characters between [: and :], though of course it rejects as
1648        unknown any "names" that contain non-letters, because all the known class
1649        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1650        for example, whereas PCRE did not - it did not recognize a POSIX character
1651        class. This seemed a bit dangerous, so the code has been changed to be
1652        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1653        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1654        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1655        Perl does, and where it didn't before.
1656    
1657    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1658        Windows environments %n is disabled by default.
1659    
1660    
1661    Version 7.4 21-Sep-07
1662    ---------------------
1663    
1664    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1665        means that a class such as [\s] counted as "explicit reference to CR or
1666        LF". That isn't really right - the whole point of the change was to try to
1667        help when there was an actual mention of one of the two characters. So now
1668        the change happens only if \r or \n (or a literal CR or LF) character is
1669        encountered.
1670    
1671    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1672        of both had grown to the point where there were only 3 bits left.
1673        Fortunately, there was spare space in the data structure, and so I have
1674        moved the internal flags into a new 16-bit field to free up more option
1675        bits.
1676    
1677    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1678        but did not set the internal JCHANGED flag - either of these is enough to
1679        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1680        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1681        start sets both bits.
1682    
1683    4.  Added options (at build time, compile time, exec time) to change \R from
1684        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1685    
1686    5.  doc/pcresyntax.html was missing from the distribution.
1687    
1688    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1689        compatibility, even though it is no longer used.
1690    
1691    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1692        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1693        windows.h file is present (where different names are used). [This was
1694        reversed later after testing - see 16 below.]
1695    
1696    8.  Changed all #include <config.h> to #include "config.h". There were also
1697        some further <pcre.h> cases that I changed to "pcre.h".
1698    
1699    9.  When pcregrep was used with the --colour option, it missed the line ending
1700        sequence off the lines that it output.
1701    
1702    10. It was pointed out to me that arrays of string pointers cause lots of
1703        relocations when a shared library is dynamically loaded. A technique of
1704        using a single long string with a table of offsets can drastically reduce
1705        these. I have refactored PCRE in four places to do this. The result is
1706        dramatic:
1707    
1708          Originally:                          290
1709          After changing UCP table:            187
1710          After changing error message table:   43
1711          After changing table of "verbs"       36
1712          After changing table of Posix names   22
1713    
1714        Thanks to the folks working on Gregex for glib for this insight.
1715    
1716    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1717        unicode-properties was also set.
1718    
1719    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1720    
1721    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1722        checked only for CRLF.
1723    
1724    14. Added casts to pcretest.c to avoid compiler warnings.
1725    
1726    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1727    
1728    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1729        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1730        entirely. This removes changes made in 7 above.
1731    
1732    17. The CMake files have been updated, and there is now more information about
1733        building with CMake in the NON-UNIX-USE document.
1734    
1735    
1736    Version 7.3 28-Aug-07
1737    ---------------------
1738    
1739     1. In the rejigging of the build system that eventually resulted in 7.1, the
1740        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1741        brackets there is not right, since it causes compilers to look for an
1742        installed pcre.h, not the version that is in the source that is being
1743        compiled (which of course may be different). I have changed it back to:
1744    
1745          #include "pcre.h"
1746    
1747        I have a vague recollection that the change was concerned with compiling in
1748        different directories, but in the new build system, that is taken care of
1749        by the VPATH setting the Makefile.
1750    
1751     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1752        when the subject happened to end in the byte 0x85 (e.g. if the last
1753        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1754        characters but of course it shouldn't be taken as a newline when it is part
1755        of another character. The bug was that, for an unlimited repeat of . in
1756        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1757        characters when looking for a newline.
1758    
1759     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1760    
1761     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1762        in debug output.
1763    
1764     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1765        long printing in the pcrecpp unittest when running under MinGW.
1766    
1767     6. ESC_K was left out of the EBCDIC table.
1768    
1769     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1770        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1771        limit also applies to "virtual nesting" when a pattern is recursive, and in
1772        this case 1000 isn't so big. I have been able to remove this limit at the
1773        expense of backing off one optimization in certain circumstances. Normally,
1774        when pcre_exec() would call its internal match() function recursively and
1775        immediately return the result unconditionally, it uses a "tail recursion"
1776        feature to save stack. However, when a subpattern that can match an empty
1777        string has an unlimited repetition quantifier, it no longer makes this
1778        optimization. That gives it a stack frame in which to save the data for
1779        checking that an empty string has been matched. Previously this was taken
1780        from the 1000-entry workspace that had been reserved. So now there is no
1781        explicit limit, but more stack is used.
1782    
1783     8. Applied Daniel's patches to solve problems with the import/export magic
1784        syntax that is required for Windows, and which was going wrong for the
1785        pcreposix and pcrecpp parts of the library. These were overlooked when this
1786        problem was solved for the main library.
1787    
1788     9. There were some crude static tests to avoid integer overflow when computing
1789        the size of patterns that contain repeated groups with explicit upper
1790        limits. As the maximum quantifier is 65535, the maximum group length was
1791        set at 30,000 so that the product of these two numbers did not overflow a
1792        32-bit integer. However, it turns out that people want to use groups that
1793        are longer than 30,000 bytes (though not repeat them that many times).
1794        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1795        made it possible to implement the integer overflow checks in a much more
1796        dynamic way, which I have now done. The artificial limitation on group
1797        length has been removed - we now have only the limit on the total length of
1798        the compiled pattern, which depends on the LINK_SIZE setting.
1799    
1800    10. Fixed a bug in the documentation for get/copy named substring when
1801        duplicate names are permitted. If none of the named substrings are set, the
1802        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1803        empty string.
1804    
1805    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1806        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1807        because the ] is interpreted as the first data character and the
1808        terminating ] is not found. PCRE has been made compatible with Perl in this
1809        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1810        cause memory overwriting.
1811    
1812    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1813        string has been matched (to stop an infinite loop). It was not recognizing
1814        a conditional subpattern that could match an empty string if that
1815        subpattern was within another subpattern. For example, it looped when
1816        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1817        condition was not nested. This bug has been fixed.
1818    
1819    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1820        past the start of the subject in the presence of bytes with the top bit
1821        set, for example "\x8aBCD".
1822    
1823    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1824        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1825    
1826    14. Optimized (?!) to (*FAIL).
1827    
1828    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1829        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1830        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1831        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1832        does: it's just the validity check that is more restrictive.
1833    
1834    16. Inserted checks for integer overflows during escape sequence (backslash)
1835        processing, and also fixed erroneous offset values for syntax errors during
1836        backslash processing.
1837    
1838    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1839        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1840    
1841    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1842        caused an overrun.
1843    
1844    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1845        something other than just ASCII characters) inside a group that had an
1846        unlimited repeat caused a loop at compile time (while checking to see
1847        whether the group could match an empty string).
1848    
1849    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1850        [\P{Any}] did so. (Error in the code for printing property names.)
1851    
1852    21. An orphan \E inside a character class could cause a crash.
1853    
1854    22. A repeated capturing bracket such as (A)? could cause a wild memory
1855        reference during compilation.
1856    
1857    23. There are several functions in pcre_compile() that scan along a compiled
1858        expression for various reasons (e.g. to see if it's fixed length for look
1859        behind). There were bugs in these functions when a repeated \p or \P was
1860        present in the pattern. These operators have additional parameters compared
1861        with \d, etc, and these were not being taken into account when moving along
1862        the compiled data. Specifically:
1863    
1864        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1865            length.
1866    
1867        (b) An item such as \pL+ within a repeated group could cause crashes or
1868            loops.
1869    
1870        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1871            "reference to non-existent subpattern" error.
1872    
1873        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1874    
1875    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1876        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1877    
1878    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1879    
1880    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1881        character were causing crashes (broken optimization).
1882    
1883    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1884        \p or \P) caused a compile-time loop.
1885    
1886    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1887        break. For example, the unstudied pattern [\r\n]A does not match the string
1888        "\r\nA" because change 7.0/46 below moves the current point on by two
1889        characters after failing to match at the start. However, the pattern \nA
1890        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1891        the same is true. There doesn't seem any very clean way out of this, but
1892        what I have chosen to do makes the common cases work: PCRE now takes note
1893        of whether there can be an explicit match for \r or \n anywhere in the
1894        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1895        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1896        pattern has explicit CR or LF references.
1897    
1898    29. Added (*CR) etc for changing newline setting at start of pattern.
1899    
1900    
1901    Version 7.2 19-Jun-07
1902    ---------------------
1903    
1904     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1905        which is apparently normally available under Windows.
1906    
1907     2. Re-jig the pcregrep tests with different newline settings in an attempt
1908        to make them independent of the local environment's newline setting.
1909    
1910     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1911    
1912     4. Some of the "internals" tests were previously cut out when the link size
1913        was not 2, because the output contained actual offsets. The recent new
1914        "Z" feature of pcretest means that these can be cut out, making the tests
1915        usable with all link sizes.
1916    
1917     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1918        stack recursion. This gives a massive performance boost under BSD, but just
1919        a small improvement under Linux. However, it saves one field in the frame
1920        in all cases.
1921    
1922     6. Added more features from the forthcoming Perl 5.10:
1923    
1924        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1925            recursion call. It refers to the nth most recently opened parentheses.
1926    
1927        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1928            to be opened parentheses.
1929    
1930        (c) Conditions that refer to capturing parentheses can be specified
1931            relatively, for example, (?(-2)... or (?(+3)...
1932    
1933        (d) \K resets the start of the current match so that everything before
1934            is not part of it.
1935    
1936        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1937    
1938        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1939            reference syntax.
1940    
1941        (g) (?| introduces a group in which the numbering of parentheses in each
1942            alternative starts with the same number.
1943    
1944        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1945    
1946     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1947        PCRE_INFO_JCHANGED.
1948    
1949     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1950        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1951        for detecting groups that can match an empty string.
1952    
1953     9. A pattern with a very large number of alternatives (more than several
1954        hundred) was running out of internal workspace during the pre-compile
1955        phase, where pcre_compile() figures out how much memory will be needed. A
1956        bit of new cunning has reduced the workspace needed for groups with
1957        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1958        workspace instead of running out of the 4096 that are available.
1959    
1960    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1961    
1962    11. Applied patch from Google to remove an optimization that didn't quite work.
1963        The report of the bug said:
1964    
1965          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1966          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1967          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1968    
1969    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1970        it matched the wrong number of bytes.
1971    
1972    
1973    Version 7.1 24-Apr-07
1974    ---------------------
1975    
1976     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1977        that is more "standard", making use of automake and other Autotools. There
1978        is some re-arrangement of the files and adjustment of comments consequent
1979        on this.
1980    
1981     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1982        for recursive directory scanning broke on some systems because the files
1983        are not scanned in any specific order and on different systems the order
1984        was different. A call to "sort" has been inserted into RunGrepTest for the
1985        approprate test as a short-term fix. In the longer term there may be an
1986        alternative.
1987    
1988     3. I had an email from Eric Raymond about problems translating some of PCRE's
1989        man pages to HTML (despite the fact that I distribute HTML pages, some
1990        people do their own conversions for various reasons). The problems
1991        concerned the use of low-level troff macros .br and .in. I have therefore
1992        removed all such uses from the man pages (some were redundant, some could
1993        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1994        HTML has been updated to handle .nf/.fi and to complain if it encounters
1995        .br or .in.
1996    
1997     4. Updated comments in configure.ac that get placed in config.h.in and also
1998        arranged for config.h to be included in the distribution, with the name
1999        config.h.generic, for the benefit of those who have to compile without
2000        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2001    
2002     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2003        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2004        makevp.bat for the current PCRE, using the additional files
2005        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2006    
2007     6. A Windows user reported a minor discrepancy with test 2, which turned out
2008        to be caused by a trailing space on an input line that had got lost in his
2009        copy. The trailing space was an accident, so I've just removed it.
2010    
2011     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2012        that is needed.
2013    
2014     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2015        as "const" (a) because they are and (b) because it helps the PHP
2016        maintainers who have recently made a script to detect big data structures
2017        in the php code that should be moved to the .rodata section. I remembered
2018        to update Builducptable as well, so it won't revert if ucptable.h is ever
2019        re-created.
2020    
2021     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2022        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2023        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2024        support is not required. This saves 1.5-2K of code, which is important in
2025        some applications.
2026    
2027        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2028        so as not to refer to the tables, even though these functions will never be
2029        called when UTF-8 support is disabled. Otherwise there are problems with a
2030        shared library.
2031    
2032    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2033    
2034        (a) It was defining its arguments as char * instead of void *.
2035    
2036        (b) It was assuming that all moves were upwards in memory; this was true
2037            a long time ago when I wrote it, but is no longer the case.
2038    
2039        The emulated memove() is provided for those environments that have neither
2040        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2041        is clearly not the case, as these two bugs were recently reported.
2042    
2043    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2044        and Detrail to create the HTML documentation, the .txt form of the man
2045        pages, and it removes trailing spaces from listed files. It also creates
2046        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2047        case, it wraps all the #defines with #ifndefs. This script should be run
2048        before "make dist".
2049    
2050    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2051        with Unicode property support.
2052    
2053        (a) For a maximizing quantifier, if the two different cases of the
2054            character were of different lengths in their UTF-8 codings (there are
2055            some cases like this - I found 11), and the matching function had to
2056            back up over a mixture of the two cases, it incorrectly assumed they
2057            were both the same length.
2058    
2059        (b) When PCRE was configured to use the heap rather than the stack for
2060            recursion during matching, it was not correctly preserving the data for
2061            the other case of a UTF-8 character when checking ahead for a match
2062            while processing a minimizing repeat. If the check also involved
2063            matching a wide character, but failed, corruption could cause an
2064            erroneous result when trying to check for a repeat of the original
2065            character.
2066    
2067    13. Some tidying changes to the testing mechanism:
2068    
2069        (a) The RunTest script now detects the internal link size and whether there
2070            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2071            values substituted by "configure". (The RunGrepTest script already did
2072            this for UTF-8.) The configure.ac script no longer substitutes the
2073            relevant variables.
2074    
2075        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2076            with length and offset values. This means that the output is different
2077            for different internal link sizes. Test 2 is skipped for link sizes
2078            other than 2 because of this, bypassing the problem. Unfortunately,
2079            there was also a test in test 3 (the locale tests) that used /B and
2080            failed for link sizes other than 2. Rather than cut the whole test out,
2081            I have added a new /Z option to pcretest that replaces the length and
2082            offset values with spaces. This is now used to make test 3 independent
2083            of link size. (Test 2 will be tidied up later.)
2084    
2085    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2086        segmentation fault instead of returning the appropriate error message.
2087    
2088    15. In multiline mode when the newline sequence was set to "any", the pattern
2089        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2090        This doesn't seem right; it now treats the CRLF combination as the line
2091        ending, and so does not match in that case. It's only a pattern such as ^$
2092        that would hit this one: something like ^ABC$ would have failed after \r
2093        and then tried again after \r\n.
2094    
2095    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2096        in an attempt to make files that differ only in their line terminators
2097        compare equal. This works on Linux.
2098    
2099    17. Under certain error circumstances pcregrep might try to free random memory
2100        as it exited. This is now fixed, thanks to valgrind.
2101    
2102    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2103        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2104        was because its rules for how to advance for /g after matching an empty
2105        string at the end of a line did not allow for this case. They now check for
2106        it specially.
2107    
2108    20. pcretest is supposed to handle patterns and data of any length, by
2109        extending its buffers when necessary. It was getting this wrong when the
2110        buffer for a data line had to be extended.
2111    
2112    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2113        CRLF as a newline sequence.
2114    
2115    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2116        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2117        I have nevertheless tidied it up.
2118    
2119    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2120    
2121    24. Added a man page for pcre-config.
2122    
2123    
2124    Version 7.0 19-Dec-06
2125    ---------------------
2126    
2127     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2128        moving to gcc 4.1.1.
2129    
2130     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2131        sys/time.h, which is documented as needed for this function. It doesn't
2132        seem to matter on Linux, but it showed up on some releases of OS X.
2133    
2134     3. It seems that there are systems where bytes whose values are greater than
2135        127 match isprint() in the "C" locale. The "C" locale should be the
2136        default when a C program starts up. In most systems, only ASCII printing
2137        characters match isprint(). This difference caused the output from pcretest
2138        to vary, making some of the tests fail. I have changed pcretest so that:
2139    
2140        (a) When it is outputting text in the compiled version of a pattern, bytes
2141            other than 32-126 are always shown as hex escapes.
2142    
2143        (b) When it is outputting text that is a matched part of a subject string,
2144            it does the same, unless a different locale has been set for the match
2145            (using the /L modifier). In this case, it uses isprint() to decide.
2146    
2147     4. Fixed a major bug that caused incorrect computation of the amount of memory
2148        required for a compiled pattern when options that changed within the
2149        pattern affected the logic of the preliminary scan that determines the
2150        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2151        that the computed length was too small. The symptoms of this bug were
2152        either the PCRE error "internal error: code overflow" from pcre_compile(),
2153        or a glibc crash with a message such as "pcretest: free(): invalid next
2154        size (fast)". Examples of patterns that provoked this bug (shown in
2155        pcretest format) are:
2156    
2157          /(?-x: )/x
2158          /(?x)(?-x: \s*#\s*)/
2159          /((?i)[\x{c0}])/8
2160          /(?i:[\x{c0}])/8
2161    
2162        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2163        is now done differently.
2164    
2165     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2166        wrapper classes; (b) implement a new function in the C++ scanner that is
2167        more efficient than the old way of doing things because it avoids levels of
2168        recursion in the regex matching; (c) add a paragraph to the documentation
2169        for the FullMatch() function.
2170    
2171     6. The escape sequence \n was being treated as whatever was defined as
2172        "newline". Not only was this contrary to the documentation, which states
2173        that \n is character 10 (hex 0A), but it also went horribly wrong when
2174        "newline" was defined as CRLF. This has been fixed.
2175    
2176     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2177        was being set to -1 for the "end of line" case (supposedly a value that no
2178        character can have). Though this value is never used (the check for end of
2179        line is "zero bytes in current character"), it caused compiler complaints.
2180        I've changed it to 0xffffffff.
2181    
2182     8. In pcre_version.c, the version string was being built by a sequence of
2183        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2184        string (as it is for production releases) called a macro with an empty
2185        argument. The C standard says the result of this is undefined. The gcc
2186        compiler treats it as an empty string (which was what was wanted) but it is
2187        reported that Visual C gives an error. The source has been hacked around to
2188        avoid this problem.
2189    
2190     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2191        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2192        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2193        of them did).
2194    
2195    10. Originally, pcretest opened its input and output without "b"; then I was
2196        told that "b" was needed in some environments, so it was added for release
2197        5.0 to both the input and output. (It makes no difference on Unix-like
2198        systems.) Later I was told that it is wrong for the input on Windows. I've
2199        now abstracted the modes into two macros, to make it easier to fiddle with
2200        them, and removed "b" from the input mode under Windows.
2201    
2202    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2203    
2204    12. Added -help and --help to pcretest as an official way of being reminded
2205        of the options.
2206    
2207    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2208        and pcrecpp.cc because they annoy compilers at high warning levels.
2209    
2210    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2211    
2212    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2213        scripts are not C programs :-) and which was not noticed because it works
2214        on Linux.
2215    
2216    16. pcretest is supposed to handle any length of pattern and data line (as one
2217        line or as a continued sequence of lines) by extending its input buffer if
2218        necessary. This feature was broken for very long pattern lines, leading to
2219        a string of junk being passed to pcre_compile() if the pattern was longer
2220        than about 50K.
2221    
2222    17. I have done a major re-factoring of the way pcre_compile() computes the
2223        amount of memory needed for a compiled pattern. Previously, there was code
2224        that made a preliminary scan of the pattern in order to do this. That was
2225        OK when PCRE was new, but as the facilities have expanded, it has become
2226        harder and harder to keep it in step with the real compile phase, and there
2227        have been a number of bugs (see for example, 4 above). I have now found a
2228        cunning way of running the real compile function in a "fake" mode that
2229        enables it to compute how much memory it would need, while actually only
2230        ever using a few hundred bytes of working memory and without too many
2231        tests of the mode. This should make future maintenance and development
2232        easier. A side effect of this work is that the limit of 200 on the nesting
2233        depth of parentheses has been removed (though this was never a serious
2234        limitation, I suspect). However, there is a downside: pcre_compile() now
2235        runs more slowly than before (30% or more, depending on the pattern). I
2236        hope this isn't a big issue. There is no effect on runtime performance.
2237    
2238    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2239        newline (only possible for the last line of a file) and it was a
2240        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2241    
2242    19. Added additional timing features to pcretest. (1) The -tm option now times
2243        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2244        separate command line item, by a number that specifies the number of
2245        repeats to use when timing. The default is 50000; this gives better
2246        precision, but takes uncomfortably long for very large patterns.
2247    
2248    20. Extended pcre_study() to be more clever in cases where a branch of a
2249        subpattern has no definite first character. For example, (a*|b*)[cd] would
2250        previously give no result from pcre_study(). Now it recognizes that the
2251        first character must be a, b, c, or d.
2252    
2253    21. There was an incorrect error "recursive call could loop indefinitely" if
2254        a subpattern (or the entire pattern) that was being tested for matching an
2255        empty string contained only one non-empty item after a nested subpattern.
2256        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2257        incorrectly, because the \d was being skipped in the check.
2258    
2259    22. The pcretest program now has a new pattern option /B and a command line
2260        option -b, which is equivalent to adding /B to every pattern. This causes
2261        it to show the compiled bytecode, without the additional information that
2262        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2263        is the same as /B/I).
2264    
2265    23. A new optimization is now able automatically to treat some sequences such
2266        as a*b as a*+b. More specifically, if something simple (such as a character
2267        or a simple class like \d) has an unlimited quantifier, and is followed by
2268        something that cannot possibly match the quantified thing, the quantifier
2269        is automatically "possessified".
2270    
2271    24. A recursive reference to a subpattern whose number was greater than 39
2272        went wrong under certain circumstances in UTF-8 mode. This bug could also
2273        have affected the operation of pcre_study().
2274    
2275    25. Realized that a little bit of performance could be had by replacing
2276        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2277    
2278    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2279    
2280    27. Possessive quantifiers such as a++ were previously implemented by turning
2281        them into atomic groups such as ($>a+). Now they have their own opcodes,
2282        which improves performance. This includes the automatically created ones
2283        from 23 above.
2284    
2285    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2286        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2287        the first matched character to be a colon. This applied both to named and
2288        numbered groups.
2289    
2290    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2291    
2292    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2293        building PCRE on Windows easier, so I have included it in the distribution.
2294    
2295    31. There is now a check in pcretest against a ridiculously large number being
2296        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2297        loop, the loop is abandoned.
2298    
2299    32. Forward references to subpatterns in conditions such as (?(2)...) where
2300        subpattern 2 is defined later cause pcre_compile() to search forwards in
2301        the pattern for the relevant set of parentheses. This search went wrong
2302        when there were unescaped parentheses in a character class, parentheses
2303        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2304    
2305    33. "Subroutine" calls and backreferences were previously restricted to
2306        referencing subpatterns earlier in the regex. This restriction has now
2307        been removed.
2308    
2309    34. Added a number of extra features that are going to be in Perl 5.10. On the
2310        whole, these are just syntactic alternatives for features that PCRE had
2311        previously implemented using the Python syntax or my own invention. The
2312        other formats are all retained for compatibility.
2313    
2314        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2315            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2316            also .NET compatible.
2317    
2318        (b) A recursion or subroutine call to a named group can now be defined as
2319            (?&name) as well as (?P>name).
2320    
2321        (c) A backreference to a named group can now be defined as \k<name> or
2322            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2323            5.10, are also .NET compatible.
2324    
2325        (d) A conditional reference to a named group can now use the syntax
2326            (?(<name>) or (?('name') as well as (?(name).
2327    
2328        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2329            groups (named and numbered) that are never evaluated inline, but can be
2330            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2331            is always false. There may be only one alternative in such a group.
2332    
2333        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2334            as the simple (?(R). The condition is true only if the most recent
2335            recursion is that of the given number or name. It does not search out
2336            through the entire recursion stack.
2337    
2338        (g) The escape \gN or \g{N} has been added, where N is a positive or
2339            negative number, specifying an absolute or relative reference.
2340    
2341    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2342        some "unreachable code" warnings.
2343    
2344    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2345        things, this adds five new scripts.
2346    
2347    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2348        There were also incompatibilities regarding the handling of \Q..\E inside
2349        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2350        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2351    
2352    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2353        matches an empty string, and forcibly breaks the loop. There were bugs in
2354        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2355        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2356        separate and independent bugs (that affected different cases) have been
2357        fixed.
2358    
2359    39. Refactored the code to abolish the use of different opcodes for small
2360        capturing bracket numbers. This is a tidy that I avoided doing when I
2361        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2362        The new approach is not only tidier, it makes it possible to reduce the
2363        memory needed to fix the previous bug (38).
2364    
2365    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2366        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2367        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2368        mode.
2369    
2370    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2371        report.
2372    
2373    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2374        copy construction and assignment in the C++ wrapper.
2375    
2376    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2377        couple of bugs that could have given wrong results in the "--newline=crlf"
2378        case.
2379    
2380    44. Added a number of casts and did some reorganization of signed/unsigned int
2381        variables following suggestions from Dair Grant. Also renamed the variable
2382        "this" as "item" because it is a C++ keyword.
2383    
2384    45. Arranged for dftables to add
2385    
2386          #include "pcre_internal.h"
2387    
2388        to pcre_chartables.c because without it, gcc 4.x may remove the array
2389        definition from the final binary if PCRE is built into a static library and
2390        dead code stripping is activated.
2391    
2392    46. For an unanchored pattern, if a match attempt fails at the start of a
2393        newline sequence, and the newline setting is CRLF or ANY, and the next two
2394        characters are CRLF, advance by two characters instead of one.
2395    
2396    
2397    Version 6.7 04-Jul-06
2398    ---------------------
2399    
2400     1. In order to handle tests when input lines are enormously long, pcretest has
2401        been re-factored so that it automatically extends its buffers when
2402        necessary. The code is crude, but this _is_ just a test program. The
2403        default size has been increased from 32K to 50K.
2404    
2405     2. The code in pcre_study() was using the value of the re argument before
2406        testing it for NULL. (Of course, in any sensible call of the function, it
2407        won't be NULL.)
2408    
2409     3. The memmove() emulation function in pcre_internal.h, which is used on
2410        systems that lack both memmove() and bcopy() - that is, hardly ever -
2411        was missing a "static" storage class specifier.
2412    
2413     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2414        containing an extended class (one that cannot be represented by a bitmap
2415        because it contains high-valued characters or Unicode property items, e.g.
2416        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2417        pattern, but PCRE should not loop if you do not (it no longer does).
2418        [Detail: two cases were found: (a) a repeated subpattern containing an
2419        extended class; (b) a recursive reference to a subpattern that followed a
2420        previous extended class. It wasn't skipping over the extended class
2421        correctly when UTF-8 mode was not set.]
2422    
2423     5. A negated single-character class was not being recognized as fixed-length
2424        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2425        compile error "lookbehind assertion is not fixed length".
2426    
2427     6. The RunPerlTest auxiliary script was showing an unexpected difference
2428        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2429        write a Perl script that can interpret lines of an input file either as
2430        byte characters or as UTF-8, which is what "perltest" was being required to
2431        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2432        can't do is switch easily at run time between having the "use utf8;" pragma
2433        or not. In the end, I fudged it by using the RunPerlTest script to insert
2434        "use utf8;" explicitly for the UTF-8 tests.
2435    
2436     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2437        the end of the subject string, contrary to the documentation and to what
2438        Perl does. This was true of both matching functions. Now it matches only at
2439        the start of the subject and immediately after *internal* newlines.
2440    
2441     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2442        a pointer to an int instead of a pointer to an unsigned long int. This
2443        caused problems on 64-bit systems.
2444    
2445     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2446        instance of the 'standard' template library not being so standard".
2447    
2448    10. There was no check on the number of named subpatterns nor the maximum
2449        length of a subpattern name. The product of these values is used to compute
2450        the size of the memory block for a compiled pattern. By supplying a very
2451        long subpattern name and a large number of named subpatterns, the size
2452        computation could be caused to overflow. This is now prevented by limiting
2453        the length of names to 32 characters, and the number of named subpatterns
2454        to 10,000.
2455    
2456    11. Subpatterns that are repeated with specific counts have to be replicated in
2457        the compiled pattern. The size of memory for this was computed from the
2458        length of the subpattern and the repeat count. The latter is limited to
2459        65535, but there was no limit on the former, meaning that integer overflow
2460        could in principle occur. The compiled length of a repeated subpattern is
2461        now limited to 30,000 bytes in order to prevent this.
2462    
2463    12. Added the optional facility to have named substrings with the same name.
2464    
2465    13. Added the ability to use a named substring as a condition, using the
2466        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2467        are numbers (not recommended). Forward references are permitted.
2468    
2469    14. Added forward references in named backreferences (if you see what I mean).
2470    
2471    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2472        pattern could run off the end of the subject. For example, the pattern
2473        "(?s)(.{1,5})"8 did this with the subject "ab".
2474    
2475    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2476        PCRE_CASELESS was set when matching characters that were quantified with ?
2477        or *.
2478    
2479    17. A character class other than a single negated character that had a minimum
2480        but no maximum quantifier - for example [ab]{6,} - was not handled
2481        correctly by pce_dfa_exec(). It would match only one character.
2482    
2483    18. A valid (though odd) pattern that looked like a POSIX character
2484        class but used an invalid character after [ (for example [[,abc,]]) caused
2485        pcre_compile() to give the error "Failed: internal error: code overflow" or
2486        in some cases to crash with a glibc free() error. This could even happen if
2487        the pattern terminated after [[ but there just happened to be a sequence of
2488        letters, a binary zero, and a closing ] in the memory that followed.
2489    
2490    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2491        over the years. Originally (before any Unicode support), just the bottom 8
2492        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2493        output from "man perlunicode" includes this:
2494    
2495          The regular expression compiler produces polymorphic opcodes.  That
2496          is, the pattern adapts to the data and automatically switches to
2497          the Unicode character scheme when presented with Unicode data--or
2498          instead uses a traditional byte scheme when presented with byte
2499          data.
2500    
2501        Sadly, a wide octal escape does not cause a switch, and in a string with
2502        no other multibyte characters, these octal escapes are treated as before.
2503        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2504        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2505        Unicode string.
2506    
2507        I have not perpetrated such confusion in PCRE. Up till now, it took just
2508        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2509        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2510        translate to the appropriate multibyte character.
2511    
2512    29. Applied some refactoring to reduce the number of warnings from Microsoft
2513        and Borland compilers. This has included removing the fudge introduced
2514        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2515        a warning about an unused variable.
2516    
2517    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2518        characters since release 4.0, because Perl (from release 5.004) does not.
2519        [Or at least, is documented not to: some releases seem to be in conflict
2520        with the documentation.] However, when a pattern was studied with
2521        pcre_study() and all its branches started with \s, PCRE still included VT
2522        as a possible starting character. Of course, this did no harm; it just
2523        caused an unnecessary match attempt.
2524    
2525    22. Removed a now-redundant internal flag bit that recorded the fact that case
2526        dependency changed within the pattern. This was once needed for "required
2527        byte" processing, but is no longer used. This recovers a now-scarce options
2528        bit. Also moved the least significant internal flag bit to the most-
2529        significant bit of the word, which was not previously used (hangover from
2530        the days when it was an int rather than a uint) to free up another bit for
2531        the future.
2532    
2533    23. Added support for CRLF line endings as well as CR and LF. As well as the
2534        default being selectable at build time, it can now be changed at runtime
2535        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2536        specify that it is scanning data with non-default line endings.
2537    
2538    24. Changed the definition of CXXLINK to make it agree with the definition of
2539        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2540    
2541    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2542        recursions. This makes a big different to stack usage for some patterns.
2543    
2544    26. If a subpattern containing a named recursion or subroutine reference such
2545        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2546        the space required for the compiled pattern went wrong and gave too small a
2547        value. Depending on the environment, this could lead to "Failed: internal
2548        error: code overflow at offset 49" or "glibc detected double free or
2549        corruption" errors.
2550    
2551    27. Applied patches from Google (a) to support the new newline modes and (b) to
2552        advance over multibyte UTF-8 characters in GlobalReplace.
2553    
2554    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2555        difference for some implementation of PCRE in some Windows version.
2556    
2557    29. Added some extra testing facilities to pcretest:
2558    
2559        \q<number>   in a data line sets the "match limit" value
2560        \Q<number>   in a data line sets the "match recursion limt" value
2561        -S <number>  sets the stack size, where <number> is in megabytes
2562    
2563        The -S option isn't available for Windows.
2564    
2565    
2566    Version 6.6 06-Feb-06
2567    ---------------------
2568    
2569     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2570        in pcreposix.h. I have copied the definition from pcre.h.
2571    
2572     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2573        because pcre.h is no longer a built file.
2574    
2575     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2576        not normally included in the compiled code.
2577    
2578    
2579    Version 6.5 01-Feb-06
2580    ---------------------
2581    
2582     1. When using the partial match feature with pcre_dfa_exec(), it was not
2583        anchoring the second and subsequent partial matches at the new starting
2584        point. This could lead to incorrect results. For example, with the pattern
2585        /1234/, partially matching against "123" and then "a4" gave a match.
2586    
2587     2. Changes to pcregrep:
2588    
2589        (a) All non-match returns from pcre_exec() were being treated as failures
2590            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2591            error message is output. Some extra information is given for the
2592            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2593            probably the only errors that are likely to be caused by users (by
2594            specifying a regex that has nested indefinite repeats, for instance).
2595            If there are more than 20 of these errors, pcregrep is abandoned.
2596    
2597        (b) A binary zero was treated as data while matching, but terminated the
2598            output line if it was written out. This has been fixed: binary zeroes
2599            are now no different to any other data bytes.
2600    
2601        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2602            used to set a locale for matching. The --locale=xxxx long option has
2603            been added (no short equivalent) to specify a locale explicitly on the
2604            pcregrep command, overriding the environment variables.
2605    
2606        (d) When -B was used with -n, some line numbers in the output were one less
2607            than they should have been.
2608    
2609        (e) Added the -o (--only-matching) option.
2610    
2611        (f) If -A or -C was used with -c (count only), some lines of context were
2612            accidentally printed for the final match.
2613    
2614        (g) Added the -H (--with-filename) option.
2615    
2616        (h) The combination of options -rh failed to suppress file names for files
2617            that were found from directory arguments.
2618    
2619        (i) Added the -D (--devices) and -d (--directories) options.
2620    
2621        (j) Added the -F (--fixed-strings) option.
2622    
2623        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2624    
2625        (l) Added the --colo(u)r option.
2626    
2627        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2628            is not present by default.
2629    
2630     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2631        items such as (?R) or (?1), when the recursion could match a number of
2632        alternatives. If it matched one of the alternatives, but subsequently,
2633        outside the recursion, there was a failure, the code tried to back up into
2634        the recursion. However, because of the way PCRE is implemented, this is not
2635        possible, and the result was an incorrect result from the match.
2636    
2637        In order to prevent this happening, the specification of recursion has
2638        been changed so that all such subpatterns are automatically treated as
2639        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2640    
2641     4. I had overlooked the fact that, in some locales, there are characters for
2642        which isalpha() is true but neither isupper() nor islower() are true. In
2643        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2644        and ordfeminine) are like this. This affected the treatment of \w and \W
2645        when they appeared in character classes, but not when they appeared outside
2646        a character class. The bit map for "word" characters is now created
2647        separately from the results of isalnum() instead of just taking it from the
2648        upper, lower, and digit maps. (Plus the underscore character, of course.)
2649    
2650     5. The above bug also affected the handling of POSIX character classes such as
2651        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2652        permanent tables. Instead, the bit maps for such a class were previously
2653        created as the appropriate unions of the upper, lower, and digit bitmaps.
2654        Now they are created by subtraction from the [[:word:]] class, which has
2655        its own bitmap.
2656    
2657     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2658        It is created by subtracting the vertical space characters (\x09, \x0a,
2659        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2660        subtraction was done in the overall bitmap for a character class, meaning
2661        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2662        be recognized. This bug has been fixed.
2663    
2664     7. Patches from the folks at Google:
2665    
2666          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2667          real life, but is still worth protecting against".
2668    
2669          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2670          regular expressions".
2671    
2672          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2673          have it.
2674    
2675          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2676          "configure" and the latter not, in order to fix a problem somebody had
2677          with compiling the Arg class on HP-UX.
2678    
2679          (e) Improve the error-handling of the C++ wrapper a little bit.
2680    
2681          (f) New tests for checking recursion limiting.
2682    
2683     8. The pcre_memmove() function, which is used only if the environment does not
2684        have a standard memmove() function (and is therefore rarely compiled),
2685        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2686        returning a result (though PCRE never actually uses the result).
2687    
2688     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2689        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2690        returned instead of calling malloc() with an overflowing number that would
2691        most likely cause subsequent chaos.
2692    
2693    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2694    
2695    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2696        with this option is matched, the nmatch and pmatch options of regexec() are
2697        ignored.
2698    
2699    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2700        provided in case anyone wants to the the POSIX interface with UTF-8
2701        strings.
2702    
2703    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2704        C++ linking (needed for some HP-UX environments).
2705    
2706    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2707        (unused parameter) and in the pcre_printint() function (omitted "default"
2708        switch label when the default is to do nothing).
2709    
2710    15. Added some code to make it possible, when PCRE is compiled as a C++
2711        library, to replace subject pointers for pcre_exec() with a smart pointer
2712        class, thus making it possible to process discontinuous strings.
2713    
2714    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2715        much the same function. They were added by different people who were trying
2716        to make PCRE easy to compile on non-Unix systems. It has been suggested
2717        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2718        for compiling on Windows systems. I have therefore replaced it with
2719        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2720        defaults to "extern" for C or "extern C" for C++, which works fine on
2721        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2722        SCOPE with something explicit in config.h. In addition:
2723    
2724        (a) pcreposix.h still had just "extern" instead of either of these macros;
2725            I have replaced it with PCRE_DATA_SCOPE.
2726    
2727        (b) Functions such as _pcre_xclass(), which are internal to the library,
2728            but external in the C sense, all had PCRE_EXPORT in their definitions.
2729            This is apparently wrong for the Windows case, so I have removed it.
2730            (It makes no difference on Unix-like systems.)
2731    
2732    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2733        of recursive calls to match(). This is different to MATCH_LIMIT because
2734        that limits the total number of calls to match(), not all of which increase
2735        the depth of recursion. Limiting the recursion depth limits the amount of
2736        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2737        when PCRE is compiled, and changed at run time. A patch from Google adds
2738        this functionality to the C++ interface.
2739    
2740    18. Changes to the handling of Unicode character properties:
2741    
2742        (a) Updated the table to Unicode 4.1.0.
2743    
2744        (b) Recognize characters that are not in the table as "Cn" (undefined).
2745    
2746        (c) I revised the way the table is implemented to a much improved format
2747            which includes recognition of ranges. It now supports the ranges that
2748            are defined in UnicodeData.txt, and it also amalgamates other
2749            characters into ranges. This has reduced the number of entries in the
2750            table from around 16,000 to around 3,000, thus reducing its size
2751            considerably. I realized I did not need to use a tree structure after
2752            all - a binary chop search is just as efficient. Having reduced the
2753            number of entries, I extended their size from 6 bytes to 8 bytes to
2754            allow for more data.
2755    
2756        (d) Added support for Unicode script names via properties such as \p{Han}.
2757    
2758    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2759        matching that character.
2760    
2761    20. When matching a repeated Unicode property with a minimum greater than zero,
2762        (for example \pL{2,}), PCRE could look past the end of the subject if it
2763        reached it while seeking the minimum number of characters. This could
2764        happen only if some of the characters were more than one byte long, because
2765        there is a check for at least the minimum number of bytes.
2766    
2767    21. Refactored the implementation of \p and \P so as to be more general, to
2768        allow for more different types of property in future. This has changed the
2769        compiled form incompatibly. Anybody with saved compiled patterns that use
2770        \p or \P will have to recompile them.
2771    
2772    22. Added "Any" and "L&" to the supported property types.
2773    
2774    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2775        but give a compile time error if the value is greater than 0xff.
2776    
2777    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2778        accidentally not being installed or uninstalled.
2779    
2780    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2781        made were to insert the current release number. This seemed silly, because
2782        it made things harder for people building PCRE on systems that don't run
2783        "configure". I have turned pcre.h into a distributed file, no longer built
2784        by "configure", with the version identification directly included. There is
2785        no longer a pcre.h.in file.
2786    
2787        However, this change necessitated a change to the pcre-config script as
2788        well. It is built from pcre-config.in, and one of the substitutions was the
2789        release number. I have updated configure.ac so that ./configure now finds
2790        the release number by grepping pcre.h.
2791    
2792    26. Added the ability to run the tests under valgrind.
2793    
2794    
2795    Version 6.4 05-Sep-05
2796    ---------------------
2797    
2798     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2799        "--" to be printed when multiple files were scanned, even when none of the
2800        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2801        consider it to be a bug, and have restored the previous behaviour.
2802    
2803     2. A couple of code tidies to get rid of compiler warnings.
2804    
2805     3. The pcretest program used to cheat by referring to symbols in the library
2806        whose names begin with _pcre_. These are internal symbols that are not
2807        really supposed to be visible externally, and in some environments it is
2808        possible to suppress them. The cheating is now confined to including
2809        certain files from the library's source, which is a bit cleaner.
2810    
2811     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2812        file's purpose clearer.
2813    
2814     5. Reorganized pcre_ucp_findchar().
2815    
2816    
2817    Version 6.3 15-Aug-05
2818    ---------------------
2819    
2820     1. The file libpcre.pc.in did not have general read permission in the tarball.
2821    
2822     2. There were some problems when building without C++ support:
2823    
2824        (a) If C++ support was not built, "make install" and "make test" still
2825            tried to test it.
2826    
2827        (b) There were problems when the value of CXX was explicitly set. Some
2828            changes have been made to try to fix these, and ...
2829    
2830        (c) --disable-cpp can now be used to explicitly disable C++ support.
2831    
2832        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2833            backslash in a target when C++ was disabled. This confuses some
2834            versions of "make", apparently. Using an intermediate variable solves
2835            this. (Same for CPP_LOBJ.)
2836    
2837     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2838        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2839        necessary on certain architectures.
2840    
2841     4. Added a setting of -export-symbols-regex to the link command to remove
2842        those symbols that are exported in the C sense, but actually are local
2843        within the library, and not documented. Their names all begin with
2844        "_pcre_". This is not a perfect job, because (a) we have to except some
2845        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2846        available (and never for static libraries). I have made a note to try to
2847        find a way round (a) in the future.
2848    
2849    
2850    Version 6.2 01-Aug-05
2851    ---------------------
2852    
2853     1. There was no test for integer overflow of quantifier values. A construction
2854        such as {1111111111111111} would give undefined results. What is worse, if
2855        a minimum quantifier for a parenthesized subpattern overflowed and became
2856        negative, the calculation of the memory size went wrong. This could have
2857        led to memory overwriting.
2858    
2859     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2860    
2861     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2862        operating environments where this matters.
2863    
2864     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2865        PCRE options from within the C++ wrapper.
2866    
2867     5. Named capturing subpatterns were not being correctly counted when a pattern
2868        was compiled. This caused two problems: (a) If there were more than 100
2869        such subpatterns, the calculation of the memory needed for the whole
2870        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2871        back references of the form \12, where the number was greater than 9, were
2872        not recognized as back references, even though there were sufficient
2873        previous subpatterns.
2874    
2875     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2876        versions of gcc, e.g. 2.95.4.
2877    
2878    
2879    Version 6.1 21-Jun-05
2880    ---------------------
2881    
2882     1. There was one reference to the variable "posix" in pcretest.c that was not
2883        surrounded by "#if !defined NOPOSIX".
2884    
2885     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2886        the cross-check on the old pcre_info() function, for the benefit of the
2887        cut-down version of PCRE that is currently imported into Exim.
2888    
2889     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2890        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2891        patterns that start (?i) but I don't think that matters. The use of (?i) is
2892        just an example; this all applies to the other options as well.
2893    
2894     4. Since libtool seems to echo the compile commands it is issuing, the output
2895        from "make" can be reduced a bit by putting "@" in front of each libtool
2896        compile command.
2897    
2898     5. Patch from the folks at Google for configure.in to be a bit more thorough
2899        in checking for a suitable C++ installation before trying to compile the
2900        C++ stuff. This should fix a reported problem when a compiler was present,
2901        but no suitable headers.
2902    
2903     6. The man pages all had just "PCRE" as their title. I have changed them to
2904        be the relevant file name. I have also arranged that these names are
2905        retained in the file doc/pcre.txt, which is a concatenation in text format
2906        of all the man pages except the little individual ones for each function.
2907    
2908     7. The NON-UNIX-USE file had not been updated for the different set of source
2909        files that come with release 6. I also added a few comments about the C++
2910        wrapper.
2911    
2912    
2913    Version 6.0 07-Jun-05
2914    ---------------------
2915    
2916     1. Some minor internal re-organization to help with my DFA experiments.
2917    
2918     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2919        didn't matter for the library itself when fully configured, but did matter
2920        when compiling without UCP support, or within Exim, where the ucp files are
2921        not imported.
2922    
2923     3. Refactoring of the library code to split up the various functions into
2924        different source modules. The addition of the new DFA matching code (see
2925        below) to a single monolithic source would have made it really too
2926        unwieldy, quite apart from causing all the code to be include in a
2927        statically linked application, when only some functions are used. This is
2928        relevant even without the DFA addition now that patterns can be compiled in
2929        one application and matched in another.
2930    
2931        The downside of splitting up is that there have to be some external
2932        functions and data tables that are used internally in different modules of
2933        the library but which are not part of the API. These have all had their
2934        names changed to start with "_pcre_" so that they are unlikely to clash
2935        with other external names.
2936    
2937     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2938        a different (DFA) algorithm. Although it is slower than the original
2939        function, it does have some advantages for certain types of matching
2940        problem.
2941    
2942     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2943        including restarting after a partial match.
2944    
2945     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2946        defined when compiling for Windows was sent to me. I have put it into the
2947        code, though I have no means of testing or verifying it.
2948    
2949     7. Added the pcre_refcount() auxiliary function.
2950    
2951     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2952        match before or at the first newline in the subject string. In pcretest,
2953        the /f option on a pattern can be used to set this.
2954    
2955     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2956        would behave wrongly. This has been present in PCRE since release 4.0.
2957    
2958    10. A number of changes to the pcregrep command:
2959    
2960        (a) Refactored how -x works; insert ^(...)$ instead of setting
2961            PCRE_ANCHORED and checking the length, in preparation for adding
2962            something similar for -w.
2963    
2964        (b) Added the -w (match as a word) option.
2965    
2966        (c) Refactored the way lines are read and buffered so as to have more
2967            than one at a time available.
2968    
2969        (d) Implemented a pcregrep test script.
2970    
2971        (e) Added the -M (multiline match) option. This allows patterns to match
2972            over several lines of the subject. The buffering ensures that at least
2973            8K, or the rest of the document (whichever is the shorter) is available
2974            for matching (and similarly the previous 8K for lookbehind assertions).
2975    
2976        (f) Changed the --help output so that it now says
2977    
2978              -w, --word-regex(p)
2979    
2980            instead of two lines, one with "regex" and the other with "regexp"
2981            because that confused at least one person since the short forms are the
2982            same. (This required a bit of code, as the output is generated
2983            automatically from a table. It wasn't just a text change.)
2984    
2985        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2986            option but starts with a hyphen. Could be a pattern or a path name
2987            starting with a hyphen, for instance.
2988    
2989        (h) "-" can be given as a file name to represent stdin.
2990    
2991        (i) When file names are being printed, "(standard input)" is used for
2992            the standard input, for compatibility with GNU grep. Previously
2993            "<stdin>" was used.
2994    
2995        (j) The option --label=xxx can be used to supply a name to be used for
2996            stdin when file names are being printed. There is no short form.
2997    
2998        (k) Re-factored the options decoding logic because we are going to add
2999            two more options that take data. Such options can now be given in four
3000            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3001    
3002        (l) Added the -A, -B, and -C options for requesting that lines of context
3003            around matches be printed.
3004    
3005        (m) Added the -L option to print the names of files that do not contain
3006            any matching lines, that is, the complement of -l.
3007    
3008        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3009            continue to scan other files.
3010    
3011        (o) The -s option was incorrectly implemented. For compatibility with other
3012            greps, it now suppresses the error message for a non-existent or non-
3013            accessible file (but not the return code). There is a new option called
3014            -q that suppresses the output of matching lines, which was what -s was
3015            previously doing.
3016    
3017        (p) Added --include and --exclude options to specify files for inclusion
3018            and exclusion when recursing.
3019    
3020    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3021        Hopefully, it now does.
3022    
3023    12. Missing cast in pcre_study().
3024    
3025    13. Added an "uninstall" target to the makefile.
3026    
3027    14. Replaced "extern" in the function prototypes in Makefile.in with
3028        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3029        world, but is set differently for Windows.
3030    
3031    15. Added a second compiling function called pcre_compile2(). The only
3032        difference is that it has an extra argument, which is a pointer to an
3033        integer error code. When there is a compile-time failure, this is set
3034        non-zero, in addition to the error test pointer being set to point to an
3035        error message. The new argument may be NULL if no error number is required
3036        (but then you may as well call pcre_compile(), which is now just a
3037        wrapper). This facility is provided because some applications need a
3038        numeric error indication, but it has also enabled me to tidy up the way
3039        compile-time errors are handled in the POSIX wrapper.
3040    
3041    16. Added VPATH=.libs to the makefile; this should help when building with one
3042        prefix path and installing with another. (Or so I'm told by someone who
3043        knows more about this stuff than I do.)
3044    
3045    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3046        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3047        match everything, including newlines. This is not POSIX-compatible, but
3048        somebody wanted the feature. From pcretest it can be activated by using
3049        both the P and the s flags.
3050    
3051    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3052    
3053    19. libpcre.pc was being incorrectly installed as executable.
3054    
3055    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3056        it now also looks for '\r' so that it will work unmodified on Windows.
3057    
3058    21. Added Google's contributed C++ wrapper to the distribution.
3059    
3060    22. Added some untidy missing memory free() calls in pcretest, to keep
3061        Electric Fence happy when testing.
3062    
3063    
3064    
3065    Version 5.0 13-Sep-04
3066    ---------------------
3067    
3068     1. Internal change: literal characters are no longer packed up into items
3069        containing multiple characters in a single byte-string. Each character
3070        is now matched using a separate opcode. However, there may be more than one
3071        byte in the character in UTF-8 mode.
3072    
3073     2. The pcre_callout_block structure has two new fields: pattern_position and
3074        next_item_length. These contain the offset in the pattern to the next match
3075        item, and its length, respectively.
3076    
3077     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3078        insertion of callouts before each pattern item. Added the /C option to
3079        pcretest to make use of this.
3080    
3081     4. On the advice of a Windows user, the lines
3082    
3083          #if defined(_WIN32) || defined(WIN32)
3084          _setmode( _fileno( stdout ), 0x8000 );
3085          #endif  /* defined(_WIN32) || defined(WIN32) */
3086    
3087        have been added to the source of pcretest. This apparently does useful
3088        magic in relation to line terminators.
3089    
3090     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3091        for the benefit of those environments where the "b" makes a difference.
3092    
3093     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3094        to know about it. I have put a hack into configure.in that adds in code
3095        to set GCC=yes if CC=icc. This seems to end up at a point in the
3096        generated configure script that is early enough to affect the setting of
3097        compiler options, which is what is needed, but I have no means of testing
3098        whether it really works. (The user who reported this had patched the
3099        generated configure script, which of course I cannot do.)
3100    
3101        LATER: After change 22 below (new libtool files), the configure script
3102        seems to know about icc (and also ecc). Therefore, I have commented out
3103        this hack in configure.in.
3104    
3105     7. Added support for pkg-config (2 patches were sent in).
3106    
3107     8. Negated POSIX character classes that used a combination of internal tables
3108        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3109        [[:^ascii]]. Typically, they would match almost any characters. The other
3110        POSIX classes were not broken in this way.
3111    
3112     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3113        to find the match, as PCRE was deluded into thinking that the match had to
3114        start at the start point or following a newline. The same bug applied to
3115        patterns with negative forward assertions or any backward assertions
3116        preceding ".*" at the start, unless the pattern required a fixed first
3117        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3118    
3119    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3120        starting at the last subject character, bytes beyond the end of the subject
3121        string were read.
3122    
3123    11. Renamed the variable "class" as "classbits" to make life easier for C++
3124        users. (Previously there was a macro definition, but it apparently wasn't
3125        enough.)
3126    
3127    12. Added the new field "tables" to the extra data so that tables can be passed
3128        in at exec time, or the internal tables can be re-selected. This allows
3129        a compiled regex to be saved and re-used at a later time by a different
3130        program that might have everything at different addresses.
3131    
3132    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3133        -R library as well as a -L library.
3134    
3135    14. The debugging options of pcretest (-d on the command line or D on a
3136        pattern) showed incorrect output for anything following an extended class
3137        that contained multibyte characters and which was followed by a quantifier.
3138    
3139    15. Added optional support for general category Unicode character properties
3140        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3141        support. It adds about 90K to the size of the library. The meanings of the
3142        inbuilt class escapes such as \d and \s have NOT been changed.
3143    
3144    16. Updated pcredemo.c to include calls to free() to release the memory for the
3145        compiled pattern.
3146    
3147    17. The generated file chartables.c was being created in the source directory
3148        instead of in the building directory. This caused the build to fail if the
3149        source directory was different from the building directory, and was
3150        read-only.
3151    
3152    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3153        file. No doubt somebody will tell me if they don't make sense... Also added
3154        Dan Mooney's comments about building on OpenVMS.
3155    
3156    19. Added support for partial matching via the PCRE_PARTIAL option for
3157        pcre_exec() and the \P data escape in pcretest.
3158    
3159    20. Extended pcretest with 3 new pattern features:
3160    
3161        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3162              write the compiled pattern to the file whose name is "rest-of-line".
3163              This is a straight binary dump of the data, with the saved pointer to
3164              the character tables forced to be NULL. The study data, if any, is
3165              written too. After writing, pcretest reads a new pattern.
3166    
3167        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3168              compiled pattern from the given file. There must not be any
3169              occurrences of "<" in the file name (pretty unlikely); if there are,
3170              pcretest will instead treat the initial "<" as a pattern delimiter.
3171              After reading in the pattern, pcretest goes on to read data lines as
3172              usual.
3173    
3174        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3175              and 16-bit fields in a compiled pattern, to simulate a pattern that
3176              was compiled on a host of opposite endianness.
3177    
3178    21. The pcre-exec() function can now cope with patterns that were compiled on
3179        hosts of opposite endianness, with this restriction:
3180    
3181          As for any compiled expression that is saved and used later, the tables
3182          pointer field cannot be preserved; the extra_data field in the arguments
3183          to pcre_exec() should be used to pass in a tables address if a value
3184          other than the default internal tables were used at compile time.
3185    
3186    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3187        now diagnosed as an error. Previously, most of the time, a negative number
3188        would have been treated as zero, but if in addition "ovector" was passed as
3189        NULL, a crash could occur.
3190    
3191    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3192        new versions from the libtool 1.5 distribution (the last one is a copy of
3193        a file called libtool.m4). This seems to have fixed the need to patch
3194        "configure" to support Darwin 1.3 (which I used to do). However, I still
3195        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3196        workstation).
3197    
3198    24. Changed the PCRE licence to be the more standard "BSD" licence.
3199    
3200    
3201    Version 4.5 01-Dec-03
3202    ---------------------
3203    
3204     1. There has been some re-arrangement of the code for the match() function so
3205        that it can be compiled in a version that does not call itself recursively.
3206        Instead, it keeps those local variables that need separate instances for
3207        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3208        needs to "recurse". Keeping track of where control must go is done by means
3209        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3210        hide most of the details from the main code, and operates only if
3211        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3212        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3213        operating.
3214    
3215        To make it easier for callers to provide specially tailored get/free
3216        functions for this usage, two new functions, pcre_stack_malloc, and
3217        pcre_stack_free, are used. They are always called in strict stacking order,
3218        and the size of block requested is always the same.
3219    
3220        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3221        PCRE has been compiled to use the stack or the heap for recursion. The
3222        -C option of pcretest uses this to show which version is compiled.
3223    
3224        A new data escape \S, is added to pcretest; it causes the amounts of store
3225        obtained and freed by both kinds of malloc/free at match time to be added
3226        to the output.
3227    
3228     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3229        what's available on my current Linux desktop machine.
3230    
3231     3. When matching a UTF-8 string, the test for a valid string at the start has
3232        been extended. If start_offset is not zero, PCRE now checks that it points
3233        to a byte that is the start of a UTF-8 character. If not, it returns
3234        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3235        this is necessary because there may be backward assertions in the pattern.
3236        When matching the same subject several times, it may save resources to use
3237        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3238    
3239     4. The code for checking the validity of UTF-8 strings has been tightened so
3240        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3241        containing "overlong sequences".
3242    
3243     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3244        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3245        should have been "&", but it just so happened that all the cases this let
3246        through by mistake were picked up later in the function.
3247    
3248     6. I had used a variable called "isblank" - this is a C99 function, causing
3249        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3250    
3251     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3252        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3253        so the version is shown; (c) stop "make test" echoing "./RunTest".
3254    
3255     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3256    
3257     9. The prototype for memmove() for systems that don't have it was using
3258        size_t, but the inclusion of the header that defines size_t was later. I've
3259        moved the #includes for the C headers earlier to avoid this.
3260    
3261    10. Added some adjustments to the code to make it easier to compiler on certain
3262        special systems:
3263    
3264          (a) Some "const" qualifiers were missing.
3265          (b) Added the macro EXPORT before all exported functions; by default this
3266              is defined to be empty.
3267          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3268              that it reads its output file name as an argument instead of writing
3269              to the standard output and assuming this can be redirected.
3270    
3271    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3272        class containing characters with values greater than 255, PCRE compilation
3273        went into a loop.
3274    
3275    12. A recursive reference to a subpattern that was within another subpattern
3276        that had a minimum quantifier of zero caused PCRE to crash. For example,
3277        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3278        recursion. If the recursively-called subpattern itself had a zero repeat,
3279        that was OK.
3280    
3281    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3282        buffer into which it was copied (for escape processing) was still set at
3283        1024, so long lines caused crashes.
3284    
3285    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3286        "internal error: code overflow...". This applied to any character class
3287        that was followed by a possessive quantifier.
3288    
3289    15. Modified the Makefile to add libpcre.la as a prerequisite for
3290        libpcreposix.la because I was told this is needed for a parallel build to
3291        work.
3292    
3293    16. If a pattern that contained .* following optional items at the start was
3294        studied, the wrong optimizing data was generated, leading to matching
3295        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3296        matching string must start with a or b or c. The correct conclusion for
3297        this pattern is that a match can start with any character.
3298    
3299    
3300    Version 4.4 13-Aug-03
3301    ---------------------
3302    
3303     1. In UTF-8 mode, a character class containing characters with values between
3304        127 and 255 was not handled correctly if the compiled pattern was studied.
3305        In fixing this, I have also improved the studying algorithm for such
3306        classes (slightly).
3307    
3308     2. Three internal functions had redundant arguments passed to them. Removal
3309        might give a very teeny performance improvement.
3310    
3311     3. Documentation bug: the value of the capture_top field in a callout is *one
3312        more than* the number of the hightest numbered captured substring.
3313    
3314     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3315        in incorrectly linking with a previously installed version. They now link
3316        explicitly with libpcre.la.
3317    
3318     5. configure.in no longer needs to recognize Cygwin specially.
3319    
3320     6. A problem in pcre.in for Windows platforms is fixed.
3321    
3322     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3323        pcretest, it used to include the size of the study block as part of its
3324        output. Unfortunately, the structure contains a field that has a different
3325        size on different hardware architectures. This meant that the tests that
3326        showed this size failed. As the block is currently always of a fixed size,
3327        this information isn't actually particularly useful in pcretest output, so
3328        I have just removed it.
3329    
3330     8. Three pre-processor statements accidentally did not start in column 1.
3331        Sadly, there are *still* compilers around that complain, even though
3332        standard C has not required this for well over a decade. Sigh.
3333    
3334     9. In pcretest, the code for checking callouts passed small integers in the
3335        callout_data field, which is a void * field. However, some picky compilers
3336        complained about the casts involved for this on 64-bit systems. Now
3337        pcretest passes the address of the small integer instead, which should get
3338        rid of the warnings.
3339    
3340    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3341        both compile and run time, and gives an error if an invalid UTF-8 sequence
3342        is found. There is a option for disabling this check in cases where the
3343        string is known to be correct and/or the maximum performance is wanted.
3344    
3345    11. In response to a bug report, I changed one line in Makefile.in from
3346    
3347            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3348        to
3349            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3350    
3351        to look similar to other lines, but I have no way of telling whether this
3352        is the right thing to do, as I do not use Windows. No doubt I'll get told
3353        if it's wrong...
3354    
3355    
3356    Version 4.3 21-May-03
3357    ---------------------
3358    
3359    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3360       Makefile.
3361    
3362    2. Some refactoring to improve the quality of the code:
3363    
3364       (i)   The utf8_table... variables are now declared "const".
3365    
3366       (ii)  The code for \cx, which used the "case flipping" table to upper case
3367             lower case letters, now just substracts 32. This is ASCII-specific,
3368             but the whole concept of \cx is ASCII-specific, so it seems
3369             reasonable.
3370    
3371       (iii) PCRE was using its character types table to recognize decimal and
3372             hexadecimal digits in the pattern. This is silly, because it handles
3373             only 0-9, a-f, and A-F, but the character types table is locale-
3374             specific, which means strange things might happen. A private
3375             table is now used for this - though it costs 256 bytes, a table is
3376             much faster than multiple explicit tests. Of course, the standard
3377             character types table is still used for matching digits in subject
3378             strings against \d.
3379    
3380       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3381             ending in _t are). So I've renamed it as ESC_tee.
3382    
3383    3. The first argument for regexec() in the POSIX wrapper should have been
3384       defined as "const".
3385    
3386    4. Changed pcretest to use malloc() for its buffers so that they can be
3387       Electric Fenced for debugging.
3388    
3389    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3390       to read one or more bytes before the start of the subject string. Often this
3391       had no effect on PCRE's behaviour, but in some circumstances it could
3392       provoke a segmentation fault.
3393    
3394    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3395       to try to read one or more bytes before the start of the subject string.
3396    
3397    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3398       UTF-8 support could misbehave in various ways if the subject string
3399       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3400       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3401       back over UTF-8 characters.)
3402    
3403    
3404    Version 4.2 14-Apr-03
3405    ---------------------
3406    
3407    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3408    
3409    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3410         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3411         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3412         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3413         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3414           and BUILD_EXEEXT
3415         Note: automatic setting of the BUILD variables is not yet working
3416         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3417           compile-time but not at link-time
3418         [LINK]: use for linking executables only
3419         make different versions for Windows and non-Windows
3420         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3421           libraries
3422         [LINK_FOR_BUILD]: new variable
3423         [OBJEXT]: use throughout
3424         [EXEEXT]: use throughout
3425         <winshared>: new target
3426         <wininstall>: new target
3427         <dftables.o>: use native compiler
3428         <dftables>: use native linker
3429         <install>: handle Windows platform correctly
3430         <clean>: ditto
3431         <check>: ditto
3432         copy DLL to top builddir before testing
3433    
3434       As part of these changes, -no-undefined was removed again. This was reported
3435       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3436       in any case.
3437    
3438    3. Some tidies to get rid of compiler warnings:
3439    
3440       . In the match_data structure, match_limit was an unsigned long int, whereas
3441         match_call_count was an int. I've made them both unsigned long ints.
3442    
3443       . In pcretest the fact that a const uschar * doesn't automatically cast to
3444         a void * provoked a warning.
3445    
3446       . Turning on some more compiler warnings threw up some "shadow" variables
3447         and a few more missing casts.
3448    
3449    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3450       option, a class that contained a single character with a value between 128
3451       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3452    
3453    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3454       option, a class that contained several characters, but with at least one
3455       whose value was between 128 and 255 caused PCRE to crash.
3456    
3457    
3458    Version 4.1 12-Mar-03
3459    ---------------------
3460    
3461    1. Compiling with gcc -pedantic found a couple of places where casts were
3462    needed, and a string in dftables.c that was longer than standard compilers are
3463    required to support.
3464    
3465    2. Compiling with Sun's compiler found a few more places where the code could
3466    be tidied up in order to avoid warnings.
3467    
3468    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3469    first of these names is deprecated in the latest Autoconf in favour of the name
3470    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3471    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3472    analogy I have changed it to CFLAGS_FOR_BUILD.
3473    
3474    4. Added -no-undefined to the linking command in the Makefile, because this is
3475    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3476    linking step for the pcreposix library.
3477    
3478    5. PCRE was failing to diagnose the case of two named groups with the same
3479    name.
3480    
3481    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3482    literal character that is needed in the subject for a match, and scans along to
3483    ensure that it is present before embarking on the full matching process. This
3484    saves time in cases of nested unlimited repeats that are never going to match.
3485    Problem: the scan can take a lot of time if the subject is very long (e.g.
3486    megabytes), thus penalizing straightforward matches. It is now done only if the
3487    amount of subject to be scanned is less than 1000 bytes.
3488    
3489    7. A lesser problem with the same optimization is that it was recording the
3490    first character of an anchored pattern as "needed", thus provoking a search
3491    right along the subject, even when the first match of the pattern was going to
3492    fail. The "needed" character is now not set for anchored patterns, unless it
3493    follows something in the pattern that is of non-fixed length. Thus, it still
3494    fulfils its original purpose of finding quick non-matches in cases of nested
3495    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3496    
3497    
3498    Version 4.0 17-Feb-03
3499    ---------------------
3500    
3501    1. If a comment in an extended regex that started immediately after a meta-item
3502    extended to the end of string, PCRE compiled incorrect data. This could lead to
3503    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3504    
3505    2. Moved to autoconf 2.53 and libtool 1.4.2.
3506    
3507    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3508    the special perltest8 script is no longer needed - all the tests can be run
3509    from a single perltest script.
3510    
3511    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3512    by \s. It has now been removed in PCRE. This means it isn't recognized as
3513    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3514    class [:space:] *does* include VT, thereby creating a mess.
3515    
3516    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3517    space and tab.
3518    
3519    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3520    its new features into the main test script, reducing the number of scripts.
3521    
3522    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3523    were backward compatible, and made the (?i) apply to the whole pattern, as if
3524    /i were given. Now it behaves more logically, and applies the option setting
3525    only to what follows. PCRE has been changed to follow suit. However, if it
3526    finds options settings right at the start of the pattern, it extracts them into
3527    the global options, as before. Thus, they show up in the info data.
3528    
3529    8. Added support for the \Q...\E escape sequence. Characters in between are
3530    treated as literals. This is slightly different from Perl in that $ and @ are
3531    also handled as literals inside the quotes. In Perl, they will cause variable
3532    interpolation. Note the following examples:
3533    
3534        Pattern            PCRE matches      Perl matches
3535    
3536        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3537        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3538        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3539    
3540    For compatibility with Perl, \Q...\E sequences are recognized inside character
3541    classes as well as outside them.
3542    
3543    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3544    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3545    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3546    signed/unsigned warnings.
3547    
3548    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3549    option for pcretest, so I've replaced it by a simple function that does just
3550    that job.
3551    
3552    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3553    "pcregrep -".
3554    
3555    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3556    Java package. This provides some syntactic sugar for simple cases of what my
3557    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3558    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3559    item, you can use this simplified notation. Note that only makes sense with
3560    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3561    greediness, whatever the setting of the PCRE_UNGREEDY option.
3562    
3563    13. A change of greediness default within a pattern was not taking effect at
3564    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3565    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3566    was abstracted outside.
3567    
3568    14. PCRE now supports the \G assertion. It is true when the current matching
3569    position is at the start point of the match. This differs from \A when the
3570    starting offset is non-zero. Used with the /g option of pcretest (or similar
3571    code), it works in the same way as it does for Perl's /g option. If all
3572    alternatives of a regex begin with \G, the expression is anchored to the start
3573    match position, and the "anchored" flag is set in the compiled expression.
3574    
3575    15. Some bugs concerning the handling of certain option changes within patterns
3576    have been fixed. These applied to options other than (?ims). For example,
3577    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3578    been the other way round. Some of this was related to change 7 above.
3579    
3580    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3581    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3582    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3583    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3584    
3585    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3586    mode. Unlike ".", it always matches newline, whatever the setting of
3587    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3588    assertions. Perl allows it, but it doesn't (in general) work because it can't
3589    calculate the length of the lookbehind. At least, that's the case for Perl
3590    5.8.0 - I've been told they are going to document that it doesn't work in
3591    future.
3592    
3593    18. Added an error diagnosis for escapes that PCRE does not support: these are
3594    \L, \l, \N, \P, \p, \U, \u, and \X.
3595    
3596    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3597    reading past the end of the pattern in cases such as /[abcd/.
3598    
3599    20. PCRE was getting more memory than necessary for patterns with classes that
3600    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3601    
3602    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3603    compiling PCRE for use with Virtual Pascal.
3604    
3605    22. Small fix to the Makefile to make it work properly if the build is done
3606    outside the source tree.
3607    
3608    23. Added a new extension: a condition to go with recursion. If a conditional
3609    subpattern starts with (?(R) the "true" branch is used if recursion has
3610    happened, whereas the "false" branch is used only at the top level.
3611    
3612    24. When there was a very long string of literal characters (over 255 bytes
3613    without UTF support, over 250 bytes with UTF support), the computation of how
3614    much memory was required could be incorrect, leading to segfaults or other
3615    strange effects.
3616    
3617    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3618    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3619    there was a subsequent back reference to those brackets. This meant that, for
3620    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3621    possible to check for precisely this case. All we can do is abandon the
3622    optimization if .* occurs inside capturing brackets when there are any back
3623    references whatsoever. (See below for a better fix that came later.)
3624    
3625    26. The handling of the optimization for finding the first character of a
3626    non-anchored pattern, and for finding a character that is required later in the
3627    match were failing in some cases. This didn't break the matching; it just
3628    failed to optimize when it could. The way this is done has been re-implemented.
3629    
3630    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3631    
3632    28. Added a new feature that provides some of the functionality that Perl
3633    provides with (?{...}). The facility is termed a "callout". The way it is done
3634    in PCRE is for the caller to provide an optional function, by setting
3635    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3636    global variable. By default it is unset, which disables all calling out. To get
3637    the function called, the regex must include (?C) at appropriate points. This
3638    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3639    This provides a means of identifying different callout points. When PCRE
3640    reaches such a point in the regex, if pcre_callout has been set, the external
3641    function is called. It is provided with data in a structure called
3642    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3643    matching continues; if it returns a non-zero value, the match at the current
3644    point fails. However, backtracking will occur if possible. [This was changed
3645    later and other features added - see item 49 below.]
3646    
3647    29. pcretest is upgraded to test the callout functionality. It provides a
3648    callout function that displays information. By default, it shows the start of
3649    the match and the current position in the text. There are some new data escapes
3650    to vary what happens:
3651    
3652        \C+         in addition, show current contents of captured substrings
3653        \C-         do not supply a callout function
3654        \C!n        return 1 when callout number n is reached
3655        \C!n!m      return 1 when callout number n is reached for the mth time
3656    
3657    30. If pcregrep was called with the -l option and just a single file name, it
3658    output "<stdin>" if a match was found, instead of the file name.
3659    
3660    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3661    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3662    pcre_exec(). This saves a malloc/free per call. The default value of
3663    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3664    when configuring.
3665    
3666    32. The default maximum size of a compiled pattern is 64K. There have been a
3667    few cases of people hitting this limit. The code now uses macros to handle the
3668    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3669    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3670    configuring. Tests 2 and 5 work only with 2-byte links because they output
3671    debugging information about compiled patterns.
3672    
3673    33. Internal code re-arrangements:
3674    
3675    (a) Moved the debugging function for printing out a compiled regex into
3676        its own source file (printint.c) and used #include to pull it into
3677        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3678        separate copies.
3679    
3680    (b) Defined the list of op-code names for debugging as a macro in
3681        internal.h so that it is next to the definition of the opcodes.
3682    
3683    (c) Defined a table of op-code lengths for simpler skipping along compiled
3684        code. This is again a macro in internal.h so that it is next to the
3685        definition of the opcodes.
3686    
3687    34. Added support for recursive calls to individual subpatterns, along the
3688    lines of Robin Houston's patch (but implemented somewhat differently).
3689    
3690    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3691    allow it to read and process whole directories in Win32. This code was
3692    contributed by Lionel Fourquaux; it has not been tested by me.
3693    
3694    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3695    used to name a group. Names consist of alphanumerics and underscores, and must
3696    be unique. Back references use the syntax (?P=name) and recursive calls use
3697    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3698    numbers. The function pcre_fullinfo() can be used after compilation to extract
3699    a name/number map. There are three relevant calls:
3700    
3701      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3702      PCRE_INFO_NAMECOUNT            yields the number of entries
3703      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3704    
3705    The map is a vector of fixed-size entries. The size of each entry depends on
3706    the length of the longest name used. The first two bytes of each entry are the
3707    group number, most significant byte first. There follows the corresponding
3708    name, zero terminated. The names are in alphabetical order.
3709    
3710    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3711    case instead of 255. Making it the same both with and without UTF-8 support
3712    means that the same test output works with both.
3713    
3714    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3715    calling malloc() with a zero argument.
3716    
3717    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3718    optimization. I've improved things by keeping a bitmap of backreferences with
3719    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3720    fact referenced, the optimization can be applied. It is unlikely that a
3721    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3722    the match to follow \n) will appear inside brackets with a number greater than
3723    31, but if it does, any back reference > 31 suppresses the optimization.
3724    
3725    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3726    of disabling numbered capturing parentheses. Any opening parenthesis that is
3727    not followed by ? behaves as if it were followed by ?: but named parentheses
3728    can still be used for capturing (and they will acquire numbers in the usual
3729    way).
3730    
3731    41. Redesigned the return codes from the match() function into yes/no/error so
3732    that errors can be passed back from deep inside the nested calls. A malloc
3733    failure while inside a recursive subpattern call now causes the
3734    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3735    
3736    42. It is now possible to set a limit on the number of times the match()
3737    function is called in a call to pcre_exec(). This facility makes it possible to
3738    limit the amount of recursion and backtracking, though not in a directly
3739    obvious way, because the match() function is used in a number of different
3740    circumstances. The count starts from zero for each position in the subject
3741    string (for non-anchored patterns). The default limit is, for compatibility, a
3742    large number, namely 10 000 000. You can change this in two ways:
3743    
3744    (a) When configuring PCRE before making, you can use --with-match-limit=n
3745        to set a default value for the compiled library.
3746    
3747    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3748        a different value is set. See 45 below.
3749    
3750    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3751    
3752    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3753    of things that can be changed at compile time. The first argument specifies
3754    what is wanted and the second points to where the information is to be placed.
3755    The current list of available information is:
3756    
3757      PCRE_CONFIG_UTF8
3758    
3759    The output is an integer that is set to one if UTF-8 support is available;
3760    otherwise it is set to zero.
3761    
3762      PCRE_CONFIG_NEWLINE
3763    
3764    The output is an integer that it set to the value of the code that is used for
3765    newline. It is either LF (10) or CR (13).
3766    
3767      PCRE_CONFIG_LINK_SIZE
3768    
3769    The output is an integer that contains the number of bytes used for internal
3770    linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3771    
3772      PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3773    
3774    The output is an integer that contains the threshold above which the POSIX
3775    interface uses malloc() for output vectors. See item 31 above.
3776    
3777      PCRE_CONFIG_MATCH_LIMIT
3778    
3779    The output is an unsigned integer that contains the default limit of the number
3780    of match() calls in a pcre_exec() execution. See 42 above.
3781    
3782    44. pcretest has been upgraded by the addition of the -C option. This causes it
3783    to extract all the available output from the new pcre_config() function, and to
3784    output it. The program then exits immediately.
3785    
3786    45. A need has arisen to pass over additional data with calls to pcre_exec() in
3787    order to support additional features. One way would have been to define
3788    pcre_exec2() (for example) with extra arguments, but this would not have been
3789    extensible, and would also have required all calls to the original function to
3790    be mapped to the new one. Instead, I have chosen to extend the mechanism that
3791    is used for passing in "extra" data from pcre_study().
3792    
3793    The pcre_extra structure is now exposed and defined in pcre.h. It currently
3794    contains the following fields:
3795    
3796      flags         a bitmap indicating which of the following fields are set
3797      study_data    opaque data from pcre_study()
3798      match_limit   a way of specifying a limit on match() calls for a specific
3799                      call to pcre_exec()
3800      callout_data  data for callouts (see 49 below)
3801    
3802    The flag bits are also defined in pcre.h, and are
3803    
3804      PCRE_EXTRA_STUDY_DATA
3805      PCRE_EXTRA_MATCH_LIMIT
3806      PCRE_EXTRA_CALLOUT_DATA
3807    
3808    The pcre_study() function now returns one of these new pcre_extra blocks, with
3809    the actual study data pointed to by the study_data field, and the
3810    PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3811    before. That is, this change is entirely upwards-compatible and requires no
3812