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

Diff of /code/trunk/ChangeLog

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

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