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

Diff of /code/trunk/ChangeLog

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

revision 13 by nigel, Sat Feb 24 21:38:21 2007 UTC revision 807 by ph10, Sun Dec 18 10:03:38 2011 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.22
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 the code for calculating the fixed length of lookbehind
11        assertions.
12    
13    
14    Version 8.21 12-Dec-2011
15    ------------------------
16    
17    1.  Updating the JIT compiler.
18    
19    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
20        are added as well.
21    
22    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
23        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
24        calling _pcre_jit_exec. Some extra comments are added.
25    
26    4.  (*MARK) settings inside atomic groups that do not contain any capturing
27        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
28        was introduced by change 18 for 8.20.
29    
30    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
31        ECMA-262 standard.
32    
33    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
34        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
35        This bug was probably introduced by change 9 of 8.13.
36    
37    7.  While fixing 6 above, I noticed that a number of other items were being
38        incorrectly rejected as "not fixed length". This arose partly because newer
39        opcodes had not been added to the fixed-length checking code. I have (a)
40        corrected the bug and added tests for these items, and (b) arranged for an
41        error to occur if an unknown opcode is encountered while checking for fixed
42        length instead of just assuming "not fixed length". The items that were
43        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
44        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
45        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
46    
47    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
48        being incorrectly compiled and would have given unpredicatble results.
49    
50    9.  A possessively repeated subpattern with minimum repeat count greater than
51        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
52        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
53        the first (A) could occur when it should not.
54    
55    10. Add a cast and remove a redundant test from the code.
56    
57    11. JIT should use pcre_malloc/pcre_free for allocation.
58    
59    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
60        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
61        is anything other than /usr, -L is still shown).
62    
63    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
64    
65    14. Perl does not support \N without a following name in a [] class; PCRE now
66        also gives an error.
67    
68    15. If a forward reference was repeated with an upper limit of around 2000,
69        it caused the error "internal error: overran compiling workspace". The
70        maximum number of forward references (including repeats) was limited by the
71        internal workspace, and dependent on the LINK_SIZE. The code has been
72        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
73        the default depends on LINK_SIZE. There is a new upper limit (for safety)
74        of around 200,000 forward references. While doing this, I also speeded up
75        the filling in of repeated forward references.
76    
77    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
78        incorrectly expecting the subject to contain another "a" after the start.
79    
80    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
81        in the match, the SKIP should be ignored. This was not happening; instead
82        the SKIP was being treated as NOMATCH. For patterns such as
83        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
84        tested.
85    
86    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
87        now much more compatible with Perl, in particular in cases where the result
88        is a non-match for a non-anchored pattern. For example, if
89        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
90        "m", where previously it did not return a name. A side effect of this
91        change is that for partial matches, the last encountered mark name is
92        returned, as for non matches. A number of tests that were previously not
93        Perl-compatible have been moved into the Perl-compatible test files. The
94        refactoring has had the pleasing side effect of removing one argument from
95        the match() function, thus reducing its stack requirements.
96    
97    19. If the /S+ option was used in pcretest to study a pattern using JIT,
98        subsequent uses of /S (without +) incorrectly behaved like /S+.
99    
100    21. Retrieve executable code size support for the JIT compiler and fixing
101        some warnings.
102    
103    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
104        not work when the shorter character appeared right at the end of the
105        subject string.
106    
107    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
108        systems.
109    
110    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
111        output it when the /M option is used in pcretest.
112    
113    25. The CheckMan script was not being included in the distribution. Also, added
114        an explicit "perl" to run Perl scripts from the PrepareRelease script
115        because this is reportedly needed in Windows.
116    
117    26. If study data was being save in a file and studying had not found a set of
118        "starts with" bytes for the pattern, the data written to the file (though
119        never used) was taken from uninitialized memory and so caused valgrind to
120        complain.
121    
122    27. Updated RunTest.bat as provided by Sheri Pierce.
123    
124    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
125    
126    29. Computation of memory usage for the table of capturing group names was
127        giving an unnecessarily large value.
128    
129    
130    Version 8.20 21-Oct-2011
131    ------------------------
132    
133    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
134        a POSIX class. After further experiments with Perl, which convinced me that
135        Perl has bugs and confusions, a closing square bracket is no longer allowed
136        in a POSIX name. This bug also affected patterns with classes that started
137        with full stops.
138    
139    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
140        captured substring, but while checking the failing first alternative,
141        substring 1 is temporarily captured. If the output vector supplied to
142        pcre_exec() was not big enough for this capture, the yield of the function
143        was still zero ("insufficient space for captured substrings"). This cannot
144        be totally fixed without adding another stack variable, which seems a lot
145        of expense for a edge case. However, I have improved the situation in cases
146        such as /(a)(b)x|abc/ matched against "abc", where the return code
147        indicates that fewer than the maximum number of slots in the ovector have
148        been set.
149    
150    3.  Related to (2) above: when there are more back references in a pattern than
151        slots in the output vector, pcre_exec() uses temporary memory during
152        matching, and copies in the captures as far as possible afterwards. It was
153        using the entire output vector, but this conflicts with the specification
154        that only 2/3 is used for passing back captured substrings. Now it uses
155        only the first 2/3, for compatibility. This is, of course, another edge
156        case.
157    
158    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
159        main code base, and can be used by building with --enable-jit. When this is
160        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
161        runtime --no-jit option is given.
162    
163    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
164        ovector, the return from the function was zero, implying that there were
165        other matches that did not fit. The correct "exactly full" value is now
166        returned.
167    
168    6.  If a subpattern that was called recursively or as a subroutine contained
169        (*PRUNE) or any other control that caused it to give a non-standard return,
170        invalid errors such as "Error -26 (nested recursion at the same subject
171        position)" or even infinite loops could occur.
172    
173    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
174        computing the minimum length on reaching *ACCEPT, and so ended up with the
175        wrong value of 1 rather than 0. Further investigation indicates that
176        computing a minimum subject length in the presence of *ACCEPT is difficult
177        (think back references, subroutine calls), and so I have changed the code
178        so that no minimum is registered for a pattern that contains *ACCEPT.
179    
180    8.  If (*THEN) was present in the first (true) branch of a conditional group,
181        it was not handled as intended. [But see 16 below.]
182    
183    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
184        Sheri Pierce.
185    
186    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
187        the first byte in a match must be "a".
188    
189    11. Change 17 for 8.13 increased the recursion depth for patterns like
190        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
191        pattern contains any instances of (*THEN). If it does not, the old
192        optimizations are restored. It would be nice to do this on a per-group
193        basis, but at the moment that is not feasible.
194    
195    12. In some environments, the output of pcretest -C is CRLF terminated. This
196        broke RunTest's code that checks for the link size. A single white space
197        character after the value is now allowed for.
198    
199    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
200        For "fr", it uses the Windows-specific input and output files.
201    
202    14. If (*THEN) appeared in a group that was called recursively or as a
203        subroutine, it did not work as intended. [But see next item.]
204    
205    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
206        pattern fragments (but not containing any | characters). If A and B are
207        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
208        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
209        D. In other words, Perl considers parentheses that do not contain any |
210        characters to be part of a surrounding alternative, whereas PCRE was
211        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
212        differently. PCRE now behaves in the same way as Perl, except in the case
213        of subroutine/recursion calls such as (?1) which have in any case always
214        been different (but PCRE had them first :-).
215    
216    16. Related to 15 above: Perl does not treat the | in a conditional group as
217        creating alternatives. Such a group is treated in the same way as an
218        ordinary group without any | characters when processing (*THEN). PCRE has
219        been changed to match Perl's behaviour.
220    
221    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
222        RunGrepTest script failed.
223    
224    18. Change 22 for version 13 caused atomic groups to use more stack. This is
225        inevitable for groups that contain captures, but it can lead to a lot of
226        stack use in large patterns. The old behaviour has been restored for atomic
227        groups that do not contain any capturing parentheses.
228    
229    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
230        suppress the check for a minimum subject length at run time. (If it was
231        given to pcre_exec() or pcre_dfa_exec() it did work.)
232    
233    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
234        fail to work when decoding hex characters in data strings in EBCDIC
235        environments.
236    
237    21. It appears that in at least one Mac OS environment, the isxdigit() function
238        is implemented as a macro that evaluates to its argument more than once,
239        contravening the C 90 Standard (I haven't checked a later standard). There
240        was an instance in pcretest which caused it to go wrong when processing
241        \x{...} escapes in subject strings. The has been rewritten to avoid using
242        things like p++ in the argument of isxdigit().
243    
244    
245    Version 8.13 16-Aug-2011
246    ------------------------
247    
248    1.  The Unicode data tables have been updated to Unicode 6.0.0.
249    
250    2.  Two minor typos in pcre_internal.h have been fixed.
251    
252    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
253        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
254        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
255    
256    4.  There were a number of related bugs in the code for matching backrefences
257        caselessly in UTF-8 mode when codes for the characters concerned were
258        different numbers of bytes. For example, U+023A and U+2C65 are an upper
259        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
260        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
261        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
262        2-byte code at the end of the subject (it thought there wasn't enough data
263        left).
264    
265    5.  Comprehensive information about what went wrong is now returned by
266        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
267        as the output vector has at least 2 elements. The offset of the start of
268        the failing character and a reason code are placed in the vector.
269    
270    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
271        now returned is for the first byte of the failing character, instead of the
272        last byte inspected. This is an incompatible change, but I hope it is small
273        enough not to be a problem. It makes the returned offset consistent with
274        pcre_exec() and pcre_dfa_exec().
275    
276    7.  pcretest now gives a text phrase as well as the error number when
277        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
278        failure, the offset and reason code are output.
279    
280    8.  When \R was used with a maximizing quantifier it failed to skip backwards
281        over a \r\n pair if the subsequent match failed. Instead, it just skipped
282        back over a single character (\n). This seems wrong (because it treated the
283        two characters as a single entity when going forwards), conflicts with the
284        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
285        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
286        has been changed.
287    
288    9.  Some internal refactoring has changed the processing so that the handling
289        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
290        time (the PCRE_DOTALL option was changed this way some time ago: version
291        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
292        which was always a bit of a fudge. It also means that there is one less
293        argument for the match() function, which reduces its stack requirements
294        slightly. This change also fixes an incompatibility with Perl: the pattern
295        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
296    
297    10. More internal refactoring has drastically reduced the number of recursive
298        calls to match() for possessively repeated groups such as (abc)++ when
299        using pcre_exec().
300    
301    11. While implementing 10, a number of bugs in the handling of groups were
302        discovered and fixed:
303    
304        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
305        (a|)*(?1) gave a compile-time internal error.
306        ((a|)+)+  did not notice that the outer group could match an empty string.
307        (^a|^)+   was not marked as anchored.
308        (.*a|.*)+ was not marked as matching at start or after a newline.
309    
310    12. Yet more internal refactoring has removed another argument from the match()
311        function. Special calls to this function are now indicated by setting a
312        value in a variable in the "match data" data block.
313    
314    13. Be more explicit in pcre_study() instead of relying on "default" for
315        opcodes that mean there is no starting character; this means that when new
316        ones are added and accidentally left out of pcre_study(), testing should
317        pick them up.
318    
319    14. The -s option of pcretest has been documented for ages as being an old
320        synonym of -m (show memory usage). I have changed it to mean "force study
321        for every regex", that is, assume /S for every regex. This is similar to -i
322        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
323        using it. It makes it easier to run collections of tests with and without
324        study enabled, and thereby test pcre_study() more easily. All the standard
325        tests are now run with and without -s (but some patterns can be marked as
326        "never study" - see 20 below).
327    
328    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
329        restoration of the capturing data to the outer values was not happening
330        correctly.
331    
332    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
333        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
334        pattern had matched an empty string, and so incorrectly returned a no
335        match.
336    
337    17. There was optimizing code for the last branch of non-capturing parentheses,
338        and also for the obeyed branch of a conditional subexpression, which used
339        tail recursion to cut down on stack usage. Unfortunately, now that there is
340        the possibility of (*THEN) occurring in these branches, tail recursion is
341        no longer possible because the return has to be checked for (*THEN). These
342        two optimizations have therefore been removed. [But see 8.20/11 above.]
343    
344    18. If a pattern containing \R was studied, it was assumed that \R always
345        matched two bytes, thus causing the minimum subject length to be
346        incorrectly computed because \R can also match just one byte.
347    
348    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
349        was incorrectly computed.
350    
351    20. If /S is present twice on a test pattern in pcretest input, it now
352        *disables* studying, thereby overriding the use of -s on the command line
353        (see 14 above). This is necessary for one or two tests to keep the output
354        identical in both cases.
355    
356    21. When (*ACCEPT) was used in an assertion that matched an empty string and
357        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
358    
359    22. When an atomic group that contained a capturing parenthesis was
360        successfully matched, but the branch in which it appeared failed, the
361        capturing was not being forgotten if a higher numbered group was later
362        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
363        group 1 to "a", when in fact it should be unset. This applied to multi-
364        branched capturing and non-capturing groups, repeated or not, and also to
365        positive assertions (capturing in negative assertions does not happen
366        in PCRE) and also to nested atomic groups.
367    
368    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
369        subject after a captured substring, to make it easier to tell which of a
370        number of identical substrings has been captured.
371    
372    24. The way atomic groups are processed by pcre_exec() has been changed so that
373        if they are repeated, backtracking one repetition now resets captured
374        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
375        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
376        "aaa". Previously, it would have been "a". As part of this code
377        refactoring, the way recursive calls are handled has also been changed.
378    
379    25. If an assertion condition captured any substrings, they were not passed
380        back unless some other capturing happened later. For example, if
381        (?(?=(a))a) was matched against "a", no capturing was returned.
382    
383    26. When studying a pattern that contained subroutine calls or assertions,
384        the code for finding the minimum length of a possible match was handling
385        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
386        group 1 called group 2 while simultaneously a separate group 2 called group
387        1). A stack overflow occurred in this case. I have fixed this by limiting
388        the recursion depth to 10.
389    
390    27. Updated RunTest.bat in the distribution to the version supplied by Tom
391        Fortmann. This supports explicit test numbers on the command line, and has
392        argument validation and error reporting.
393    
394    28. An instance of \X with an unlimited repeat could fail if at any point the
395        first character it looked at was a mark character.
396    
397    29. Some minor code refactoring concerning Unicode properties and scripts
398        should reduce the stack requirement of match() slightly.
399    
400    30. Added the '=' option to pcretest to check the setting of unused capturing
401        slots at the end of the pattern, which are documented as being -1, but are
402        not included in the return count.
403    
404    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
405        compiled something random. Now it gives a compile-time error (as does
406        Perl).
407    
408    32. A *MARK encountered during the processing of a positive assertion is now
409        recorded and passed back (compatible with Perl).
410    
411    33. If --only-matching or --colour was set on a pcregrep call whose pattern
412        had alternative anchored branches, the search for a second match in a line
413        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
414        matched the line "0102" twice. The same bug affected patterns that started
415        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
416        twice.
417    
418    34. Previously, PCRE did not allow quantification of assertions. However, Perl
419        does, and because of capturing effects, quantifying parenthesized
420        assertions may at times be useful. Quantifiers are now allowed for
421        parenthesized assertions.
422    
423    35. A minor code tidy in pcre_compile() when checking options for \R usage.
424    
425    36. \g was being checked for fancy things in a character class, when it should
426        just be a literal "g".
427    
428    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
429        appearance of a nested POSIX class supersedes an apparent external class.
430        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
431        unescaped square brackets may also appear as part of class names. For
432        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
433        more like Perl. (But see 8.20/1 above.)
434    
435    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
436        was because it thought it was \N{name}, which is not supported).
437    
438    39. Add minix to OS list not supporting the -S option in pcretest.
439    
440    40. PCRE tries to detect cases of infinite recursion at compile time, but it
441        cannot analyze patterns in sufficient detail to catch mutual recursions
442        such as ((?1))((?2)). There is now a runtime test that gives an error if a
443        subgroup is called recursively as a subpattern for a second time at the
444        same position in the subject string. In previous releases this might have
445        been caught by the recursion limit, or it might have run out of stack.
446    
447    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
448        happen only once. PCRE was, however incorrectly giving a compile time error
449        "recursive call could loop indefinitely" because it cannot analyze the
450        pattern in sufficient detail. The compile time test no longer happens when
451        PCRE is compiling a conditional subpattern, but actual runaway loops are
452        now caught at runtime (see 40 above).
453    
454    42. It seems that Perl allows any characters other than a closing parenthesis
455        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
456        has been changed to be the same.
457    
458    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
459        as not to get warnings when autogen.sh is called. Also changed
460        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
461    
462    44. To help people who use pcregrep to scan files containing exceedingly long
463        lines, the following changes have been made:
464    
465        (a) The default value of the buffer size parameter has been increased from
466            8K to 20K. (The actual buffer used is three times this size.)
467    
468        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
469            PCRE is built.
470    
471        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
472            to be set at run time.
473    
474        (d) Numerical values in pcregrep options can be followed by K or M, for
475            example --buffer-size=50K.
476    
477        (e) If a line being scanned overflows pcregrep's buffer, an error is now
478            given and the return code is set to 2.
479    
480    45. Add a pointer to the latest mark to the callout data block.
481    
482    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
483        partial match of an empty string instead of no match. This was specific to
484        the use of ".".
485    
486    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
487        complete match instead of a partial match. This bug was dependent on both
488        the PCRE_UTF8 and PCRE_DOTALL options being set.
489    
490    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
491        starting byte set, because \b was not being ignored.
492    
493    
494    Version 8.12 15-Jan-2011
495    ------------------------
496    
497    1.  Fixed some typos in the markup of the man pages, and wrote a script that
498        checks for such things as part of the documentation building process.
499    
500    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
501        --match-limit and --recursion-limit options (added for 8.11). In
502        particular, this made one of the standard tests fail. (The integer value
503        went into the wrong half of a long int.)
504    
505    3.  If the --colour option was given to pcregrep with -v (invert match), it
506        did strange things, either producing crazy output, or crashing. It should,
507        of course, ignore a request for colour when reporting lines that do not
508        match.
509    
510    4.  Another pcregrep bug caused similar problems if --colour was specified with
511        -M (multiline) and the pattern match finished with a line ending.
512    
513    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
514        matched in multiline mode, the following line was shown as part of the
515        match. This seems wrong, so I have changed it.
516    
517    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
518        the check for further matches in the same line (so they could be coloured)
519        to overrun the end of the current line. If another match was found, it was
520        incorrectly shown (and then shown again when found in the next line).
521    
522    7.  If pcregrep was compiled under Windows, there was a reference to the
523        function pcregrep_exit() before it was defined. I am assuming this was
524        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
525        reported by a user. I've moved the definition above the reference.
526    
527    
528    Version 8.11 10-Dec-2010
529    ------------------------
530    
531    1.  (*THEN) was not working properly if there were untried alternatives prior
532        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
533        backtracked to try for "b" instead of moving to the next alternative branch
534        at the same level (in this case, to look for "c"). The Perl documentation
535        is clear that when (*THEN) is backtracked onto, it goes to the "next
536        alternative in the innermost enclosing group".
537    
538    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
539        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
540        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
541        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
542        (*THEN).
543    
544    3.  If \s appeared in a character class, it removed the VT character from
545        the class, even if it had been included by some previous item, for example
546        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
547        of \s, but is part of the POSIX "space" class.)
548    
549    4.  A partial match never returns an empty string (because you can always
550        match an empty string at the end of the subject); however the checking for
551        an empty string was starting at the "start of match" point. This has been
552        changed to the "earliest inspected character" point, because the returned
553        data for a partial match starts at this character. This means that, for
554        example, /(?<=abc)def/ gives a partial match for the subject "abc"
555        (previously it gave "no match").
556    
557    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
558        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
559        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
560        has an implication that the given string is incomplete (because a partial
561        match is preferred over a full match). For this reason, these items now
562        give a partial match in this situation. [Aside: previously, the one case
563        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
564        match rather than a full match, which was wrong by the old rules, but is
565        now correct.]
566    
567    6.  There was a bug in the handling of #-introduced comments, recognized when
568        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
569        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
570        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
571        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
572        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
573        places in pcre_compile().
574    
575    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
576        comments when looking ahead for named forward references to subpatterns,
577        the only newline sequence it recognized was NL. It now handles newlines
578        according to the set newline convention.
579    
580    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
581        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
582        cater for a lack of strerror(). These oversights have been fixed.
583    
584    9.  Added --match-limit and --recursion-limit to pcregrep.
585    
586    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
587    
588    11. When the -o option was used, pcregrep was setting a return code of 1, even
589        when matches were found, and --line-buffered was not being honoured.
590    
591    12. Added an optional parentheses number to the -o and --only-matching options
592        of pcregrep.
593    
594    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
595        can match an empty string. The code to do it in pcretest and pcredemo
596        needed fixing:
597    
598        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
599            only one byte after an empty string match just before CRLF (this case
600            just got forgotten; "any" and "anycrlf" were OK).
601    
602        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
603            mode when an empty string match preceded an ASCII character followed by
604            a non-ASCII character. (The code for advancing by one character rather
605            than one byte was nonsense.)
606    
607        (c) The pcredemo.c sample program did not have any code at all to handle
608            the cases when CRLF is a valid newline sequence.
609    
610    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
611        as a starting offset was within the subject string. There is now a new
612        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
613        negative or greater than the length of the string. In order to test this,
614        pcretest is extended to allow the setting of negative starting offsets.
615    
616    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
617        starting offset points to the beginning of a UTF-8 character was
618        unnecessarily clumsy. I tidied it up.
619    
620    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
621        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
622    
623    17. Nobody had reported that the --include_dir option, which was added in
624        release 7.7 should have been called --include-dir (hyphen, not underscore)
625        for compatibility with GNU grep. I have changed it to --include-dir, but
626        left --include_dir as an undocumented synonym, and the same for
627        --exclude-dir, though that is not available in GNU grep, at least as of
628        release 2.5.4.
629    
630    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
631        characters from a string of bytes) have been redefined so as not to use
632        loops, in order to improve performance in some environments. At the same
633        time, I abstracted some of the common code into auxiliary macros to save
634        repetition (this should not affect the compiled code).
635    
636    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
637        compile-time error is now given if \c is not followed by an ASCII
638        character, that is, a byte less than 128. (In EBCDIC mode, the code is
639        different, and any byte value is allowed.)
640    
641    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
642        START_OPTIMIZE option, which is now allowed at compile time - but just
643        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
644        to pcregrep and other applications that have no direct access to PCRE
645        options. The new /Y option in pcretest sets this option when calling
646        pcre_compile().
647    
648    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
649        back references. Groups containing recursive back references were forced to
650        be atomic by that change, but in the case of named groups, the amount of
651        memory required was incorrectly computed, leading to "Failed: internal
652        error: code overflow". This has been fixed.
653    
654    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
655        pcretest.c, to avoid build problems in some Borland environments.
656    
657    
658    Version 8.10 25-Jun-2010
659    ------------------------
660    
661    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
662        THEN.
663    
664    2.  (*ACCEPT) was not working when inside an atomic group.
665    
666    3.  Inside a character class, \B is treated as a literal by default, but
667        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
668        causes the error). The code is unchanged, but I tidied the documentation.
669    
670    4.  Inside a character class, PCRE always treated \R and \X as literals,
671        whereas Perl faults them if its -w option is set. I have changed PCRE so
672        that it faults them when PCRE_EXTRA is set.
673    
674    5.  Added support for \N, which always matches any character other than
675        newline. (It is the same as "." when PCRE_DOTALL is not set.)
676    
677    6.  When compiling pcregrep with newer versions of gcc which may have
678        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
679        declared with attribute warn_unused_result" were given. Just casting the
680        result to (void) does not stop the warnings; a more elaborate fudge is
681        needed. I've used a macro to implement this.
682    
683    7.  Minor change to pcretest.c to avoid a compiler warning.
684    
685    8.  Added four artifical Unicode properties to help with an option to make
686        \s etc use properties (see next item). The new properties are: Xan
687        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
688    
689    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
690        use Unicode properties. (*UCP) at the start of a pattern can be used to set
691        this option. Modified pcretest to add /W to test this facility. Added
692        REG_UCP to make it available via the POSIX interface.
693    
694    10. Added --line-buffered to pcregrep.
695    
696    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
697        studied, and the match started with a letter with a code point greater than
698        127 whose first byte was different to the first byte of the other case of
699        the letter, the other case of this starting letter was not recognized
700        (#976).
701    
702    12. If a pattern that was studied started with a repeated Unicode property
703        test, for example, \p{Nd}+, there was the theoretical possibility of
704        setting up an incorrect bitmap of starting bytes, but fortunately it could
705        not have actually happened in practice until change 8 above was made (it
706        added property types that matched character-matching opcodes).
707    
708    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
709        possible starting bytes for non-anchored patterns.
710    
711    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
712        \R, and also a number of cases that involve Unicode properties, both
713        explicit and implicit when PCRE_UCP is set.
714    
715    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
716        input, it could crash or give wrong results if characters with values
717        greater than 0xc0 were present in the subject string. (Detail: it assumed
718        UTF-8 input when processing these items.)
719    
720    16. Added a lot of (int) casts to avoid compiler warnings in systems where
721        size_t is 64-bit (#991).
722    
723    17. Added a check for running out of memory when PCRE is compiled with
724        --disable-stack-for-recursion (#990).
725    
726    18. If the last data line in a file for pcretest does not have a newline on
727        the end, a newline was missing in the output.
728    
729    19. The default pcre_chartables.c file recognizes only ASCII characters (values
730        less than 128) in its various bitmaps. However, there is a facility for
731        generating tables according to the current locale when PCRE is compiled. It
732        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
733        characters, are recognized by isspace() and therefore were getting set in
734        these tables, and indeed these tables seem to approximate to ISO 8859. This
735        caused a problem in UTF-8 mode when pcre_study() was used to create a list
736        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
737        which of course cannot start UTF-8 characters. I have changed the code so
738        that only real ASCII characters (less than 128) and the correct starting
739        bytes for UTF-8 encodings are set for characters greater than 127 when in
740        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
741        altogether.)
742    
743    20. Added the /T option to pcretest so as to be able to run tests with non-
744        standard character tables, thus making it possible to include the tests
745        used for 19 above in the standard set of tests.
746    
747    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
748        reference to a subpattern the other side of a comment that contains an
749        opening parenthesis caused either an internal compiling error, or a
750        reference to the wrong subpattern.
751    
752    
753    Version 8.02 19-Mar-2010
754    ------------------------
755    
756    1.  The Unicode data tables have been updated to Unicode 5.2.0.
757    
758    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
759        configured.
760    
761    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
762        original author of that file, following a query about its status.
763    
764    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
765        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
766    
767    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
768        quantifier applied to a forward-referencing subroutine call, could compile
769        incorrect code or give the error "internal error: previously-checked
770        referenced subpattern not found".
771    
772    6.  Both MS Visual Studio and Symbian OS have problems with initializing
773        variables to point to external functions. For these systems, therefore,
774        pcre_malloc etc. are now initialized to local functions that call the
775        relevant global functions.
776    
777    7.  There were two entries missing in the vectors called coptable and poptable
778        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
779        I've fixed the data, and added a kludgy way of testing at compile time that
780        the lengths are correct (equal to the number of opcodes).
781    
782    8.  Following on from 7, I added a similar kludge to check the length of the
783        eint vector in pcreposix.c.
784    
785    9.  Error texts for pcre_compile() are held as one long string to avoid too
786        much relocation at load time. To find a text, the string is searched,
787        counting zeros. There was no check for running off the end of the string,
788        which could happen if a new error number was added without updating the
789        string.
790    
791    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
792    
793    11. \K was not working if it appeared in an atomic group or in a group that
794        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
795        \K is "not well defined" if used in an assertion. PCRE now accepts it if
796        the assertion is positive, but not if it is negative.
797    
798    12. Change 11 fortuitously reduced the size of the stack frame used in the
799        "match()" function of pcre_exec.c by one pointer. Forthcoming
800        implementation of support for (*MARK) will need an extra pointer on the
801        stack; I have reserved it now, so that the stack frame size does not
802        decrease.
803    
804    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
805        item in branch that calls a recursion is a subroutine call - as in the
806        second branch in the above example - was incorrectly given the compile-
807        time error "recursive call could loop indefinitely" because pcre_compile()
808        was not correctly checking the subroutine for matching a non-empty string.
809    
810    14. The checks for overrunning compiling workspace could trigger after an
811        overrun had occurred. This is a "should never occur" error, but it can be
812        triggered by pathological patterns such as hundreds of nested parentheses.
813        The checks now trigger 100 bytes before the end of the workspace.
814    
815    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
816    
817    
818    Version 8.01 19-Jan-2010
819    ------------------------
820    
821    1.  If a pattern contained a conditional subpattern with only one branch (in
822        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
823        computed the wrong minimum data length (which is of course zero for such
824        subpatterns). This could cause incorrect "no match" results.
825    
826    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
827        the pattern is reset in the first branch, pcre_compile() failed with
828        "internal error: code overflow at offset...". This happened only when
829        the reset was to the original external option setting. (An optimization
830        abstracts leading options settings into an external setting, which was the
831        cause of this.)
832    
833    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
834        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
835        assertion pattern did not match (meaning that the assertion was true), it
836        was incorrectly treated as false if the SKIP had been reached during the
837        matching. This also applied to assertions used as conditions.
838    
839    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
840        assertion subpattern, including such a pattern used as a condition,
841        unpredictable results occurred, instead of the error return
842        PCRE_ERROR_DFA_UITEM.
843    
844    5.  The C++ GlobalReplace function was not working like Perl for the special
845        situation when an empty string is matched. It now does the fancy magic
846        stuff that is necessary.
847    
848    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
849        removed. (These were left over from very, very early versions of PCRE.)
850    
851    7.  Some cosmetic changes to the code to make life easier when compiling it
852        as part of something else:
853    
854        (a) Change DEBUG to PCRE_DEBUG.
855    
856        (b) In pcre_compile(), rename the member of the "branch_chain" structure
857            called "current" as "current_branch", to prevent a collision with the
858            Linux macro when compiled as a kernel module.
859    
860        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
861            prevent a collision with the Linux macro when compiled as a kernel
862            module.
863    
864    8.  In pcre_compile() there are some checks for integer overflows that used to
865        cast potentially large values to (double). This has been changed to that
866        when building, a check for int64_t is made, and if it is found, it is used
867        instead, thus avoiding the use of floating point arithmetic. (There is no
868        other use of FP in PCRE.) If int64_t is not found, the fallback is to
869        double.
870    
871    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
872        2005 (difference between two addresses compared to an unsigned value).
873    
874    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
875        custom one, because of the following reported problem in Windows:
876    
877          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
878              under Win32.
879          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
880              therefore missing the function definition.
881          - The compiler thus generates a "C" signature for the test function.
882          - The linker fails to find the "C" function.
883          - PCRE fails to configure if asked to do so against libbz2.
884    
885    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
886        messages were output:
887    
888          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
889          rerunning libtoolize, to keep the correct libtool macros in-tree.
890          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
891    
892        I have done both of these things.
893    
894    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
895        most of the time, it *can* run out if it is given a pattern that contains a
896        runaway infinite recursion. I updated the discussion in the pcrestack man
897        page.
898    
899    13. Now that we have gone to the x.xx style of version numbers, the minor
900        version may start with zero. Using 08 or 09 is a bad idea because users
901        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
902        interpreted as invalid octal numbers. I've updated the previous comment in
903        configure.ac, and also added a check that gives an error if 08 or 09 are
904        used.
905    
906    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
907        causing partial matching to fail when the end of the subject matched \W
908        in a UTF-8 pattern where \W was quantified with a minimum of 3.
909    
910    15. There were some discrepancies between the declarations in pcre_internal.h
911        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
912        their definitions. The declarations used "const uschar *" and the
913        definitions used USPTR. Even though USPTR is normally defined as "const
914        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
915        reported that: "This difference in casting confuses some C++ compilers, for
916        example, SunCC recognizes above declarations as different functions and
917        generates broken code for hbpcre." I have changed the declarations to use
918        USPTR.
919    
920    16. GNU libtool is named differently on some systems. The autogen.sh script now
921        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
922        (FreeBSD).
923    
924    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
925        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
926        comment: "Figure out how to create a longlong from a string: strtoll and
927        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
928        instance, but it only takes 2 args instead of 3!"
929    
930    18. A subtle bug concerned with back references has been fixed by a change of
931        specification, with a corresponding code fix. A pattern such as
932        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
933        refers, was giving matches when it shouldn't. For example, xa=xaaa would
934        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
935        same bug. Such groups have to be quantified to be useful, or contained
936        inside another quantified group. (If there's no repetition, the reference
937        can never match.) The problem arises because, having left the group and
938        moved on to the rest of the pattern, a later failure that backtracks into
939        the group uses the captured value from the final iteration of the group
940        rather than the correct earlier one. I have fixed this in PCRE by forcing
941        any group that contains a reference to itself to be an atomic group; that
942        is, there cannot be any backtracking into it once it has completed. This is
943        similar to recursive and subroutine calls.
944    
945    
946    Version 8.00 19-Oct-09
947    ----------------------
948    
949    1.  The table for translating pcre_compile() error codes into POSIX error codes
950        was out-of-date, and there was no check on the pcre_compile() error code
951        being within the table. This could lead to an OK return being given in
952        error.
953    
954    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
955        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
956        in a Windows environment.
957    
958    3.  The pcregrep --count option prints the count for each file even when it is
959        zero, as does GNU grep. However, pcregrep was also printing all files when
960        --files-with-matches was added. Now, when both options are given, it prints
961        counts only for those files that have at least one match. (GNU grep just
962        prints the file name in this circumstance, but including the count seems
963        more useful - otherwise, why use --count?) Also ensured that the
964        combination -clh just lists non-zero counts, with no names.
965    
966    4.  The long form of the pcregrep -F option was incorrectly implemented as
967        --fixed_strings instead of --fixed-strings. This is an incompatible change,
968        but it seems right to fix it, and I didn't think it was worth preserving
969        the old behaviour.
970    
971    5.  The command line items --regex=pattern and --regexp=pattern were not
972        recognized by pcregrep, which required --regex pattern or --regexp pattern
973        (with a space rather than an '='). The man page documented the '=' forms,
974        which are compatible with GNU grep; these now work.
975    
976    6.  No libpcreposix.pc file was created for pkg-config; there was just
977        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
978    
979    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
980        when UCP support is not needed, by modifying the Python script that
981        generates it from Unicode data files. This should not matter if the module
982        is correctly used as a library, but I received one complaint about 50K of
983        unwanted data. My guess is that the person linked everything into his
984        program rather than using a library. Anyway, it does no harm.
985    
986    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
987        was a minimum greater than 1 for a wide character in a possessive
988        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
989        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
990        character. Chaos in the form of incorrect output or a compiling loop could
991        result.
992    
993    9.  The restrictions on what a pattern can contain when partial matching is
994        requested for pcre_exec() have been removed. All patterns can now be
995        partially matched by this function. In addition, if there are at least two
996        slots in the offset vector, the offset of the earliest inspected character
997        for the match and the offset of the end of the subject are set in them when
998        PCRE_ERROR_PARTIAL is returned.
999    
1000    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1001        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1002        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1003        and may be more useful for multi-segment matching.
1004    
1005    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1006        used to be given if ever the end of the subject was reached; now it is
1007        given only if matching could not proceed because another character was
1008        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1009        string "Z", which now yields "no match" instead of "partial match". In the
1010        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1011        final character ended with (*FAIL).
1012    
1013    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1014        if the pattern had a "must contain" character that was already found in the
1015        earlier partial match, unless partial matching was again requested. For
1016        example, with the pattern /dog.(body)?/, the "must contain" character is
1017        "g". If the first part-match was for the string "dog", restarting with
1018        "sbody" failed. This bug has been fixed.
1019    
1020    13. The string returned by pcre_dfa_exec() after a partial match has been
1021        changed so that it starts at the first inspected character rather than the
1022        first character of the match. This makes a difference only if the pattern
1023        starts with a lookbehind assertion or \b or \B (\K is not supported by
1024        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1025        matching functions compatible, and I think it's the right thing to do.
1026    
1027    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1028        so that the demonstration program is easily available in environments where
1029        PCRE has not been installed from source.
1030    
1031    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1032        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1033        library.
1034    
1035    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1036        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1037        is not the first non-POSIX option to be added. Clearly some people find
1038        these options useful.
1039    
1040    17. If a caller to the POSIX matching function regexec() passes a non-zero
1041        value for nmatch with a NULL value for pmatch, the value of
1042        nmatch is forced to zero.
1043    
1044    18. RunGrepTest did not have a test for the availability of the -u option of
1045        the diff command, as RunTest does. It now checks in the same way as
1046        RunTest, and also checks for the -b option.
1047    
1048    19. If an odd number of negated classes containing just a single character
1049        interposed, within parentheses, between a forward reference to a named
1050        subpattern and the definition of the subpattern, compilation crashed with
1051        an internal error, complaining that it could not find the referenced
1052        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1053        [The bug was that it was starting one character too far in when skipping
1054        over the character class, thus treating the ] as data rather than
1055        terminating the class. This meant it could skip too much.]
1056    
1057    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1058        /g option in pcretest when the pattern contains \K, which makes it possible
1059        to have an empty string match not at the start, even when the pattern is
1060        anchored. Updated pcretest and pcredemo to use this option.
1061    
1062    21. If the maximum number of capturing subpatterns in a recursion was greater
1063        than the maximum at the outer level, the higher number was returned, but
1064        with unset values at the outer level. The correct (outer level) value is
1065        now given.
1066    
1067    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1068        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1069        make it do so. The string so far is captured, making this feature
1070        compatible with Perl.
1071    
1072    23. The tests have been re-organized, adding tests 11 and 12, to make it
1073        possible to check the Perl 5.10 features against Perl 5.10.
1074    
1075    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1076        pattern matches a fixed length string. PCRE did not allow this; now it
1077        does. Neither allows recursion.
1078    
1079    25. I finally figured out how to implement a request to provide the minimum
1080        length of subject string that was needed in order to match a given pattern.
1081        (It was back references and recursion that I had previously got hung up
1082        on.) This code has now been added to pcre_study(); it finds a lower bound
1083        to the length of subject needed. It is not necessarily the greatest lower
1084        bound, but using it to avoid searching strings that are too short does give
1085        some useful speed-ups. The value is available to calling programs via
1086        pcre_fullinfo().
1087    
1088    26. While implementing 25, I discovered to my embarrassment that pcretest had
1089        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1090        study optimizations had never been tested with that matching function.
1091        Oops. What is worse, even when it was passed study data, there was a bug in
1092        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1093        were also very few tests of studied patterns with pcre_dfa_exec().
1094    
1095    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1096        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1097        on the other side of the coin, they are no longer allowed to have different
1098        names, because these cannot be distinguished in PCRE, and this has caused
1099        confusion. (This is a difference from Perl.)
1100    
1101    28. When duplicate subpattern names are present (necessarily with different
1102        numbers, as required by 27 above), and a test is made by name in a
1103        conditional pattern, either for a subpattern having been matched, or for
1104        recursion in such a pattern, all the associated numbered subpatterns are
1105        tested, and the overall condition is true if the condition is true for any
1106        one of them. This is the way Perl works, and is also more like the way
1107        testing by number works.
1108    
1109    
1110    Version 7.9 11-Apr-09
1111    ---------------------
1112    
1113    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1114        (pcretest), all targets were linked against these libraries. This included
1115        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1116        libraries. This caused unwanted dependencies to be created. This problem
1117        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1118        pcretest is linked with readline.
1119    
1120    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1121        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1122        moved outside it again, because FALSE and TRUE are already defined in AIX,
1123        but BOOL is not.
1124    
1125    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1126        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1127    
1128    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1129        hyphens) following file names and line numbers when outputting matching
1130        lines. This is not true; no spaces are inserted. I have also clarified the
1131        wording for the --colour (or --color) option.
1132    
1133    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1134        was not coloured; this is different to GNU grep, so I have changed it to be
1135        the same.
1136    
1137    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1138        each matching line was coloured. Now it goes on to look for further matches
1139        of any of the test patterns, which is the same behaviour as GNU grep.
1140    
1141    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1142        doesn't make sense to accept an empty string match in pcregrep, so I have
1143        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1144        seems to be how GNU grep behaves.
1145    
1146    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1147        start or after a newline", because the conditional assertion was not being
1148        correctly handled. The rule now is that both the assertion and what follows
1149        in the first alternative must satisfy the test.
1150    
1151    9.  If auto-callout was enabled in a pattern with a conditional group whose
1152        condition was an assertion, PCRE could crash during matching, both with
1153        pcre_exec() and pcre_dfa_exec().
1154    
1155    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1156        used for matching.
1157    
1158    11. Unicode property support in character classes was not working for
1159        characters (bytes) greater than 127 when not in UTF-8 mode.
1160    
1161    12. Added the -M command line option to pcretest.
1162    
1163    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1164    
1165    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1166    
1167    16. Added comments and documentation about mis-use of no_arg in the C++
1168        wrapper.
1169    
1170    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1171        from Martin Jerabek that uses macro names for all relevant character and
1172        string constants.
1173    
1174    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1175        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1176        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1177        these, but not everybody uses configure.
1178    
1179    19. A conditional group that had only one branch was not being correctly
1180        recognized as an item that could match an empty string. This meant that an
1181        enclosing group might also not be so recognized, causing infinite looping
1182        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1183        with the subject "ab", where knowledge that the repeated group can match
1184        nothing is needed in order to break the loop.
1185    
1186    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1187        exec(), but without supplying a callout function, matching went wrong.
1188    
1189    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1190        leak if the size of the offset vector was greater than 30. When the vector
1191        is smaller, the saved offsets during recursion go onto a local stack
1192        vector, but for larger vectors malloc() is used. It was failing to free
1193        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1194        error, in fact).
1195    
1196    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1197        heapframe that is used only when UTF-8 support is enabled. This caused no
1198        problem, but was untidy.
1199    
1200    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1201        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1202        included within another project.
1203    
1204    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1205        slightly modified by me:
1206    
1207          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1208              not building pcregrep.
1209    
1210          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1211              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1212    
1213    25. Forward references, both numeric and by name, in patterns that made use of
1214        duplicate group numbers, could behave incorrectly or give incorrect errors,
1215        because when scanning forward to find the reference group, PCRE was not
1216        taking into account the duplicate group numbers. A pattern such as
1217        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1218    
1219    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1220        the feature of a custom pointer more persuasive (as requested by a user).
1221    
1222    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1223        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1224        pre-defined.
1225    
1226    28. Added support for (*UTF8) at the start of a pattern.
1227    
1228    29. Arrange for flags added by the "release type" setting in CMake to be shown
1229        in the configuration summary.
1230    
1231    
1232    Version 7.8 05-Sep-08
1233    ---------------------
1234    
1235    1.  Replaced UCP searching code with optimized version as implemented for Ad
1236        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1237        stage table and inline lookup instead of a function, giving speed ups of 2
1238        to 5 times on some simple patterns that I tested. Permission was given to
1239        distribute the MultiStage2.py script that generates the tables (it's not in
1240        the tarball, but is in the Subversion repository).
1241    
1242    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1243        scripts.
1244    
1245    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1246        a group with a zero qualifier. The result of the study could be incorrect,
1247        or the function might crash, depending on the pattern.
1248    
1249    4.  Caseless matching was not working for non-ASCII characters in back
1250        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1251        It now works when Unicode Property Support is available.
1252    
1253    5.  In pcretest, an escape such as \x{de} in the data was always generating
1254        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1255        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1256        truncation.
1257    
1258    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1259    
1260    7.  Added two (int) casts to pcregrep when printing the difference of two
1261        pointers, in case they are 64-bit values.
1262    
1263    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1264        test 2 if it fails.
1265    
1266    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1267        and a #define of that name to empty if it is not externally set. This is to
1268        allow users of MSVC to set it if necessary.
1269    
1270    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1271        the convenience functions in the pcre_get.c source file.
1272    
1273    11. An option change at the start of a pattern that had top-level alternatives
1274        could cause overwriting and/or a crash. This command provoked a crash in
1275        some environments:
1276    
1277          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1278    
1279        This potential security problem was recorded as CVE-2008-2371.
1280    
1281    12. For a pattern where the match had to start at the beginning or immediately
1282        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1283        pcre_dfa_exec() could read past the end of the passed subject if there was
1284        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1285        pcretest so that it places the subject at the end of its malloc-ed buffer.
1286    
1287    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1288        exec() might read past the end of the data buffer in UTF-8 mode.
1289    
1290    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1291        the data contained the byte 0x85 as part of a UTF-8 character within its
1292        first line. This applied both to normal and DFA matching.
1293    
1294    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1295        /^[^d]*?$/8 failed to match "abc".
1296    
1297    16. Added a missing copyright notice to pcrecpp_internal.h.
1298    
1299    17. Make it more clear in the documentation that values returned from
1300        pcre_exec() in ovector are byte offsets, not character counts.
1301    
1302    18. Tidied a few places to stop certain compilers from issuing warnings.
1303    
1304    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1305        supplied by Stefan Weber. I made a further small update for 7.8 because
1306        there is a change of source arrangements: the pcre_searchfuncs.c module is
1307        replaced by pcre_ucd.c.
1308    
1309    
1310    Version 7.7 07-May-08
1311    ---------------------
1312    
1313    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1314        a string to a long long, pretend we don't even have a long long." This is
1315        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1316    
1317    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1318        pre-7.6 versions, which defined a global no_arg variable instead of putting
1319        it in the RE class. (See also #8 below.)
1320    
1321    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1322        Lopes.
1323    
1324    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1325    
1326        (1) The include/exclude patterns were being applied to the whole pathnames
1327            of files, instead of just to the final components.
1328    
1329        (2) If there was more than one level of directory, the subdirectories were
1330            skipped unless they satisfied the include/exclude conditions. This is
1331            inconsistent with GNU grep (and could even be seen as contrary to the
1332            pcregrep specification - which I improved to make it absolutely clear).
1333            The action now is always to scan all levels of directory, and just
1334            apply the include/exclude patterns to regular files.
1335    
1336    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1337        --exclude_dir in the tests to avoid scanning .svn directories.
1338    
1339    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1340        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1341        doesn't support NULs in patterns.
1342    
1343    7.  Added some missing "const"s to declarations of static tables in
1344        pcre_compile.c and pcre_dfa_exec.c.
1345    
1346    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1347        caused by fix #2  above. (Subsequently also a second patch to fix the
1348        first patch. And a third patch - this was a messy problem.)
1349    
1350    9.  Applied Craig's patch to remove the use of push_back().
1351    
1352    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1353        matching function regexec().
1354    
1355    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1356        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1357        references. PCRE supports relative numbers with this syntax (I don't think
1358        Oniguruma does).
1359    
1360    12. Previously, a group with a zero repeat such as (...){0} was completely
1361        omitted from the compiled regex. However, this means that if the group
1362        was called as a subroutine from elsewhere in the pattern, things went wrong
1363        (an internal error was given). Such groups are now left in the compiled
1364        pattern, with a new opcode that causes them to be skipped at execution
1365        time.
1366    
1367    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1368        to the way PCRE behaves:
1369    
1370        (a) A lone ] character is dis-allowed (Perl treats it as data).
1371    
1372        (b) A back reference to an unmatched subpattern matches an empty string
1373            (Perl fails the current match path).
1374    
1375        (c) A data ] in a character class must be notated as \] because if the
1376            first data character in a class is ], it defines an empty class. (In
1377            Perl it is not possible to have an empty class.) The empty class []
1378            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1379            The negative empty class [^] matches any one character, independently
1380            of the DOTALL setting.
1381    
1382    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1383        non-existent subpattern following a character class starting with ']' and
1384        containing () gave an internal compiling error instead of "reference to
1385        non-existent subpattern". Fortunately, when the pattern did exist, the
1386        compiled code was correct. (When scanning forwards to check for the
1387        existencd of the subpattern, it was treating the data ']' as terminating
1388        the class, so got the count wrong. When actually compiling, the reference
1389        was subsequently set up correctly.)
1390    
1391    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1392        it was being rejected as not supported by pcre_dfa_exec(), even though
1393        other assertions are supported. I have made pcre_dfa_exec() support
1394        (*FAIL).
1395    
1396    16. The implementation of 13c above involved the invention of a new opcode,
1397        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1398        cannot be changed at match time, I realized I could make a small
1399        improvement to matching performance by compiling OP_ALLANY instead of
1400        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1401        on the OP_ANY path.
1402    
1403    17. Compiling pcretest on Windows with readline support failed without the
1404        following two fixes: (1) Make the unistd.h include conditional on
1405        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1406    
1407    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1408        ncurses library to be included for pcretest when ReadLine support is
1409        requested, but also to allow for it to be overridden. This patch came from
1410        Daniel Bergström.
1411    
1412    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1413        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1414        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1415        spotting this.
1416    
1417    
1418    Version 7.6 28-Jan-08
1419    ---------------------
1420    
1421    1.  A character class containing a very large number of characters with
1422        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1423        overflow.
1424    
1425    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1426        HAVE_LONG_LONG is not defined.
1427    
1428    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1429        bring them up to date and include new features. This patch includes:
1430    
1431        - Fixed PH's badly added libz and libbz2 support.
1432        - Fixed a problem with static linking.
1433        - Added pcredemo. [But later removed - see 7 below.]
1434        - Fixed dftables problem and added an option.
1435        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1436            HAVE_LONG_LONG.
1437        - Added readline support for pcretest.
1438        - Added an listing of the option settings after cmake has run.
1439    
1440    4.  A user submitted a patch to Makefile that makes it easy to create
1441        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1442        Makefile.am that cause it to include this special target, without
1443        affecting anything else. Note that the same mingw target plus all
1444        the other distribution libraries and programs are now supported
1445        when configuring with CMake (see 6 below) instead of with
1446        Configure/Make.
1447    
1448    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1449        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1450        exported in the Windows port". It has not yet been confirmed that the patch
1451        solves the problem, but it does no harm.
1452    
1453    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1454        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1455        with CMake, and also correct the comment about stack recursion.
1456    
1457    7.  Remove the automatic building of pcredemo from the ./configure system and
1458        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1459        of a program that users should build themselves after PCRE is installed, so
1460        building it automatically is not really right. What is more, it gave
1461        trouble in some build environments.
1462    
1463    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1464    
1465    
1466    Version 7.5 10-Jan-08
1467    ---------------------
1468    
1469    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1470        values in parens when parsing an RE using the C++ wrapper."
1471    
1472    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1473        Characters greater than 255 were excluded from the class instead of being
1474        included.
1475    
1476    3.  The same bug as (2) above applied to negated POSIX classes such as
1477        [:^space:].
1478    
1479    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1480        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1481        I have changed it.
1482    
1483    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1484        first named subpattern) and a construct such as (?&a) would reference the
1485        first named subpattern whose name started with "a" (in other words, the
1486        length check was missing). Both these problems are fixed. "Subpattern name
1487        expected" is now given for (?&) (a zero-length name), and this patch also
1488        makes it give the same error for \k'' (previously it complained that that
1489        was a reference to a non-existent subpattern).
1490    
1491    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1492        this is right because (?- can be followed by option settings as well as by
1493        digits. I have, however, made the messages clearer.
1494    
1495    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1496        than the number used in the conditional) now cause a compile-time error.
1497        This is actually not compatible with Perl, which accepts such patterns, but
1498        treats the conditional as always being FALSE (as PCRE used to), but it
1499        seems to me that giving a diagnostic is better.
1500    
1501    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1502        and messages.
1503    
1504    9.  Fix two occurrences of "backslash" in comments that should have been
1505        "backspace".
1506    
1507    10. Remove two redundant lines of code that can never be obeyed (their function
1508        was moved elsewhere).
1509    
1510    11. The program that makes PCRE's Unicode character property table had a bug
1511        which caused it to generate incorrect table entries for sequences of
1512        characters that have the same character type, but are in different scripts.
1513        It amalgamated them into a single range, with the script of the first of
1514        them. In other words, some characters were in the wrong script. There were
1515        thirteen such cases, affecting characters in the following ranges:
1516    
1517          U+002b0 - U+002c1
1518          U+0060c - U+0060d
1519          U+0061e - U+00612
1520          U+0064b - U+0065e
1521          U+0074d - U+0076d
1522          U+01800 - U+01805
1523          U+01d00 - U+01d77
1524          U+01d9b - U+01dbf
1525          U+0200b - U+0200f
1526          U+030fc - U+030fe
1527          U+03260 - U+0327f
1528          U+0fb46 - U+0fbb1
1529          U+10450 - U+1049d
1530    
1531    12. The -o option (show only the matching part of a line) for pcregrep was not
1532        compatible with GNU grep in that, if there was more than one match in a
1533        line, it showed only the first of them. It now behaves in the same way as
1534        GNU grep.
1535    
1536    13. If the -o and -v options were combined for pcregrep, it printed a blank
1537        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1538        does the same. The return code can be used to tell if there were any
1539        non-matching lines.
1540    
1541    14. Added --file-offsets and --line-offsets to pcregrep.
1542    
1543    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1544        infinitely looping recursion. The bug was that positive lookaheads were not
1545        being skipped when checking for a possible empty match (negative lookaheads
1546        and both kinds of lookbehind were skipped).
1547    
1548    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1549        inclusion of <windows.h> to before rather than after the definition of
1550        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1551    
1552    17. Specifying a possessive quantifier with a specific limit for a Unicode
1553        character property caused pcre_compile() to compile bad code, which led at
1554        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1555        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1556        caused the error; without that there was no problem.
1557    
1558    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1559    
1560    19. Added --enable-pcretest-libreadline.
1561    
1562    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1563        RE::GlobalReplace(). As a result, the number of replacements returned was
1564        double what it should be. I removed one of the increments, but Craig sent a
1565        later patch that removed the other one (the right fix) and added unit tests
1566        that check the return values (which was not done before).
1567    
1568    21. Several CMake things:
1569    
1570        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1571            the names libpcre and libpcreposix, not just pcre and pcreposix.
1572    
1573        (2) The above change means that pcretest and pcregrep are now correctly
1574            linked with the newly-built libraries, not previously installed ones.
1575    
1576        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1577    
1578    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1579        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1580        UTF-8 newline character). The key issue is that the pattern starts .*;
1581        this means that the match must be either at the beginning, or after a
1582        newline. The bug was in the code for advancing after a failed match and
1583        checking that the new position followed a newline. It was not taking
1584        account of UTF-8 characters correctly.
1585    
1586    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1587        character classes. PCRE was not treating the sequence [:...:] as a
1588        character class unless the ... were all letters. Perl, however, seems to
1589        allow any characters between [: and :], though of course it rejects as
1590        unknown any "names" that contain non-letters, because all the known class
1591        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1592        for example, whereas PCRE did not - it did not recognize a POSIX character
1593        class. This seemed a bit dangerous, so the code has been changed to be
1594        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1595        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1596        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1597        Perl does, and where it didn't before.
1598    
1599    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1600        Windows environments %n is disabled by default.
1601    
1602    
1603    Version 7.4 21-Sep-07
1604    ---------------------
1605    
1606    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1607        means that a class such as [\s] counted as "explicit reference to CR or
1608        LF". That isn't really right - the whole point of the change was to try to
1609        help when there was an actual mention of one of the two characters. So now
1610        the change happens only if \r or \n (or a literal CR or LF) character is
1611        encountered.
1612    
1613    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1614        of both had grown to the point where there were only 3 bits left.
1615        Fortunately, there was spare space in the data structure, and so I have
1616        moved the internal flags into a new 16-bit field to free up more option
1617        bits.
1618    
1619    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1620        but did not set the internal JCHANGED flag - either of these is enough to
1621        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1622        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1623        start sets both bits.
1624    
1625    4.  Added options (at build time, compile time, exec time) to change \R from
1626        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1627    
1628    5.  doc/pcresyntax.html was missing from the distribution.
1629    
1630    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1631        compatibility, even though it is no longer used.
1632    
1633    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1634        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1635        windows.h file is present (where different names are used). [This was
1636        reversed later after testing - see 16 below.]
1637    
1638    8.  Changed all #include <config.h> to #include "config.h". There were also
1639        some further <pcre.h> cases that I changed to "pcre.h".
1640    
1641    9.  When pcregrep was used with the --colour option, it missed the line ending
1642        sequence off the lines that it output.
1643    
1644    10. It was pointed out to me that arrays of string pointers cause lots of
1645        relocations when a shared library is dynamically loaded. A technique of
1646        using a single long string with a table of offsets can drastically reduce
1647        these. I have refactored PCRE in four places to do this. The result is
1648        dramatic:
1649    
1650          Originally:                          290
1651          After changing UCP table:            187
1652          After changing error message table:   43
1653          After changing table of "verbs"       36
1654          After changing table of Posix names   22
1655    
1656        Thanks to the folks working on Gregex for glib for this insight.
1657    
1658    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1659        unicode-properties was also set.
1660    
1661    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1662    
1663    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1664        checked only for CRLF.
1665    
1666    14. Added casts to pcretest.c to avoid compiler warnings.
1667    
1668    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1669    
1670    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1671        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1672        entirely. This removes changes made in 7 above.
1673    
1674    17. The CMake files have been updated, and there is now more information about
1675        building with CMake in the NON-UNIX-USE document.
1676    
1677    
1678    Version 7.3 28-Aug-07
1679    ---------------------
1680    
1681     1. In the rejigging of the build system that eventually resulted in 7.1, the
1682        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1683        brackets there is not right, since it causes compilers to look for an
1684        installed pcre.h, not the version that is in the source that is being
1685        compiled (which of course may be different). I have changed it back to:
1686    
1687          #include "pcre.h"
1688    
1689        I have a vague recollection that the change was concerned with compiling in
1690        different directories, but in the new build system, that is taken care of
1691        by the VPATH setting the Makefile.
1692    
1693     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1694        when the subject happened to end in the byte 0x85 (e.g. if the last
1695        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1696        characters but of course it shouldn't be taken as a newline when it is part
1697        of another character. The bug was that, for an unlimited repeat of . in
1698        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1699        characters when looking for a newline.
1700    
1701     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1702    
1703     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1704        in debug output.
1705    
1706     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1707        long printing in the pcrecpp unittest when running under MinGW.
1708    
1709     6. ESC_K was left out of the EBCDIC table.
1710    
1711     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1712        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1713        limit also applies to "virtual nesting" when a pattern is recursive, and in
1714        this case 1000 isn't so big. I have been able to remove this limit at the
1715        expense of backing off one optimization in certain circumstances. Normally,
1716        when pcre_exec() would call its internal match() function recursively and
1717        immediately return the result unconditionally, it uses a "tail recursion"
1718        feature to save stack. However, when a subpattern that can match an empty
1719        string has an unlimited repetition quantifier, it no longer makes this
1720        optimization. That gives it a stack frame in which to save the data for
1721        checking that an empty string has been matched. Previously this was taken
1722        from the 1000-entry workspace that had been reserved. So now there is no
1723        explicit limit, but more stack is used.
1724    
1725     8. Applied Daniel's patches to solve problems with the import/export magic
1726        syntax that is required for Windows, and which was going wrong for the
1727        pcreposix and pcrecpp parts of the library. These were overlooked when this
1728        problem was solved for the main library.
1729    
1730     9. There were some crude static tests to avoid integer overflow when computing
1731        the size of patterns that contain repeated groups with explicit upper
1732        limits. As the maximum quantifier is 65535, the maximum group length was
1733        set at 30,000 so that the product of these two numbers did not overflow a
1734        32-bit integer. However, it turns out that people want to use groups that
1735        are longer than 30,000 bytes (though not repeat them that many times).
1736        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1737        made it possible to implement the integer overflow checks in a much more
1738        dynamic way, which I have now done. The artificial limitation on group
1739        length has been removed - we now have only the limit on the total length of
1740        the compiled pattern, which depends on the LINK_SIZE setting.
1741    
1742    10. Fixed a bug in the documentation for get/copy named substring when
1743        duplicate names are permitted. If none of the named substrings are set, the
1744        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1745        empty string.
1746    
1747    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1748        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1749        because the ] is interpreted as the first data character and the
1750        terminating ] is not found. PCRE has been made compatible with Perl in this
1751        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1752        cause memory overwriting.
1753    
1754    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1755        string has been matched (to stop an infinite loop). It was not recognizing
1756        a conditional subpattern that could match an empty string if that
1757        subpattern was within another subpattern. For example, it looped when
1758        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1759        condition was not nested. This bug has been fixed.
1760    
1761    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1762        past the start of the subject in the presence of bytes with the top bit
1763        set, for example "\x8aBCD".
1764    
1765    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1766        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1767    
1768    14. Optimized (?!) to (*FAIL).
1769    
1770    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1771        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1772        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1773        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1774        does: it's just the validity check that is more restrictive.
1775    
1776    16. Inserted checks for integer overflows during escape sequence (backslash)
1777        processing, and also fixed erroneous offset values for syntax errors during
1778        backslash processing.
1779    
1780    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1781        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1782    
1783    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1784        caused an overrun.
1785    
1786    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1787        something other than just ASCII characters) inside a group that had an
1788        unlimited repeat caused a loop at compile time (while checking to see
1789        whether the group could match an empty string).
1790    
1791    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1792        [\P{Any}] did so. (Error in the code for printing property names.)
1793    
1794    21. An orphan \E inside a character class could cause a crash.
1795    
1796    22. A repeated capturing bracket such as (A)? could cause a wild memory
1797        reference during compilation.
1798    
1799    23. There are several functions in pcre_compile() that scan along a compiled
1800        expression for various reasons (e.g. to see if it's fixed length for look
1801        behind). There were bugs in these functions when a repeated \p or \P was
1802        present in the pattern. These operators have additional parameters compared
1803        with \d, etc, and these were not being taken into account when moving along
1804        the compiled data. Specifically:
1805    
1806        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1807            length.
1808    
1809        (b) An item such as \pL+ within a repeated group could cause crashes or
1810            loops.
1811    
1812        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1813            "reference to non-existent subpattern" error.
1814    
1815        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1816    
1817    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1818        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1819    
1820    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1821    
1822    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1823        character were causing crashes (broken optimization).
1824    
1825    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1826        \p or \P) caused a compile-time loop.
1827    
1828    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1829        break. For example, the unstudied pattern [\r\n]A does not match the string
1830        "\r\nA" because change 7.0/46 below moves the current point on by two
1831        characters after failing to match at the start. However, the pattern \nA
1832        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1833        the same is true. There doesn't seem any very clean way out of this, but
1834        what I have chosen to do makes the common cases work: PCRE now takes note
1835        of whether there can be an explicit match for \r or \n anywhere in the
1836        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1837        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1838        pattern has explicit CR or LF references.
1839    
1840    29. Added (*CR) etc for changing newline setting at start of pattern.
1841    
1842    
1843    Version 7.2 19-Jun-07
1844    ---------------------
1845    
1846     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1847        which is apparently normally available under Windows.
1848    
1849     2. Re-jig the pcregrep tests with different newline settings in an attempt
1850        to make them independent of the local environment's newline setting.
1851    
1852     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1853    
1854     4. Some of the "internals" tests were previously cut out when the link size
1855        was not 2, because the output contained actual offsets. The recent new
1856        "Z" feature of pcretest means that these can be cut out, making the tests
1857        usable with all link sizes.
1858    
1859     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1860        stack recursion. This gives a massive performance boost under BSD, but just
1861        a small improvement under Linux. However, it saves one field in the frame
1862        in all cases.
1863    
1864     6. Added more features from the forthcoming Perl 5.10:
1865    
1866        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1867            recursion call. It refers to the nth most recently opened parentheses.
1868    
1869        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1870            to be opened parentheses.
1871    
1872        (c) Conditions that refer to capturing parentheses can be specified
1873            relatively, for example, (?(-2)... or (?(+3)...
1874    
1875        (d) \K resets the start of the current match so that everything before
1876            is not part of it.
1877    
1878        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1879    
1880        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1881            reference syntax.
1882    
1883        (g) (?| introduces a group in which the numbering of parentheses in each
1884            alternative starts with the same number.
1885    
1886        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1887    
1888     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1889        PCRE_INFO_JCHANGED.
1890    
1891     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1892        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1893        for detecting groups that can match an empty string.
1894    
1895     9. A pattern with a very large number of alternatives (more than several
1896        hundred) was running out of internal workspace during the pre-compile
1897        phase, where pcre_compile() figures out how much memory will be needed. A
1898        bit of new cunning has reduced the workspace needed for groups with
1899        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1900        workspace instead of running out of the 4096 that are available.
1901    
1902    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1903    
1904    11. Applied patch from Google to remove an optimization that didn't quite work.
1905        The report of the bug said:
1906    
1907          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1908          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1909          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1910    
1911    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1912        it matched the wrong number of bytes.
1913    
1914    
1915    Version 7.1 24-Apr-07
1916    ---------------------
1917    
1918     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1919        that is more "standard", making use of automake and other Autotools. There
1920        is some re-arrangement of the files and adjustment of comments consequent
1921        on this.
1922    
1923     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1924        for recursive directory scanning broke on some systems because the files
1925        are not scanned in any specific order and on different systems the order
1926        was different. A call to "sort" has been inserted into RunGrepTest for the
1927        approprate test as a short-term fix. In the longer term there may be an
1928        alternative.
1929    
1930     3. I had an email from Eric Raymond about problems translating some of PCRE's
1931        man pages to HTML (despite the fact that I distribute HTML pages, some
1932        people do their own conversions for various reasons). The problems
1933        concerned the use of low-level troff macros .br and .in. I have therefore
1934        removed all such uses from the man pages (some were redundant, some could
1935        be replaced by .nf/.fi pairs). The 132html script that I use to generate
1936        HTML has been updated to handle .nf/.fi and to complain if it encounters
1937        .br or .in.
1938    
1939     4. Updated comments in configure.ac that get placed in config.h.in and also
1940        arranged for config.h to be included in the distribution, with the name
1941        config.h.generic, for the benefit of those who have to compile without
1942        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1943    
1944     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1945        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1946        makevp.bat for the current PCRE, using the additional files
1947        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1948    
1949     6. A Windows user reported a minor discrepancy with test 2, which turned out
1950        to be caused by a trailing space on an input line that had got lost in his
1951        copy. The trailing space was an accident, so I've just removed it.
1952    
1953     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1954        that is needed.
1955    
1956     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1957        as "const" (a) because they are and (b) because it helps the PHP
1958        maintainers who have recently made a script to detect big data structures
1959        in the php code that should be moved to the .rodata section. I remembered
1960        to update Builducptable as well, so it won't revert if ucptable.h is ever
1961        re-created.
1962    
1963     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1964        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1965        order to be able to cut out the UTF-8 tables in the latter when UTF-8
1966        support is not required. This saves 1.5-2K of code, which is important in
1967        some applications.
1968    
1969        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1970        so as not to refer to the tables, even though these functions will never be
1971        called when UTF-8 support is disabled. Otherwise there are problems with a
1972        shared library.
1973    
1974    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1975    
1976        (a) It was defining its arguments as char * instead of void *.
1977    
1978        (b) It was assuming that all moves were upwards in memory; this was true
1979            a long time ago when I wrote it, but is no longer the case.
1980    
1981        The emulated memove() is provided for those environments that have neither
1982        memmove() nor bcopy(). I didn't think anyone used it these days, but that
1983        is clearly not the case, as these two bugs were recently reported.
1984    
1985    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1986        and Detrail to create the HTML documentation, the .txt form of the man
1987        pages, and it removes trailing spaces from listed files. It also creates
1988        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1989        case, it wraps all the #defines with #ifndefs. This script should be run
1990        before "make dist".
1991    
1992    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1993        with Unicode property support.
1994    
1995        (a) For a maximizing quantifier, if the two different cases of the
1996            character were of different lengths in their UTF-8 codings (there are
1997            some cases like this - I found 11), and the matching function had to
1998            back up over a mixture of the two cases, it incorrectly assumed they
1999            were both the same length.
2000    
2001        (b) When PCRE was configured to use the heap rather than the stack for
2002            recursion during matching, it was not correctly preserving the data for
2003            the other case of a UTF-8 character when checking ahead for a match
2004            while processing a minimizing repeat. If the check also involved
2005            matching a wide character, but failed, corruption could cause an
2006            erroneous result when trying to check for a repeat of the original
2007            character.
2008    
2009    13. Some tidying changes to the testing mechanism:
2010    
2011        (a) The RunTest script now detects the internal link size and whether there
2012            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2013            values substituted by "configure". (The RunGrepTest script already did
2014            this for UTF-8.) The configure.ac script no longer substitutes the
2015            relevant variables.
2016    
2017        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2018            with length and offset values. This means that the output is different
2019            for different internal link sizes. Test 2 is skipped for link sizes
2020            other than 2 because of this, bypassing the problem. Unfortunately,
2021            there was also a test in test 3 (the locale tests) that used /B and
2022            failed for link sizes other than 2. Rather than cut the whole test out,
2023            I have added a new /Z option to pcretest that replaces the length and
2024            offset values with spaces. This is now used to make test 3 independent
2025            of link size. (Test 2 will be tidied up later.)
2026    
2027    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2028        segmentation fault instead of returning the appropriate error message.
2029    
2030    15. In multiline mode when the newline sequence was set to "any", the pattern
2031        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2032        This doesn't seem right; it now treats the CRLF combination as the line
2033        ending, and so does not match in that case. It's only a pattern such as ^$
2034        that would hit this one: something like ^ABC$ would have failed after \r
2035        and then tried again after \r\n.
2036    
2037    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2038        in an attempt to make files that differ only in their line terminators
2039        compare equal. This works on Linux.
2040    
2041    17. Under certain error circumstances pcregrep might try to free random memory
2042        as it exited. This is now fixed, thanks to valgrind.
2043    
2044    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2045        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2046        was because its rules for how to advance for /g after matching an empty
2047        string at the end of a line did not allow for this case. They now check for
2048        it specially.
2049    
2050    20. pcretest is supposed to handle patterns and data of any length, by
2051        extending its buffers when necessary. It was getting this wrong when the
2052        buffer for a data line had to be extended.
2053    
2054    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2055        CRLF as a newline sequence.
2056    
2057    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2058        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2059        I have nevertheless tidied it up.
2060    
2061    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2062    
2063    24. Added a man page for pcre-config.
2064    
2065    
2066    Version 7.0 19-Dec-06
2067    ---------------------
2068    
2069     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2070        moving to gcc 4.1.1.
2071    
2072     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2073        sys/time.h, which is documented as needed for this function. It doesn't
2074        seem to matter on Linux, but it showed up on some releases of OS X.
2075    
2076     3. It seems that there are systems where bytes whose values are greater than
2077        127 match isprint() in the "C" locale. The "C" locale should be the
2078        default when a C program starts up. In most systems, only ASCII printing
2079        characters match isprint(). This difference caused the output from pcretest
2080        to vary, making some of the tests fail. I have changed pcretest so that:
2081    
2082        (a) When it is outputting text in the compiled version of a pattern, bytes
2083            other than 32-126 are always shown as hex escapes.
2084    
2085        (b) When it is outputting text that is a matched part of a subject string,
2086            it does the same, unless a different locale has been set for the match
2087            (using the /L modifier). In this case, it uses isprint() to decide.
2088    
2089     4. Fixed a major bug that caused incorrect computation of the amount of memory
2090        required for a compiled pattern when options that changed within the
2091        pattern affected the logic of the preliminary scan that determines the
2092        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2093        that the computed length was too small. The symptoms of this bug were
2094        either the PCRE error "internal error: code overflow" from pcre_compile(),
2095        or a glibc crash with a message such as "pcretest: free(): invalid next
2096        size (fast)". Examples of patterns that provoked this bug (shown in
2097        pcretest format) are:
2098    
2099          /(?-x: )/x
2100          /(?x)(?-x: \s*#\s*)/
2101          /((?i)[\x{c0}])/8
2102          /(?i:[\x{c0}])/8
2103    
2104        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2105        is now done differently.
2106    
2107     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2108        wrapper classes; (b) implement a new function in the C++ scanner that is
2109        more efficient than the old way of doing things because it avoids levels of
2110        recursion in the regex matching; (c) add a paragraph to the documentation
2111        for the FullMatch() function.
2112    
2113     6. The escape sequence \n was being treated as whatever was defined as
2114        "newline". Not only was this contrary to the documentation, which states
2115        that \n is character 10 (hex 0A), but it also went horribly wrong when
2116        "newline" was defined as CRLF. This has been fixed.
2117    
2118     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2119        was being set to -1 for the "end of line" case (supposedly a value that no
2120        character can have). Though this value is never used (the check for end of
2121        line is "zero bytes in current character"), it caused compiler complaints.
2122        I've changed it to 0xffffffff.
2123    
2124     8. In pcre_version.c, the version string was being built by a sequence of
2125        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2126        string (as it is for production releases) called a macro with an empty
2127        argument. The C standard says the result of this is undefined. The gcc
2128        compiler treats it as an empty string (which was what was wanted) but it is
2129        reported that Visual C gives an error. The source has been hacked around to
2130        avoid this problem.
2131    
2132     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2133        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2134        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2135        of them did).
2136    
2137    10. Originally, pcretest opened its input and output without "b"; then I was
2138        told that "b" was needed in some environments, so it was added for release
2139        5.0 to both the input and output. (It makes no difference on Unix-like
2140        systems.) Later I was told that it is wrong for the input on Windows. I've
2141        now abstracted the modes into two macros, to make it easier to fiddle with
2142        them, and removed "b" from the input mode under Windows.
2143    
2144    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2145    
2146    12. Added -help and --help to pcretest as an official way of being reminded
2147        of the options.
2148    
2149    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2150        and pcrecpp.cc because they annoy compilers at high warning levels.
2151    
2152    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2153    
2154    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2155        scripts are not C programs :-) and which was not noticed because it works
2156        on Linux.
2157    
2158    16. pcretest is supposed to handle any length of pattern and data line (as one
2159        line or as a continued sequence of lines) by extending its input buffer if
2160        necessary. This feature was broken for very long pattern lines, leading to
2161        a string of junk being passed to pcre_compile() if the pattern was longer
2162        than about 50K.
2163    
2164    17. I have done a major re-factoring of the way pcre_compile() computes the
2165        amount of memory needed for a compiled pattern. Previously, there was code
2166        that made a preliminary scan of the pattern in order to do this. That was
2167        OK when PCRE was new, but as the facilities have expanded, it has become
2168        harder and harder to keep it in step with the real compile phase, and there
2169        have been a number of bugs (see for example, 4 above). I have now found a
2170        cunning way of running the real compile function in a "fake" mode that
2171        enables it to compute how much memory it would need, while actually only
2172        ever using a few hundred bytes of working memory and without too many
2173        tests of the mode. This should make future maintenance and development
2174        easier. A side effect of this work is that the limit of 200 on the nesting
2175        depth of parentheses has been removed (though this was never a serious
2176        limitation, I suspect). However, there is a downside: pcre_compile() now
2177        runs more slowly than before (30% or more, depending on the pattern). I
2178        hope this isn't a big issue. There is no effect on runtime performance.
2179    
2180    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2181        newline (only possible for the last line of a file) and it was a
2182        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2183    
2184    19. Added additional timing features to pcretest. (1) The -tm option now times
2185        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2186        separate command line item, by a number that specifies the number of
2187        repeats to use when timing. The default is 50000; this gives better
2188        precision, but takes uncomfortably long for very large patterns.
2189    
2190    20. Extended pcre_study() to be more clever in cases where a branch of a
2191        subpattern has no definite first character. For example, (a*|b*)[cd] would
2192        previously give no result from pcre_study(). Now it recognizes that the
2193        first character must be a, b, c, or d.
2194    
2195    21. There was an incorrect error "recursive call could loop indefinitely" if
2196        a subpattern (or the entire pattern) that was being tested for matching an
2197        empty string contained only one non-empty item after a nested subpattern.
2198        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2199        incorrectly, because the \d was being skipped in the check.
2200    
2201    22. The pcretest program now has a new pattern option /B and a command line
2202        option -b, which is equivalent to adding /B to every pattern. This causes
2203        it to show the compiled bytecode, without the additional information that
2204        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2205        is the same as /B/I).
2206    
2207    23. A new optimization is now able automatically to treat some sequences such
2208        as a*b as a*+b. More specifically, if something simple (such as a character
2209        or a simple class like \d) has an unlimited quantifier, and is followed by
2210        something that cannot possibly match the quantified thing, the quantifier
2211        is automatically "possessified".
2212    
2213    24. A recursive reference to a subpattern whose number was greater than 39
2214        went wrong under certain circumstances in UTF-8 mode. This bug could also
2215        have affected the operation of pcre_study().
2216    
2217    25. Realized that a little bit of performance could be had by replacing
2218        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2219    
2220    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2221    
2222    27. Possessive quantifiers such as a++ were previously implemented by turning
2223        them into atomic groups such as ($>a+). Now they have their own opcodes,
2224        which improves performance. This includes the automatically created ones
2225        from 23 above.
2226    
2227    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2228        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2229        the first matched character to be a colon. This applied both to named and
2230        numbered groups.
2231    
2232    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2233    
2234    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2235        building PCRE on Windows easier, so I have included it in the distribution.
2236    
2237    31. There is now a check in pcretest against a ridiculously large number being
2238        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2239        loop, the loop is abandoned.
2240    
2241    32. Forward references to subpatterns in conditions such as (?(2)...) where
2242        subpattern 2 is defined later cause pcre_compile() to search forwards in
2243        the pattern for the relevant set of parentheses. This search went wrong
2244        when there were unescaped parentheses in a character class, parentheses
2245        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2246    
2247    33. "Subroutine" calls and backreferences were previously restricted to
2248        referencing subpatterns earlier in the regex. This restriction has now
2249        been removed.
2250    
2251    34. Added a number of extra features that are going to be in Perl 5.10. On the
2252        whole, these are just syntactic alternatives for features that PCRE had
2253        previously implemented using the Python syntax or my own invention. The
2254        other formats are all retained for compatibility.
2255    
2256        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2257            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2258            also .NET compatible.
2259    
2260        (b) A recursion or subroutine call to a named group can now be defined as
2261            (?&name) as well as (?P>name).
2262    
2263        (c) A backreference to a named group can now be defined as \k<name> or
2264            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2265            5.10, are also .NET compatible.
2266    
2267        (d) A conditional reference to a named group can now use the syntax
2268            (?(<name>) or (?('name') as well as (?(name).
2269    
2270        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2271            groups (named and numbered) that are never evaluated inline, but can be
2272            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2273            is always false. There may be only one alternative in such a group.
2274    
2275        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2276            as the simple (?(R). The condition is true only if the most recent
2277            recursion is that of the given number or name. It does not search out
2278            through the entire recursion stack.
2279    
2280        (g) The escape \gN or \g{N} has been added, where N is a positive or
2281            negative number, specifying an absolute or relative reference.
2282    
2283    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2284        some "unreachable code" warnings.
2285    
2286    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2287        things, this adds five new scripts.
2288    
2289    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2290        There were also incompatibilities regarding the handling of \Q..\E inside
2291        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2292        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2293    
2294    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2295        matches an empty string, and forcibly breaks the loop. There were bugs in
2296        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2297        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2298        separate and independent bugs (that affected different cases) have been
2299        fixed.
2300    
2301    39. Refactored the code to abolish the use of different opcodes for small
2302        capturing bracket numbers. This is a tidy that I avoided doing when I
2303        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2304        The new approach is not only tidier, it makes it possible to reduce the
2305        memory needed to fix the previous bug (38).
2306    
2307    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2308        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2309        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2310        mode.
2311    
2312    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2313        report.
2314    
2315    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2316        copy construction and assignment in the C++ wrapper.
2317    
2318    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2319        couple of bugs that could have given wrong results in the "--newline=crlf"
2320        case.
2321    
2322    44. Added a number of casts and did some reorganization of signed/unsigned int
2323        variables following suggestions from Dair Grant. Also renamed the variable
2324        "this" as "item" because it is a C++ keyword.
2325    
2326    45. Arranged for dftables to add
2327    
2328          #include "pcre_internal.h"
2329    
2330        to pcre_chartables.c because without it, gcc 4.x may remove the array
2331        definition from the final binary if PCRE is built into a static library and
2332        dead code stripping is activated.
2333    
2334    46. For an unanchored pattern, if a match attempt fails at the start of a
2335        newline sequence, and the newline setting is CRLF or ANY, and the next two
2336        characters are CRLF, advance by two characters instead of one.
2337    
2338    
2339    Version 6.7 04-Jul-06
2340    ---------------------
2341    
2342     1. In order to handle tests when input lines are enormously long, pcretest has
2343        been re-factored so that it automatically extends its buffers when
2344        necessary. The code is crude, but this _is_ just a test program. The
2345        default size has been increased from 32K to 50K.
2346    
2347     2. The code in pcre_study() was using the value of the re argument before
2348        testing it for NULL. (Of course, in any sensible call of the function, it
2349        won't be NULL.)
2350    
2351     3. The memmove() emulation function in pcre_internal.h, which is used on
2352        systems that lack both memmove() and bcopy() - that is, hardly ever -
2353        was missing a "static" storage class specifier.
2354    
2355     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2356        containing an extended class (one that cannot be represented by a bitmap
2357        because it contains high-valued characters or Unicode property items, e.g.
2358        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2359        pattern, but PCRE should not loop if you do not (it no longer does).
2360        [Detail: two cases were found: (a) a repeated subpattern containing an
2361        extended class; (b) a recursive reference to a subpattern that followed a
2362        previous extended class. It wasn't skipping over the extended class
2363        correctly when UTF-8 mode was not set.]
2364    
2365     5. A negated single-character class was not being recognized as fixed-length
2366        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2367        compile error "lookbehind assertion is not fixed length".
2368    
2369     6. The RunPerlTest auxiliary script was showing an unexpected difference
2370        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2371        write a Perl script that can interpret lines of an input file either as
2372        byte characters or as UTF-8, which is what "perltest" was being required to
2373        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2374        can't do is switch easily at run time between having the "use utf8;" pragma
2375        or not. In the end, I fudged it by using the RunPerlTest script to insert
2376        "use utf8;" explicitly for the UTF-8 tests.
2377    
2378     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2379        the end of the subject string, contrary to the documentation and to what
2380        Perl does. This was true of both matching functions. Now it matches only at
2381        the start of the subject and immediately after *internal* newlines.
2382    
2383     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2384        a pointer to an int instead of a pointer to an unsigned long int. This
2385        caused problems on 64-bit systems.
2386    
2387     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2388        instance of the 'standard' template library not being so standard".
2389    
2390    10. There was no check on the number of named subpatterns nor the maximum
2391        length of a subpattern name. The product of these values is used to compute
2392        the size of the memory block for a compiled pattern. By supplying a very
2393        long subpattern name and a large number of named subpatterns, the size
2394        computation could be caused to overflow. This is now prevented by limiting
2395        the length of names to 32 characters, and the number of named subpatterns
2396        to 10,000.
2397    
2398    11. Subpatterns that are repeated with specific counts have to be replicated in
2399        the compiled pattern. The size of memory for this was computed from the
2400        length of the subpattern and the repeat count. The latter is limited to
2401        65535, but there was no limit on the former, meaning that integer overflow
2402        could in principle occur. The compiled length of a repeated subpattern is
2403        now limited to 30,000 bytes in order to prevent this.
2404    
2405    12. Added the optional facility to have named substrings with the same name.
2406    
2407    13. Added the ability to use a named substring as a condition, using the
2408        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2409        are numbers (not recommended). Forward references are permitted.
2410    
2411    14. Added forward references in named backreferences (if you see what I mean).
2412    
2413    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2414        pattern could run off the end of the subject. For example, the pattern
2415        "(?s)(.{1,5})"8 did this with the subject "ab".
2416    
2417    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2418        PCRE_CASELESS was set when matching characters that were quantified with ?
2419        or *.
2420    
2421    17. A character class other than a single negated character that had a minimum
2422        but no maximum quantifier - for example [ab]{6,} - was not handled
2423        correctly by pce_dfa_exec(). It would match only one character.
2424    
2425    18. A valid (though odd) pattern that looked like a POSIX character
2426        class but used an invalid character after [ (for example [[,abc,]]) caused
2427        pcre_compile() to give the error "Failed: internal error: code overflow" or
2428        in some cases to crash with a glibc free() error. This could even happen if
2429        the pattern terminated after [[ but there just happened to be a sequence of
2430        letters, a binary zero, and a closing ] in the memory that followed.
2431    
2432    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2433        over the years. Originally (before any Unicode support), just the bottom 8
2434        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2435        output from "man perlunicode" includes this:
2436    
2437          The regular expression compiler produces polymorphic opcodes.  That
2438          is, the pattern adapts to the data and automatically switches to
2439          the Unicode character scheme when presented with Unicode data--or
2440          instead uses a traditional byte scheme when presented with byte
2441          data.
2442    
2443        Sadly, a wide octal escape does not cause a switch, and in a string with
2444        no other multibyte characters, these octal escapes are treated as before.
2445        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2446        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2447        Unicode string.
2448    
2449        I have not perpetrated such confusion in PCRE. Up till now, it took just
2450        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2451        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2452        translate to the appropriate multibyte character.
2453    
2454    29. Applied some refactoring to reduce the number of warnings from Microsoft
2455        and Borland compilers. This has included removing the fudge introduced
2456        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2457        a warning about an unused variable.
2458    
2459    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2460        characters since release 4.0, because Perl (from release 5.004) does not.
2461        [Or at least, is documented not to: some releases seem to be in conflict
2462        with the documentation.] However, when a pattern was studied with
2463        pcre_study() and all its branches started with \s, PCRE still included VT
2464        as a possible starting character. Of course, this did no harm; it just
2465        caused an unnecessary match attempt.
2466    
2467    22. Removed a now-redundant internal flag bit that recorded the fact that case
2468        dependency changed within the pattern. This was once needed for "required
2469        byte" processing, but is no longer used. This recovers a now-scarce options
2470        bit. Also moved the least significant internal flag bit to the most-
2471        significant bit of the word, which was not previously used (hangover from
2472        the days when it was an int rather than a uint) to free up another bit for
2473        the future.
2474    
2475    23. Added support for CRLF line endings as well as CR and LF. As well as the
2476        default being selectable at build time, it can now be changed at runtime
2477        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2478        specify that it is scanning data with non-default line endings.
2479    
2480    24. Changed the definition of CXXLINK to make it agree with the definition of
2481        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2482    
2483    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2484        recursions. This makes a big different to stack usage for some patterns.
2485    
2486    26. If a subpattern containing a named recursion or subroutine reference such
2487        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2488        the space required for the compiled pattern went wrong and gave too small a
2489        value. Depending on the environment, this could lead to "Failed: internal
2490        error: code overflow at offset 49" or "glibc detected double free or
2491        corruption" errors.
2492    
2493    27. Applied patches from Google (a) to support the new newline modes and (b) to
2494        advance over multibyte UTF-8 characters in GlobalReplace.
2495    
2496    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2497        difference for some implementation of PCRE in some Windows version.
2498    
2499    29. Added some extra testing facilities to pcretest:
2500    
2501        \q<number>   in a data line sets the "match limit" value
2502        \Q<number>   in a data line sets the "match recursion limt" value
2503        -S <number>  sets the stack size, where <number> is in megabytes
2504    
2505        The -S option isn't available for Windows.
2506    
2507    
2508    Version 6.6 06-Feb-06
2509    ---------------------
2510    
2511     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2512        in pcreposix.h. I have copied the definition from pcre.h.
2513    
2514     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2515        because pcre.h is no longer a built file.
2516    
2517     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2518        not normally included in the compiled code.
2519    
2520    
2521    Version 6.5 01-Feb-06
2522    ---------------------
2523    
2524     1. When using the partial match feature with pcre_dfa_exec(), it was not
2525        anchoring the second and subsequent partial matches at the new starting
2526        point. This could lead to incorrect results. For example, with the pattern
2527        /1234/, partially matching against "123" and then "a4" gave a match.
2528    
2529     2. Changes to pcregrep:
2530    
2531        (a) All non-match returns from pcre_exec() were being treated as failures
2532            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2533            error message is output. Some extra information is given for the
2534            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2535            probably the only errors that are likely to be caused by users (by
2536            specifying a regex that has nested indefinite repeats, for instance).
2537            If there are more than 20 of these errors, pcregrep is abandoned.
2538    
2539        (b) A binary zero was treated as data while matching, but terminated the
2540            output line if it was written out. This has been fixed: binary zeroes
2541            are now no different to any other data bytes.
2542    
2543        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2544            used to set a locale for matching. The --locale=xxxx long option has
2545            been added (no short equivalent) to specify a locale explicitly on the
2546            pcregrep command, overriding the environment variables.
2547    
2548        (d) When -B was used with -n, some line numbers in the output were one less
2549            than they should have been.
2550    
2551        (e) Added the -o (--only-matching) option.
2552    
2553        (f) If -A or -C was used with -c (count only), some lines of context were
2554            accidentally printed for the final match.
2555    
2556        (g) Added the -H (--with-filename) option.
2557    
2558        (h) The combination of options -rh failed to suppress file names for files
2559            that were found from directory arguments.
2560    
2561        (i) Added the -D (--devices) and -d (--directories) options.
2562    
2563        (j) Added the -F (--fixed-strings) option.
2564    
2565        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2566    
2567        (l) Added the --colo(u)r option.
2568    
2569        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2570            is not present by default.
2571    
2572     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2573        items such as (?R) or (?1), when the recursion could match a number of
2574        alternatives. If it matched one of the alternatives, but subsequently,
2575        outside the recursion, there was a failure, the code tried to back up into
2576        the recursion. However, because of the way PCRE is implemented, this is not
2577        possible, and the result was an incorrect result from the match.
2578    
2579        In order to prevent this happening, the specification of recursion has
2580        been changed so that all such subpatterns are automatically treated as
2581        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2582    
2583     4. I had overlooked the fact that, in some locales, there are characters for
2584        which isalpha() is true but neither isupper() nor islower() are true. In
2585        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2586        and ordfeminine) are like this. This affected the treatment of \w and \W
2587        when they appeared in character classes, but not when they appeared outside
2588        a character class. The bit map for "word" characters is now created
2589        separately from the results of isalnum() instead of just taking it from the
2590        upper, lower, and digit maps. (Plus the underscore character, of course.)
2591    
2592     5. The above bug also affected the handling of POSIX character classes such as
2593        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2594        permanent tables. Instead, the bit maps for such a class were previously
2595        created as the appropriate unions of the upper, lower, and digit bitmaps.
2596        Now they are created by subtraction from the [[:word:]] class, which has
2597        its own bitmap.
2598    
2599     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2600        It is created by subtracting the vertical space characters (\x09, \x0a,
2601        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2602        subtraction was done in the overall bitmap for a character class, meaning
2603        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2604        be recognized. This bug has been fixed.
2605    
2606     7. Patches from the folks at Google:
2607    
2608          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2609          real life, but is still worth protecting against".
2610    
2611          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2612          regular expressions".
2613    
2614          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2615          have it.
2616    
2617          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2618          "configure" and the latter not, in order to fix a problem somebody had
2619          with compiling the Arg class on HP-UX.
2620    
2621          (e) Improve the error-handling of the C++ wrapper a little bit.
2622    
2623          (f) New tests for checking recursion limiting.
2624    
2625     8. The pcre_memmove() function, which is used only if the environment does not
2626        have a standard memmove() function (and is therefore rarely compiled),
2627        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2628        returning a result (though PCRE never actually uses the result).
2629    
2630     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2631        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2632        returned instead of calling malloc() with an overflowing number that would
2633        most likely cause subsequent chaos.
2634    
2635    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2636    
2637    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2638        with this option is matched, the nmatch and pmatch options of regexec() are
2639        ignored.
2640    
2641    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2642        provided in case anyone wants to the the POSIX interface with UTF-8
2643        strings.
2644    
2645    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2646        C++ linking (needed for some HP-UX environments).
2647    
2648    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2649        (unused parameter) and in the pcre_printint() function (omitted "default"
2650        switch label when the default is to do nothing).
2651    
2652    15. Added some code to make it possible, when PCRE is compiled as a C++
2653        library, to replace subject pointers for pcre_exec() with a smart pointer
2654        class, thus making it possible to process discontinuous strings.
2655    
2656    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2657        much the same function. They were added by different people who were trying
2658        to make PCRE easy to compile on non-Unix systems. It has been suggested
2659        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2660        for compiling on Windows systems. I have therefore replaced it with
2661        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2662        defaults to "extern" for C or "extern C" for C++, which works fine on
2663        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2664        SCOPE with something explicit in config.h. In addition:
2665    
2666        (a) pcreposix.h still had just "extern" instead of either of these macros;
2667            I have replaced it with PCRE_DATA_SCOPE.
2668    
2669        (b) Functions such as _pcre_xclass(), which are internal to the library,
2670            but external in the C sense, all had PCRE_EXPORT in their definitions.
2671            This is apparently wrong for the Windows case, so I have removed it.
2672            (It makes no difference on Unix-like systems.)
2673    
2674    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2675        of recursive calls to match(). This is different to MATCH_LIMIT because
2676        that limits the total number of calls to match(), not all of which increase
2677        the depth of recursion. Limiting the recursion depth limits the amount of
2678        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2679        when PCRE is compiled, and changed at run time. A patch from Google adds
2680        this functionality to the C++ interface.
2681    
2682    18. Changes to the handling of Unicode character properties:
2683    
2684        (a) Updated the table to Unicode 4.1.0.
2685    
2686        (b) Recognize characters that are not in the table as "Cn" (undefined).
2687    
2688        (c) I revised the way the table is implemented to a much improved format
2689            which includes recognition of ranges. It now supports the ranges that
2690            are defined in UnicodeData.txt, and it also amalgamates other
2691            characters into ranges. This has reduced the number of entries in the
2692            table from around 16,000 to around 3,000, thus reducing its size
2693            considerably. I realized I did not need to use a tree structure after
2694            all - a binary chop search is just as efficient. Having reduced the
2695            number of entries, I extended their size from 6 bytes to 8 bytes to
2696            allow for more data.
2697    
2698        (d) Added support for Unicode script names via properties such as \p{Han}.
2699    
2700    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2701        matching that character.
2702    
2703    20. When matching a repeated Unicode property with a minimum greater than zero,
2704        (for example \pL{2,}), PCRE could look past the end of the subject if it
2705        reached it while seeking the minimum number of characters. This could
2706        happen only if some of the characters were more than one byte long, because
2707        there is a check for at least the minimum number of bytes.
2708    
2709    21. Refactored the implementation of \p and \P so as to be more general, to
2710        allow for more different types of property in future. This has changed the
2711        compiled form incompatibly. Anybody with saved compiled patterns that use
2712        \p or \P will have to recompile them.
2713    
2714    22. Added "Any" and "L&" to the supported property types.
2715    
2716    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2717        but give a compile time error if the value is greater than 0xff.
2718    
2719    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2720        accidentally not being installed or uninstalled.
2721    
2722    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2723        made were to insert the current release number. This seemed silly, because
2724        it made things harder for people building PCRE on systems that don't run
2725        "configure". I have turned pcre.h into a distributed file, no longer built
2726        by "configure", with the version identification directly included. There is
2727        no longer a pcre.h.in file.
2728    
2729        However, this change necessitated a change to the pcre-config script as
2730        well. It is built from pcre-config.in, and one of the substitutions was the
2731        release number. I have updated configure.ac so that ./configure now finds
2732        the release number by grepping pcre.h.
2733    
2734    26. Added the ability to run the tests under valgrind.
2735    
2736    
2737    Version 6.4 05-Sep-05
2738    ---------------------
2739    
2740     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2741        "--" to be printed when multiple files were scanned, even when none of the
2742        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2743        consider it to be a bug, and have restored the previous behaviour.
2744    
2745     2. A couple of code tidies to get rid of compiler warnings.
2746    
2747     3. The pcretest program used to cheat by referring to symbols in the library
2748        whose names begin with _pcre_. These are internal symbols that are not
2749        really supposed to be visible externally, and in some environments it is
2750        possible to suppress them. The cheating is now confined to including
2751        certain files from the library's source, which is a bit cleaner.
2752    
2753     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2754        file's purpose clearer.
2755    
2756     5. Reorganized pcre_ucp_findchar().
2757    
2758    
2759    Version 6.3 15-Aug-05
2760    ---------------------
2761    
2762     1. The file libpcre.pc.in did not have general read permission in the tarball.
2763    
2764     2. There were some problems when building without C++ support:
2765    
2766        (a) If C++ support was not built, "make install" and "make test" still
2767            tried to test it.
2768    
2769        (b) There were problems when the value of CXX was explicitly set. Some
2770            changes have been made to try to fix these, and ...
2771    
2772        (c) --disable-cpp can now be used to explicitly disable C++ support.
2773    
2774        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2775            backslash in a target when C++ was disabled. This confuses some
2776            versions of "make", apparently. Using an intermediate variable solves
2777            this. (Same for CPP_LOBJ.)
2778    
2779     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2780        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2781        necessary on certain architectures.
2782    
2783     4. Added a setting of -export-symbols-regex to the link command to remove
2784        those symbols that are exported in the C sense, but actually are local
2785        within the library, and not documented. Their names all begin with
2786        "_pcre_". This is not a perfect job, because (a) we have to except some
2787        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2788        available (and never for static libraries). I have made a note to try to
2789        find a way round (a) in the future.
2790    
2791    
2792    Version 6.2 01-Aug-05
2793    ---------------------
2794    
2795     1. There was no test for integer overflow of quantifier values. A construction
2796        such as {1111111111111111} would give undefined results. What is worse, if
2797        a minimum quantifier for a parenthesized subpattern overflowed and became
2798        negative, the calculation of the memory size went wrong. This could have
2799        led to memory overwriting.
2800    
2801     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2802    
2803     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2804        operating environments where this matters.
2805    
2806     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2807        PCRE options from within the C++ wrapper.
2808    
2809     5. Named capturing subpatterns were not being correctly counted when a pattern
2810        was compiled. This caused two problems: (a) If there were more than 100
2811        such subpatterns, the calculation of the memory needed for the whole
2812        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2813        back references of the form \12, where the number was greater than 9, were
2814        not recognized as back references, even though there were sufficient
2815        previous subpatterns.
2816    
2817     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2818        versions of gcc, e.g. 2.95.4.
2819    
2820    
2821    Version 6.1 21-Jun-05
2822    ---------------------
2823    
2824     1. There was one reference to the variable "posix" in pcretest.c that was not
2825        surrounded by "#if !defined NOPOSIX".
2826    
2827     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2828        the cross-check on the old pcre_info() function, for the benefit of the
2829        cut-down version of PCRE that is currently imported into Exim.
2830    
2831     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2832        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2833        patterns that start (?i) but I don't think that matters. The use of (?i) is
2834        just an example; this all applies to the other options as well.
2835    
2836     4. Since libtool seems to echo the compile commands it is issuing, the output
2837        from "make" can be reduced a bit by putting "@" in front of each libtool
2838        compile command.
2839    
2840     5. Patch from the folks at Google for configure.in to be a bit more thorough
2841        in checking for a suitable C++ installation before trying to compile the
2842        C++ stuff. This should fix a reported problem when a compiler was present,
2843        but no suitable headers.
2844    
2845     6. The man pages all had just "PCRE" as their title. I have changed them to
2846        be the relevant file name. I have also arranged that these names are
2847        retained in the file doc/pcre.txt, which is a concatenation in text format
2848        of all the man pages except the little individual ones for each function.
2849    
2850     7. The NON-UNIX-USE file had not been updated for the different set of source
2851        files that come with release 6. I also added a few comments about the C++
2852        wrapper.
2853    
2854    
2855    Version 6.0 07-Jun-05
2856    ---------------------
2857    
2858     1. Some minor internal re-organization to help with my DFA experiments.
2859    
2860     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2861        didn't matter for the library itself when fully configured, but did matter
2862        when compiling without UCP support, or within Exim, where the ucp files are
2863        not imported.
2864    
2865     3. Refactoring of the library code to split up the various functions into
2866        different source modules. The addition of the new DFA matching code (see
2867        below) to a single monolithic source would have made it really too
2868        unwieldy, quite apart from causing all the code to be include in a
2869        statically linked application, when only some functions are used. This is
2870        relevant even without the DFA addition now that patterns can be compiled in
2871        one application and matched in another.
2872    
2873        The downside of splitting up is that there have to be some external
2874        functions and data tables that are used internally in different modules of
2875        the library but which are not part of the API. These have all had their
2876        names changed to start with "_pcre_" so that they are unlikely to clash
2877        with other external names.
2878    
2879     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2880        a different (DFA) algorithm. Although it is slower than the original
2881        function, it does have some advantages for certain types of matching
2882        problem.
2883    
2884     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2885        including restarting after a partial match.
2886    
2887     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2888        defined when compiling for Windows was sent to me. I have put it into the
2889        code, though I have no means of testing or verifying it.
2890    
2891     7. Added the pcre_refcount() auxiliary function.
2892    
2893     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2894        match before or at the first newline in the subject string. In pcretest,
2895        the /f option on a pattern can be used to set this.
2896    
2897     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2898        would behave wrongly. This has been present in PCRE since release 4.0.
2899    
2900    10. A number of changes to the pcregrep command:
2901    
2902        (a) Refactored how -x works; insert ^(...)$ instead of setting
2903            PCRE_ANCHORED and checking the length, in preparation for adding
2904            something similar for -w.
2905    
2906        (b) Added the -w (match as a word) option.
2907    
2908        (c) Refactored the way lines are read and buffered so as to have more
2909            than one at a time available.
2910    
2911        (d) Implemented a pcregrep test script.
2912    
2913        (e) Added the -M (multiline match) option. This allows patterns to match
2914            over several lines of the subject. The buffering ensures that at least
2915            8K, or the rest of the document (whichever is the shorter) is available
2916            for matching (and similarly the previous 8K for lookbehind assertions).
2917    
2918        (f) Changed the --help output so that it now says
2919    
2920              -w, --word-regex(p)
2921    
2922            instead of two lines, one with "regex" and the other with "regexp"
2923            because that confused at least one person since the short forms are the
2924            same. (This required a bit of code, as the output is generated
2925            automatically from a table. It wasn't just a text change.)
2926    
2927        (g) -- can be used to terminate pcregrep options if the next thing isn't an
2928            option but starts with a hyphen. Could be a pattern or a path name
2929            starting with a hyphen, for instance.
2930    
2931        (h) "-" can be given as a file name to represent stdin.
2932    
2933        (i) When file names are being printed, "(standard input)" is used for
2934            the standard input, for compatibility with GNU grep. Previously
2935            "<stdin>" was used.
2936    
2937        (j) The option --label=xxx can be used to supply a name to be used for
2938            stdin when file names are being printed. There is no short form.
2939    
2940        (k) Re-factored the options decoding logic because we are going to add
2941            two more options that take data. Such options can now be given in four
2942            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2943    
2944        (l) Added the -A, -B, and -C options for requesting that lines of context
2945            around matches be printed.
2946    
2947        (m) Added the -L option to print the names of files that do not contain
2948            any matching lines, that is, the complement of -l.
2949    
2950        (n) The return code is 2 if any file cannot be opened, but pcregrep does
2951            continue to scan other files.
2952    
2953        (o) The -s option was incorrectly implemented. For compatibility with other
2954            greps, it now suppresses the error message for a non-existent or non-
2955            accessible file (but not the return code). There is a new option called
2956            -q that suppresses the output of matching lines, which was what -s was
2957            previously doing.
2958    
2959        (p) Added --include and --exclude options to specify files for inclusion
2960            and exclusion when recursing.
2961    
2962    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2963        Hopefully, it now does.
2964    
2965    12. Missing cast in pcre_study().
2966    
2967    13. Added an "uninstall" target to the makefile.
2968    
2969    14. Replaced "extern" in the function prototypes in Makefile.in with
2970        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2971        world, but is set differently for Windows.
2972    
2973    15. Added a second compiling function called pcre_compile2(). The only
2974        difference is that it has an extra argument, which is a pointer to an
2975        integer error code. When there is a compile-time failure, this is set
2976        non-zero, in addition to the error test pointer being set to point to an
2977        error message. The new argument may be NULL if no error number is required
2978        (but then you may as well call pcre_compile(), which is now just a
2979        wrapper). This facility is provided because some applications need a
2980        numeric error indication, but it has also enabled me to tidy up the way
2981        compile-time errors are handled in the POSIX wrapper.
2982    
2983    16. Added VPATH=.libs to the makefile; this should help when building with one
2984        prefix path and installing with another. (Or so I'm told by someone who
2985        knows more about this stuff than I do.)
2986    
2987    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2988        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2989        match everything, including newlines. This is not POSIX-compatible, but
2990        somebody wanted the feature. From pcretest it can be activated by using
2991        both the P and the s flags.
2992    
2993    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2994    
2995    19. libpcre.pc was being incorrectly installed as executable.
2996    
2997    20. A couple of places in pcretest check for end-of-line by looking for '\n';
2998        it now also looks for '\r' so that it will work unmodified on Windows.
2999    
3000    21. Added Google's contributed C++ wrapper to the distribution.
3001    
3002    22. Added some untidy missing memory free() calls in pcretest, to keep
3003        Electric Fence happy when testing.
3004    
3005    
3006    
3007    Version 5.0 13-Sep-04
3008    ---------------------
3009    
3010     1. Internal change: literal characters are no longer packed up into items
3011        containing multiple characters in a single byte-string. Each character
3012        is now matched using a separate opcode. However, there may be more than one
3013        byte in the character in UTF-8 mode.
3014    
3015     2. The pcre_callout_block structure has two new fields: pattern_position and
3016        next_item_length. These contain the offset in the pattern to the next match
3017        item, and its length, respectively.
3018    
3019     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3020        insertion of callouts before each pattern item. Added the /C option to
3021        pcretest to make use of this.
3022    
3023     4. On the advice of a Windows user, the lines
3024    
3025          #if defined(_WIN32) || defined(WIN32)
3026          _setmode( _fileno( stdout ), 0x8000 );
3027          #endif  /* defined(_WIN32) || defined(WIN32) */
3028    
3029        have been added to the source of pcretest. This apparently does useful
3030        magic in relation to line terminators.
3031    
3032     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3033        for the benefit of those environments where the "b" makes a difference.
3034    
3035     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3036        to know about it. I have put a hack into configure.in that adds in code
3037        to set GCC=yes if CC=icc. This seems to end up at a point in the
3038        generated configure script that is early enough to affect the setting of
3039        compiler options, which is what is needed, but I have no means of testing
3040        whether it really works. (The user who reported this had patched the
3041        generated configure script, which of course I cannot do.)
3042    
3043        LATER: After change 22 below (new libtool files), the configure script
3044        seems to know about icc (and also ecc). Therefore, I have commented out
3045        this hack in configure.in.
3046    
3047     7. Added support for pkg-config (2 patches were sent in).
3048    
3049     8. Negated POSIX character classes that used a combination of internal tables
3050        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3051        [[:^ascii]]. Typically, they would match almost any characters. The other
3052        POSIX classes were not broken in this way.
3053    
3054     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3055        to find the match, as PCRE was deluded into thinking that the match had to
3056        start at the start point or following a newline. The same bug applied to
3057        patterns with negative forward assertions or any backward assertions
3058        preceding ".*" at the start, unless the pattern required a fixed first
3059        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3060    
3061    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3062        starting at the last subject character, bytes beyond the end of the subject
3063        string were read.
3064    
3065    11. Renamed the variable "class" as "classbits" to make life easier for C++
3066        users. (Previously there was a macro definition, but it apparently wasn't
3067        enough.)
3068    
3069    12. Added the new field "tables" to the extra data so that tables can be passed
3070        in at exec time, or the internal tables can be re-selected. This allows
3071        a compiled regex to be saved and re-used at a later time by a different
3072        program that might have everything at different addresses.
3073    
3074    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3075        -R library as well as a -L library.
3076    
3077    14. The debugging options of pcretest (-d on the command line or D on a
3078        pattern) showed incorrect output for anything following an extended class
3079        that contained multibyte characters and which was followed by a quantifier.
3080    
3081    15. Added optional support for general category Unicode character properties
3082        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3083        support. It adds about 90K to the size of the library. The meanings of the
3084        inbuilt class escapes such as \d and \s have NOT been changed.
3085    
3086    16. Updated pcredemo.c to include calls to free() to release the memory for the
3087        compiled pattern.
3088    
3089    17. The generated file chartables.c was being created in the source directory
3090        instead of in the building directory. This caused the build to fail if the
3091        source directory was different from the building directory, and was
3092        read-only.
3093    
3094    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3095        file. No doubt somebody will tell me if they don't make sense... Also added
3096        Dan Mooney's comments about building on OpenVMS.
3097    
3098    19. Added support for partial matching via the PCRE_PARTIAL option for
3099        pcre_exec() and the \P data escape in pcretest.
3100    
3101    20. Extended pcretest with 3 new pattern features:
3102    
3103        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3104              write the compiled pattern to the file whose name is "rest-of-line".
3105              This is a straight binary dump of the data, with the saved pointer to
3106              the character tables forced to be NULL. The study data, if any, is
3107              written too. After writing, pcretest reads a new pattern.
3108    
3109        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3110              compiled pattern from the given file. There must not be any
3111              occurrences of "<" in the file name (pretty unlikely); if there are,
3112              pcretest will instead treat the initial "<" as a pattern delimiter.
3113              After reading in the pattern, pcretest goes on to read data lines as
3114              usual.
3115    
3116        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3117              and 16-bit fields in a compiled pattern, to simulate a pattern that
3118              was compiled on a host of opposite endianness.
3119    
3120    21. The pcre-exec() function can now cope with patterns that were compiled on
3121        hosts of opposite endianness, with this restriction:
3122    
3123          As for any compiled expression that is saved and used later, the tables
3124          pointer field cannot be preserved; the extra_data field in the arguments
3125          to pcre_exec() should be used to pass in a tables address if a value
3126          other than the default internal tables were used at compile time.
3127    
3128    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3129        now diagnosed as an error. Previously, most of the time, a negative number
3130        would have been treated as zero, but if in addition "ovector" was passed as
3131        NULL, a crash could occur.
3132    
3133    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3134        new versions from the libtool 1.5 distribution (the last one is a copy of
3135        a file called libtool.m4). This seems to have fixed the need to patch
3136        "configure" to support Darwin 1.3 (which I used to do). However, I still
3137        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3138        workstation).
3139    
3140    24. Changed the PCRE licence to be the more standard "BSD" licence.
3141    
3142    
3143    Version 4.5 01-Dec-03
3144    ---------------------
3145    
3146     1. There has been some re-arrangement of the code for the match() function so
3147        that it can be compiled in a version that does not call itself recursively.
3148        Instead, it keeps those local variables that need separate instances for
3149        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3150        needs to "recurse". Keeping track of where control must go is done by means
3151        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3152        hide most of the details from the main code, and operates only if
3153        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3154        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3155        operating.
3156    
3157        To make it easier for callers to provide specially tailored get/free
3158        functions for this usage, two new functions, pcre_stack_malloc, and
3159        pcre_stack_free, are used. They are always called in strict stacking order,
3160        and the size of block requested is always the same.
3161    
3162        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3163        PCRE has been compiled to use the stack or the heap for recursion. The
3164        -C option of pcretest uses this to show which version is compiled.
3165    
3166        A new data escape \S, is added to pcretest; it causes the amounts of store
3167        obtained and freed by both kinds of malloc/free at match time to be added
3168        to the output.
3169    
3170     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3171        what's available on my current Linux desktop machine.
3172    
3173     3. When matching a UTF-8 string, the test for a valid string at the start has
3174        been extended. If start_offset is not zero, PCRE now checks that it points
3175        to a byte that is the start of a UTF-8 character. If not, it returns
3176        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3177        this is necessary because there may be backward assertions in the pattern.
3178        When matching the same subject several times, it may save resources to use
3179        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3180    
3181     4. The code for checking the validity of UTF-8 strings has been tightened so
3182        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3183        containing "overlong sequences".
3184    
3185     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3186        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3187        should have been "&", but it just so happened that all the cases this let
3188        through by mistake were picked up later in the function.
3189    
3190     6. I had used a variable called "isblank" - this is a C99 function, causing
3191        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3192    
3193     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3194        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3195        so the version is shown; (c) stop "make test" echoing "./RunTest".
3196    
3197     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3198    
3199     9. The prototype for memmove() for systems that don't have it was using
3200        size_t, but the inclusion of the header that defines size_t was later. I've
3201        moved the #includes for the C headers earlier to avoid this.
3202    
3203    10. Added some adjustments to the code to make it easier to compiler on certain
3204        special systems:
3205    
3206          (a) Some "const" qualifiers were missing.
3207          (b) Added the macro EXPORT before all exported functions; by default this
3208              is defined to be empty.
3209          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3210              that it reads its output file name as an argument instead of writing
3211              to the standard output and assuming this can be redirected.
3212    
3213    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3214        class containing characters with values greater than 255, PCRE compilation
3215        went into a loop.
3216    
3217    12. A recursive reference to a subpattern that was within another subpattern
3218        that had a minimum quantifier of zero caused PCRE to crash. For example,
3219        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3220        recursion. If the recursively-called subpattern itself had a zero repeat,
3221        that was OK.
3222    
3223    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3224        buffer into which it was copied (for escape processing) was still set at
3225        1024, so long lines caused crashes.
3226    
3227    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3228        "internal error: code overflow...". This applied to any character class
3229        that was followed by a possessive quantifier.
3230    
3231    15. Modified the Makefile to add libpcre.la as a prerequisite for
3232        libpcreposix.la because I was told this is needed for a parallel build to
3233        work.
3234    
3235    16. If a pattern that contained .* following optional items at the start was
3236        studied, the wrong optimizing data was generated, leading to matching
3237        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3238        matching string must start with a or b or c. The correct conclusion for
3239        this pattern is that a match can start with any character.
3240    
3241    
3242    Version 4.4 13-Aug-03
3243    ---------------------
3244    
3245     1. In UTF-8 mode, a character class containing characters with values between
3246        127 and 255 was not handled correctly if the compiled pattern was studied.
3247        In fixing this, I have also improved the studying algorithm for such
3248        classes (slightly).
3249    
3250     2. Three internal functions had redundant arguments passed to them. Removal
3251        might give a very teeny performance improvement.
3252    
3253     3. Documentation bug: the value of the capture_top field in a callout is *one
3254        more than* the number of the hightest numbered captured substring.
3255    
3256     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3257        in incorrectly linking with a previously installed version. They now link
3258        explicitly with libpcre.la.
3259    
3260     5. configure.in no longer needs to recognize Cygwin specially.
3261    
3262     6. A problem in pcre.in for Windows platforms is fixed.
3263    
3264     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3265        pcretest, it used to include the size of the study block as part of its
3266        output. Unfortunately, the structure contains a field that has a different
3267        size on different hardware architectures. This meant that the tests that
3268        showed this size failed. As the block is currently always of a fixed size,
3269        this information isn't actually particularly useful in pcretest output, so
3270        I have just removed it.
3271    
3272     8. Three pre-processor statements accidentally did not start in column 1.
3273        Sadly, there are *still* compilers around that complain, even though
3274        standard C has not required this for well over a decade. Sigh.
3275    
3276     9. In pcretest, the code for checking callouts passed small integers in the
3277        callout_data field, which is a void * field. However, some picky compilers
3278        complained about the casts involved for this on 64-bit systems. Now
3279        pcretest passes the address of the small integer instead, which should get
3280        rid of the warnings.
3281    
3282    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3283        both compile and run time, and gives an error if an invalid UTF-8 sequence
3284        is found. There is a option for disabling this check in cases where the
3285        string is known to be correct and/or the maximum performance is wanted.
3286    
3287    11. In response to a bug report, I changed one line in Makefile.in from
3288    
3289            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3290        to
3291            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3292    
3293        to look similar to other lines, but I have no way of telling whether this
3294        is the right thing to do, as I do not use Windows. No doubt I'll get told
3295        if it's wrong...
3296    
3297    
3298    Version 4.3 21-May-03
3299    ---------------------
3300    
3301    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3302       Makefile.
3303    
3304    2. Some refactoring to improve the quality of the code:
3305    
3306       (i)   The utf8_table... variables are now declared "const".
3307    
3308       (ii)  The code for \cx, which used the "case flipping" table to upper case
3309             lower case letters, now just substracts 32. This is ASCII-specific,
3310             but the whole concept of \cx is ASCII-specific, so it seems
3311             reasonable.
3312    
3313       (iii) PCRE was using its character types table to recognize decimal and
3314             hexadecimal digits in the pattern. This is silly, because it handles
3315             only 0-9, a-f, and A-F, but the character types table is locale-
3316             specific, which means strange things might happen. A private
3317             table is now used for this - though it costs 256 bytes, a table is
3318             much faster than multiple explicit tests. Of course, the standard
3319             character types table is still used for matching digits in subject
3320             strings against \d.
3321    
3322       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3323             ending in _t are). So I've renamed it as ESC_tee.
3324    
3325    3. The first argument for regexec() in the POSIX wrapper should have been
3326       defined as "const".
3327    
3328    4. Changed pcretest to use malloc() for its buffers so that they can be
3329       Electric Fenced for debugging.
3330    
3331    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3332       to read one or more bytes before the start of the subject string. Often this
3333       had no effect on PCRE's behaviour, but in some circumstances it could
3334       provoke a segmentation fault.
3335    
3336    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3337       to try to read one or more bytes before the start of the subject string.
3338    
3339    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3340       UTF-8 support could misbehave in various ways if the subject string
3341       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3342       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3343       back over UTF-8 characters.)
3344    
3345    
3346    Version 4.2 14-Apr-03
3347    ---------------------
3348    
3349    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3350    
3351    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3352         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3353         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3354         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3355         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3356           and BUILD_EXEEXT
3357         Note: automatic setting of the BUILD variables is not yet working
3358         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3359           compile-time but not at link-time
3360         [LINK]: use for linking executables only
3361         make different versions for Windows and non-Windows
3362         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3363           libraries
3364         [LINK_FOR_BUILD]: new variable
3365         [OBJEXT]: use throughout
3366         [EXEEXT]: use throughout
3367         <winshared>: new target
3368         <wininstall>: new target
3369         <dftables.o>: use native compiler
3370         <dftables>: use native linker
3371         <install>: handle Windows platform correctly
3372         <clean>: ditto
3373         <check>: ditto
3374         copy DLL to top builddir before testing
3375    
3376       As part of these changes, -no-undefined was removed again. This was reported
3377       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3378       in any case.
3379    
3380    3. Some tidies to get rid of compiler warnings:
3381    
3382       . In the match_data structure, match_limit was an unsigned long int, whereas
3383         match_call_count was an int. I've made them both unsigned long ints.
3384    
3385       . In pcretest the fact that a const uschar * doesn't automatically cast to
3386         a void * provoked a warning.
3387    
3388       . Turning on some more compiler warnings threw up some "shadow" variables
3389         and a few more missing casts.
3390    
3391    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3392       option, a class that contained a single character with a value between 128
3393       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3394    
3395    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3396       option, a class that contained several characters, but with at least one
3397       whose value was between 128 and 255 caused PCRE to crash.
3398    
3399    
3400    Version 4.1 12-Mar-03
3401    ---------------------
3402    
3403    1. Compiling with gcc -pedantic found a couple of places where casts were
3404    needed, and a string in dftables.c that was longer than standard compilers are
3405    required to support.
3406    
3407    2. Compiling with Sun's compiler found a few more places where the code could
3408    be tidied up in order to avoid warnings.
3409    
3410    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3411    first of these names is deprecated in the latest Autoconf in favour of the name
3412    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3413    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3414    analogy I have changed it to CFLAGS_FOR_BUILD.
3415    
3416    4. Added -no-undefined to the linking command in the Makefile, because this is
3417    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3418    linking step for the pcreposix library.
3419    
3420    5. PCRE was failing to diagnose the case of two named groups with the same
3421    name.
3422    
3423    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3424    literal character that is needed in the subject for a match, and scans along to
3425    ensure that it is present before embarking on the full matching process. This
3426    saves time in cases of nested unlimited repeats that are never going to match.
3427    Problem: the scan can take a lot of time if the subject is very long (e.g.
3428    megabytes), thus penalizing straightforward matches. It is now done only if the
3429    amount of subject to be scanned is less than 1000 bytes.
3430    
3431    7. A lesser problem with the same optimization is that it was recording the
3432    first character of an anchored pattern as "needed", thus provoking a search
3433    right along the subject, even when the first match of the pattern was going to
3434    fail. The "needed" character is now not set for anchored patterns, unless it
3435    follows something in the pattern that is of non-fixed length. Thus, it still
3436    fulfils its original purpose of finding quick non-matches in cases of nested
3437    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3438    
3439    
3440    Version 4.0 17-Feb-03
3441    ---------------------
3442    
3443    1. If a comment in an extended regex that started immediately after a meta-item
3444    extended to the end of string, PCRE compiled incorrect data. This could lead to
3445    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3446    
3447    2. Moved to autoconf 2.53 and libtool 1.4.2.
3448    
3449    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3450    the special perltest8 script is no longer needed - all the tests can be run
3451    from a single perltest script.
3452    
3453    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3454    by \s. It has now been removed in PCRE. This means it isn't recognized as
3455    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3456    class [:space:] *does* include VT, thereby creating a mess.
3457    
3458    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3459    space and tab.
3460    
3461    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3462    its new features into the main test script, reducing the number of scripts.
3463    
3464    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3465    were backward compatible, and made the (?i) apply to the whole pattern, as if
3466    /i were given. Now it behaves more logically, and applies the option setting
3467    only to what follows. PCRE has been changed to follow suit. However, if it
3468    finds options settings right at the start of the pattern, it extracts them into
3469    the global options, as before. Thus, they show up in the info data.
3470    
3471    8. Added support for the \Q...\E escape sequence. Characters in between are
3472    treated as literals. This is slightly different from Perl in that $ and @ are
3473    also handled as literals inside the quotes. In Perl, they will cause variable
3474    interpolation. Note the following examples:
3475    
3476        Pattern            PCRE matches      Perl matches
3477    
3478        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3479        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3480        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3481    
3482    For compatibility with Perl, \Q...\E sequences are recognized inside character
3483    classes as well as outside them.
3484    
3485    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3486    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3487    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3488    signed/unsigned warnings.
3489    
3490    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3491    option for pcretest, so I've replaced it by a simple function that does just
3492    that job.
3493    
3494    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3495    "pcregrep -".
3496    
3497    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3498    Java package. This provides some syntactic sugar for simple cases of what my
3499    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3500    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3501    item, you can use this simplified notation. Note that only makes sense with
3502    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3503    greediness, whatever the setting of the PCRE_UNGREEDY option.
3504    
3505    13. A change of greediness default within a pattern was not taking effect at
3506    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3507    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3508    was abstracted outside.
3509    
3510    14. PCRE now supports the \G assertion. It is true when the current matching
3511    position is at the start point of the match. This differs from \A when the
3512    starting offset is non-zero. Used with the /g option of pcretest (or similar
3513    code), it works in the same way as it does for Perl's /g option. If all
3514    alternatives of a regex begin with \G, the expression is anchored to the start
3515    match position, and the "anchored" flag is set in the compiled expression.
3516    
3517    15. Some bugs concerning the handling of certain option changes within patterns
3518    have been fixed. These applied to options other than (?ims). For example,
3519    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3520    been the other way round. Some of this was related to change 7 above.
3521    
3522    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3523    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3524    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3525    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3526    
3527    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3528    mode. Unlike ".", it always matches newline, whatever the setting of
3529    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3530    assertions. Perl allows it, but it doesn't (in general) work because it can't
3531    calculate the length of the lookbehind. At least, that's the case for Perl
3532    5.8.0 - I've been told they are going to document that it doesn't work in
3533    future.
3534    
3535    18. Added an error diagnosis for escapes that PCRE does not support: these are
3536    \L, \l, \N, \P, \p, \U, \u, and \X.
3537    
3538    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3539    reading past the end of the pattern in cases such as /[abcd/.
3540    
3541    20. PCRE was getting more memory than necessary for patterns with classes that
3542    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3543    
3544    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3545    compiling PCRE for use with Virtual Pascal.
3546    
3547    22. Small fix to the Makefile to make it work properly if the build is done
3548    outside the source tree.
3549    
3550    23. Added a new extension: a condition to go with recursion. If a conditional
3551    subpattern starts with (?(R) the "true" branch is used if recursion has
3552    happened, whereas the "false" branch is used only at the top level.
3553    
3554    24. When there was a very long string of literal characters (over 255 bytes
3555    without UTF support, over 250 bytes with UTF support), the computation of how
3556    much memory was required could be incorrect, leading to segfaults or other
3557    strange effects.
3558    
3559    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3560    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3561    there was a subsequent back reference to those brackets. This meant that, for
3562    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3563    possible to check for precisely this case. All we can do is abandon the
3564    optimization if .* occurs inside capturing brackets when there are any back
3565    references whatsoever. (See below for a better fix that came later.)
3566    
3567    26. The handling of the optimization for finding the first character of a
3568    non-anchored pattern, and for finding a character that is required later in the
3569    match were failing in some cases. This didn't break the matching; it just
3570    failed to optimize when it could. The way this is done has been re-implemented.
3571    
3572    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3573    
3574    28. Added a new feature that provides some of the functionality that Perl
3575    provides with (?{...}). The facility is termed a "callout". The way it is done
3576    in PCRE is for the caller to provide an optional function, by setting
3577    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3578    global variable. By default it is unset, which disables all calling out. To get
3579    the function called, the regex must include (?C) at appropriate points. This
3580    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3581    This provides a means of identifying different callout points. When PCRE
3582    reaches such a point in the regex, if pcre_callout has been set, the external
3583    function is called. It is provided with data in a structure called
3584    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3585    matching continues; if it returns a non-zero value, the match at the current
3586    point fails. However, backtracking will occur if possible. [This was changed
3587    later and other features added - see item 49 below.]
3588    
3589    29. pcretest is upgraded to test the callout functionality. It provides a
3590    callout function that displays information. By default, it shows the start of
3591    the match and the current position in the text. There are some new data escapes
3592    to vary what happens:
3593    
3594        \C+         in addition, show current contents of captured substrings
3595        \C-         do not supply a callout function
3596        \C!n        return 1 when callout number n is reached
3597        \C!n!m      return 1 when callout number n is reached for the mth time