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

Diff of /code/trunk/ChangeLog

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

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