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

Diff of /code/trunk/ChangeLog

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

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

Legend:
Removed from v.81  
changed lines
  Added in v.915

  ViewVC Help
Powered by ViewVC 1.1.5