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

Diff of /code/trunk/ChangeLog

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

revision 41 by nigel, Sat Feb 24 21:39:17 2007 UTC revision 922 by ph10, Mon Feb 20 18:44:42 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7  Version 2.09 14-Sep-99  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.  Fixed several bugs concerned with partial matching of items that consist
18        of more than one character:
19    
20        (a) /^(..)\1/ did not partially match "aba" because checking references was
21            done on an "all or nothing" basis. This also applied to repeated
22            references.
23    
24        (b) \R did not give a hard partial match if \r was found at the end of the
25            subject.
26    
27        (c) \X did not give a hard partial match after matching one or more
28            characters at the end of the subject.
29    
30        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
31            a partial match for the string "\r".
32    
33        (e) When newline was set to CRLF, the metacharacter "." did not recognize
34            a partial match for a CR character at the end of the subject string.
35    
36    6.  When matching is performed using JIT-compiled code, the flag bit
37        PCRE_EXTRA_USED_JIT is set in the pcre_extra block and can be tested by the
38        caller. This bit is checked by pcretest if JIT is requested by /S++ or -s++
39        (instead of just /S+ or -s+) and the text "(JIT)" added to the output if
40        the bit is set.
41    
42    
43    Version 8.30 04-February-2012
44    -----------------------------
45    
46    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
47        name is defined in ctype.h.
48    
49    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
50        only in quite long subpatterns.
51    
52    3.  Removed the function pcre_info(), which has been obsolete and deprecated
53        since it was replaced by pcre_fullinfo() in February 2000.
54    
55    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
56        match a (*MARK), and the match failed at the start of the subject, a
57        reference to memory before the start of the subject could occur. This bug
58        was introduced by fix 17 of release 8.21.
59    
60    5.  A reference to an unset group with zero minimum repetition was giving
61        totally wrong answers (in non-JavaScript-compatibility mode). For example,
62        /(another)?(\1?)test/ matched against "hello world test". This bug was
63        introduced in release 8.13.
64    
65    6.  Add support for 16-bit character strings (a large amount of work involving
66        many changes and refactorings).
67    
68    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
69        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
70        from a file.
71    
72    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
73        rounding is not applied in this particular case).
74    
75    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
76        if they appear, or are escaped, in patterns.
77    
78    10. Get rid of a number of -Wunused-but-set-variable warnings.
79    
80    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
81        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
82        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
83        also returns the mark "x". This bug applied to capturing parentheses,
84        non-capturing parentheses, and atomic parentheses. It also applied to some
85        assertions.
86    
87    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
88        information out of configure.ac instead of relying on pcre.h.generic, which
89        is not stored in the repository.
90    
91    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
92        -lreadline.
93    
94    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
95    
96    15. Applied Graycode's patch to put the top-level frame on the stack rather
97        than the heap when not using the stack for recursion. This gives a
98        performance improvement in many cases when recursion is not deep.
99    
100    16. Experimental code added to "pcretest -C" to output the stack frame size.
101    
102    
103    Version 8.21 12-Dec-2011
104    ------------------------
105    
106    1.  Updating the JIT compiler.
107    
108    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
109        are added as well.
110    
111    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
112        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
113        calling _pcre_jit_exec. Some extra comments are added.
114    
115    4.  (*MARK) settings inside atomic groups that do not contain any capturing
116        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
117        was introduced by change 18 for 8.20.
118    
119    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
120        ECMA-262 standard.
121    
122    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
123        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
124        This bug was probably introduced by change 9 of 8.13.
125    
126    7.  While fixing 6 above, I noticed that a number of other items were being
127        incorrectly rejected as "not fixed length". This arose partly because newer
128        opcodes had not been added to the fixed-length checking code. I have (a)
129        corrected the bug and added tests for these items, and (b) arranged for an
130        error to occur if an unknown opcode is encountered while checking for fixed
131        length instead of just assuming "not fixed length". The items that were
132        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
133        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
134        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
135    
136    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
137        being incorrectly compiled and would have given unpredicatble results.
138    
139    9.  A possessively repeated subpattern with minimum repeat count greater than
140        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
141        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
142        the first (A) could occur when it should not.
143    
144    10. Add a cast and remove a redundant test from the code.
145    
146    11. JIT should use pcre_malloc/pcre_free for allocation.
147    
148    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
149        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
150        is anything other than /usr, -L is still shown).
151    
152    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
153    
154    14. Perl does not support \N without a following name in a [] class; PCRE now
155        also gives an error.
156    
157    15. If a forward reference was repeated with an upper limit of around 2000,
158        it caused the error "internal error: overran compiling workspace". The
159        maximum number of forward references (including repeats) was limited by the
160        internal workspace, and dependent on the LINK_SIZE. The code has been
161        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
162        the default depends on LINK_SIZE. There is a new upper limit (for safety)
163        of around 200,000 forward references. While doing this, I also speeded up
164        the filling in of repeated forward references.
165    
166    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
167        incorrectly expecting the subject to contain another "a" after the start.
168    
169    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
170        in the match, the SKIP should be ignored. This was not happening; instead
171        the SKIP was being treated as NOMATCH. For patterns such as
172        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
173        tested.
174    
175    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
176        now much more compatible with Perl, in particular in cases where the result
177        is a non-match for a non-anchored pattern. For example, if
178        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
179        "m", where previously it did not return a name. A side effect of this
180        change is that for partial matches, the last encountered mark name is
181        returned, as for non matches. A number of tests that were previously not
182        Perl-compatible have been moved into the Perl-compatible test files. The
183        refactoring has had the pleasing side effect of removing one argument from
184        the match() function, thus reducing its stack requirements.
185    
186    19. If the /S+ option was used in pcretest to study a pattern using JIT,
187        subsequent uses of /S (without +) incorrectly behaved like /S+.
188    
189    21. Retrieve executable code size support for the JIT compiler and fixing
190        some warnings.
191    
192    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
193        not work when the shorter character appeared right at the end of the
194        subject string.
195    
196    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
197        systems.
198    
199    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
200        output it when the /M option is used in pcretest.
201    
202    25. The CheckMan script was not being included in the distribution. Also, added
203        an explicit "perl" to run Perl scripts from the PrepareRelease script
204        because this is reportedly needed in Windows.
205    
206    26. If study data was being save in a file and studying had not found a set of
207        "starts with" bytes for the pattern, the data written to the file (though
208        never used) was taken from uninitialized memory and so caused valgrind to
209        complain.
210    
211    27. Updated RunTest.bat as provided by Sheri Pierce.
212    
213    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
214    
215    29. Computation of memory usage for the table of capturing group names was
216        giving an unnecessarily large value.
217    
218    
219    Version 8.20 21-Oct-2011
220    ------------------------
221    
222    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
223        a POSIX class. After further experiments with Perl, which convinced me that
224        Perl has bugs and confusions, a closing square bracket is no longer allowed
225        in a POSIX name. This bug also affected patterns with classes that started
226        with full stops.
227    
228    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
229        captured substring, but while checking the failing first alternative,
230        substring 1 is temporarily captured. If the output vector supplied to
231        pcre_exec() was not big enough for this capture, the yield of the function
232        was still zero ("insufficient space for captured substrings"). This cannot
233        be totally fixed without adding another stack variable, which seems a lot
234        of expense for a edge case. However, I have improved the situation in cases
235        such as /(a)(b)x|abc/ matched against "abc", where the return code
236        indicates that fewer than the maximum number of slots in the ovector have
237        been set.
238    
239    3.  Related to (2) above: when there are more back references in a pattern than
240        slots in the output vector, pcre_exec() uses temporary memory during
241        matching, and copies in the captures as far as possible afterwards. It was
242        using the entire output vector, but this conflicts with the specification
243        that only 2/3 is used for passing back captured substrings. Now it uses
244        only the first 2/3, for compatibility. This is, of course, another edge
245        case.
246    
247    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
248        main code base, and can be used by building with --enable-jit. When this is
249        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
250        runtime --no-jit option is given.
251    
252    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
253        ovector, the return from the function was zero, implying that there were
254        other matches that did not fit. The correct "exactly full" value is now
255        returned.
256    
257    6.  If a subpattern that was called recursively or as a subroutine contained
258        (*PRUNE) or any other control that caused it to give a non-standard return,
259        invalid errors such as "Error -26 (nested recursion at the same subject
260        position)" or even infinite loops could occur.
261    
262    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
263        computing the minimum length on reaching *ACCEPT, and so ended up with the
264        wrong value of 1 rather than 0. Further investigation indicates that
265        computing a minimum subject length in the presence of *ACCEPT is difficult
266        (think back references, subroutine calls), and so I have changed the code
267        so that no minimum is registered for a pattern that contains *ACCEPT.
268    
269    8.  If (*THEN) was present in the first (true) branch of a conditional group,
270        it was not handled as intended. [But see 16 below.]
271    
272    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
273        Sheri Pierce.
274    
275    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
276        the first byte in a match must be "a".
277    
278    11. Change 17 for 8.13 increased the recursion depth for patterns like
279        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
280        pattern contains any instances of (*THEN). If it does not, the old
281        optimizations are restored. It would be nice to do this on a per-group
282        basis, but at the moment that is not feasible.
283    
284    12. In some environments, the output of pcretest -C is CRLF terminated. This
285        broke RunTest's code that checks for the link size. A single white space
286        character after the value is now allowed for.
287    
288    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
289        For "fr", it uses the Windows-specific input and output files.
290    
291    14. If (*THEN) appeared in a group that was called recursively or as a
292        subroutine, it did not work as intended. [But see next item.]
293    
294    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
295        pattern fragments (but not containing any | characters). If A and B are
296        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
297        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
298        D. In other words, Perl considers parentheses that do not contain any |
299        characters to be part of a surrounding alternative, whereas PCRE was
300        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
301        differently. PCRE now behaves in the same way as Perl, except in the case
302        of subroutine/recursion calls such as (?1) which have in any case always
303        been different (but PCRE had them first :-).
304    
305    16. Related to 15 above: Perl does not treat the | in a conditional group as
306        creating alternatives. Such a group is treated in the same way as an
307        ordinary group without any | characters when processing (*THEN). PCRE has
308        been changed to match Perl's behaviour.
309    
310    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
311        RunGrepTest script failed.
312    
313    18. Change 22 for version 13 caused atomic groups to use more stack. This is
314        inevitable for groups that contain captures, but it can lead to a lot of
315        stack use in large patterns. The old behaviour has been restored for atomic
316        groups that do not contain any capturing parentheses.
317    
318    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
319        suppress the check for a minimum subject length at run time. (If it was
320        given to pcre_exec() or pcre_dfa_exec() it did work.)
321    
322    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
323        fail to work when decoding hex characters in data strings in EBCDIC
324        environments.
325    
326    21. It appears that in at least one Mac OS environment, the isxdigit() function
327        is implemented as a macro that evaluates to its argument more than once,
328        contravening the C 90 Standard (I haven't checked a later standard). There
329        was an instance in pcretest which caused it to go wrong when processing
330        \x{...} escapes in subject strings. The has been rewritten to avoid using
331        things like p++ in the argument of isxdigit().
332    
333    
334    Version 8.13 16-Aug-2011
335    ------------------------
336    
337    1.  The Unicode data tables have been updated to Unicode 6.0.0.
338    
339    2.  Two minor typos in pcre_internal.h have been fixed.
340    
341    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
342        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
343        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
344    
345    4.  There were a number of related bugs in the code for matching backrefences
346        caselessly in UTF-8 mode when codes for the characters concerned were
347        different numbers of bytes. For example, U+023A and U+2C65 are an upper
348        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
349        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
350        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
351        2-byte code at the end of the subject (it thought there wasn't enough data
352        left).
353    
354    5.  Comprehensive information about what went wrong is now returned by
355        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
356        as the output vector has at least 2 elements. The offset of the start of
357        the failing character and a reason code are placed in the vector.
358    
359    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
360        now returned is for the first byte of the failing character, instead of the
361        last byte inspected. This is an incompatible change, but I hope it is small
362        enough not to be a problem. It makes the returned offset consistent with
363        pcre_exec() and pcre_dfa_exec().
364    
365    7.  pcretest now gives a text phrase as well as the error number when
366        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
367        failure, the offset and reason code are output.
368    
369    8.  When \R was used with a maximizing quantifier it failed to skip backwards
370        over a \r\n pair if the subsequent match failed. Instead, it just skipped
371        back over a single character (\n). This seems wrong (because it treated the
372        two characters as a single entity when going forwards), conflicts with the
373        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
374        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
375        has been changed.
376    
377    9.  Some internal refactoring has changed the processing so that the handling
378        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
379        time (the PCRE_DOTALL option was changed this way some time ago: version
380        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
381        which was always a bit of a fudge. It also means that there is one less
382        argument for the match() function, which reduces its stack requirements
383        slightly. This change also fixes an incompatibility with Perl: the pattern
384        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
385    
386    10. More internal refactoring has drastically reduced the number of recursive
387        calls to match() for possessively repeated groups such as (abc)++ when
388        using pcre_exec().
389    
390    11. While implementing 10, a number of bugs in the handling of groups were
391        discovered and fixed:
392    
393        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
394        (a|)*(?1) gave a compile-time internal error.
395        ((a|)+)+  did not notice that the outer group could match an empty string.
396        (^a|^)+   was not marked as anchored.
397        (.*a|.*)+ was not marked as matching at start or after a newline.
398    
399    12. Yet more internal refactoring has removed another argument from the match()
400        function. Special calls to this function are now indicated by setting a
401        value in a variable in the "match data" data block.
402    
403    13. Be more explicit in pcre_study() instead of relying on "default" for
404        opcodes that mean there is no starting character; this means that when new
405        ones are added and accidentally left out of pcre_study(), testing should
406        pick them up.
407    
408    14. The -s option of pcretest has been documented for ages as being an old
409        synonym of -m (show memory usage). I have changed it to mean "force study
410        for every regex", that is, assume /S for every regex. This is similar to -i
411        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
412        using it. It makes it easier to run collections of tests with and without
413        study enabled, and thereby test pcre_study() more easily. All the standard
414        tests are now run with and without -s (but some patterns can be marked as
415        "never study" - see 20 below).
416    
417    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
418        restoration of the capturing data to the outer values was not happening
419        correctly.
420    
421    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
422        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
423        pattern had matched an empty string, and so incorrectly returned a no
424        match.
425    
426    17. There was optimizing code for the last branch of non-capturing parentheses,
427        and also for the obeyed branch of a conditional subexpression, which used
428        tail recursion to cut down on stack usage. Unfortunately, now that there is
429        the possibility of (*THEN) occurring in these branches, tail recursion is
430        no longer possible because the return has to be checked for (*THEN). These
431        two optimizations have therefore been removed. [But see 8.20/11 above.]
432    
433    18. If a pattern containing \R was studied, it was assumed that \R always
434        matched two bytes, thus causing the minimum subject length to be
435        incorrectly computed because \R can also match just one byte.
436    
437    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
438        was incorrectly computed.
439    
440    20. If /S is present twice on a test pattern in pcretest input, it now
441        *disables* studying, thereby overriding the use of -s on the command line
442        (see 14 above). This is necessary for one or two tests to keep the output
443        identical in both cases.
444    
445    21. When (*ACCEPT) was used in an assertion that matched an empty string and
446        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
447    
448    22. When an atomic group that contained a capturing parenthesis was
449        successfully matched, but the branch in which it appeared failed, the
450        capturing was not being forgotten if a higher numbered group was later
451        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
452        group 1 to "a", when in fact it should be unset. This applied to multi-
453        branched capturing and non-capturing groups, repeated or not, and also to
454        positive assertions (capturing in negative assertions does not happen
455        in PCRE) and also to nested atomic groups.
456    
457    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
458        subject after a captured substring, to make it easier to tell which of a
459        number of identical substrings has been captured.
460    
461    24. The way atomic groups are processed by pcre_exec() has been changed so that
462        if they are repeated, backtracking one repetition now resets captured
463        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
464        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
465        "aaa". Previously, it would have been "a". As part of this code
466        refactoring, the way recursive calls are handled has also been changed.
467    
468    25. If an assertion condition captured any substrings, they were not passed
469        back unless some other capturing happened later. For example, if
470        (?(?=(a))a) was matched against "a", no capturing was returned.
471    
472    26. When studying a pattern that contained subroutine calls or assertions,
473        the code for finding the minimum length of a possible match was handling
474        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
475        group 1 called group 2 while simultaneously a separate group 2 called group
476        1). A stack overflow occurred in this case. I have fixed this by limiting
477        the recursion depth to 10.
478    
479    27. Updated RunTest.bat in the distribution to the version supplied by Tom
480        Fortmann. This supports explicit test numbers on the command line, and has
481        argument validation and error reporting.
482    
483    28. An instance of \X with an unlimited repeat could fail if at any point the
484        first character it looked at was a mark character.
485    
486    29. Some minor code refactoring concerning Unicode properties and scripts
487        should reduce the stack requirement of match() slightly.
488    
489    30. Added the '=' option to pcretest to check the setting of unused capturing
490        slots at the end of the pattern, which are documented as being -1, but are
491        not included in the return count.
492    
493    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
494        compiled something random. Now it gives a compile-time error (as does
495        Perl).
496    
497    32. A *MARK encountered during the processing of a positive assertion is now
498        recorded and passed back (compatible with Perl).
499    
500    33. If --only-matching or --colour was set on a pcregrep call whose pattern
501        had alternative anchored branches, the search for a second match in a line
502        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
503        matched the line "0102" twice. The same bug affected patterns that started
504        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
505        twice.
506    
507    34. Previously, PCRE did not allow quantification of assertions. However, Perl
508        does, and because of capturing effects, quantifying parenthesized
509        assertions may at times be useful. Quantifiers are now allowed for
510        parenthesized assertions.
511    
512    35. A minor code tidy in pcre_compile() when checking options for \R usage.
513    
514    36. \g was being checked for fancy things in a character class, when it should
515        just be a literal "g".
516    
517    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
518        appearance of a nested POSIX class supersedes an apparent external class.
519        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
520        unescaped square brackets may also appear as part of class names. For
521        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
522        more like Perl. (But see 8.20/1 above.)
523    
524    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
525        was because it thought it was \N{name}, which is not supported).
526    
527    39. Add minix to OS list not supporting the -S option in pcretest.
528    
529    40. PCRE tries to detect cases of infinite recursion at compile time, but it
530        cannot analyze patterns in sufficient detail to catch mutual recursions
531        such as ((?1))((?2)). There is now a runtime test that gives an error if a
532        subgroup is called recursively as a subpattern for a second time at the
533        same position in the subject string. In previous releases this might have
534        been caught by the recursion limit, or it might have run out of stack.
535    
536    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
537        happen only once. PCRE was, however incorrectly giving a compile time error
538        "recursive call could loop indefinitely" because it cannot analyze the
539        pattern in sufficient detail. The compile time test no longer happens when
540        PCRE is compiling a conditional subpattern, but actual runaway loops are
541        now caught at runtime (see 40 above).
542    
543    42. It seems that Perl allows any characters other than a closing parenthesis
544        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
545        has been changed to be the same.
546    
547    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
548        as not to get warnings when autogen.sh is called. Also changed
549        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
550    
551    44. To help people who use pcregrep to scan files containing exceedingly long
552        lines, the following changes have been made:
553    
554        (a) The default value of the buffer size parameter has been increased from
555            8K to 20K. (The actual buffer used is three times this size.)
556    
557        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
558            PCRE is built.
559    
560        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
561            to be set at run time.
562    
563        (d) Numerical values in pcregrep options can be followed by K or M, for
564            example --buffer-size=50K.
565    
566        (e) If a line being scanned overflows pcregrep's buffer, an error is now
567            given and the return code is set to 2.
568    
569    45. Add a pointer to the latest mark to the callout data block.
570    
571    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
572        partial match of an empty string instead of no match. This was specific to
573        the use of ".".
574    
575    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
576        complete match instead of a partial match. This bug was dependent on both
577        the PCRE_UTF8 and PCRE_DOTALL options being set.
578    
579    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
580        starting byte set, because \b was not being ignored.
581    
582    
583    Version 8.12 15-Jan-2011
584    ------------------------
585    
586    1.  Fixed some typos in the markup of the man pages, and wrote a script that
587        checks for such things as part of the documentation building process.
588    
589    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
590        --match-limit and --recursion-limit options (added for 8.11). In
591        particular, this made one of the standard tests fail. (The integer value
592        went into the wrong half of a long int.)
593    
594    3.  If the --colour option was given to pcregrep with -v (invert match), it
595        did strange things, either producing crazy output, or crashing. It should,
596        of course, ignore a request for colour when reporting lines that do not
597        match.
598    
599    4.  Another pcregrep bug caused similar problems if --colour was specified with
600        -M (multiline) and the pattern match finished with a line ending.
601    
602    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
603        matched in multiline mode, the following line was shown as part of the
604        match. This seems wrong, so I have changed it.
605    
606    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
607        the check for further matches in the same line (so they could be coloured)
608        to overrun the end of the current line. If another match was found, it was
609        incorrectly shown (and then shown again when found in the next line).
610    
611    7.  If pcregrep was compiled under Windows, there was a reference to the
612        function pcregrep_exit() before it was defined. I am assuming this was
613        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
614        reported by a user. I've moved the definition above the reference.
615    
616    
617    Version 8.11 10-Dec-2010
618    ------------------------
619    
620    1.  (*THEN) was not working properly if there were untried alternatives prior
621        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
622        backtracked to try for "b" instead of moving to the next alternative branch
623        at the same level (in this case, to look for "c"). The Perl documentation
624        is clear that when (*THEN) is backtracked onto, it goes to the "next
625        alternative in the innermost enclosing group".
626    
627    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
628        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
629        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
630        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
631        (*THEN).
632    
633    3.  If \s appeared in a character class, it removed the VT character from
634        the class, even if it had been included by some previous item, for example
635        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
636        of \s, but is part of the POSIX "space" class.)
637    
638    4.  A partial match never returns an empty string (because you can always
639        match an empty string at the end of the subject); however the checking for
640        an empty string was starting at the "start of match" point. This has been
641        changed to the "earliest inspected character" point, because the returned
642        data for a partial match starts at this character. This means that, for
643        example, /(?<=abc)def/ gives a partial match for the subject "abc"
644        (previously it gave "no match").
645    
646    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
647        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
648        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
649        has an implication that the given string is incomplete (because a partial
650        match is preferred over a full match). For this reason, these items now
651        give a partial match in this situation. [Aside: previously, the one case
652        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
653        match rather than a full match, which was wrong by the old rules, but is
654        now correct.]
655    
656    6.  There was a bug in the handling of #-introduced comments, recognized when
657        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
658        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
659        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
660        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
661        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
662        places in pcre_compile().
663    
664    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
665        comments when looking ahead for named forward references to subpatterns,
666        the only newline sequence it recognized was NL. It now handles newlines
667        according to the set newline convention.
668    
669    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
670        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
671        cater for a lack of strerror(). These oversights have been fixed.
672    
673    9.  Added --match-limit and --recursion-limit to pcregrep.
674    
675    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
676    
677    11. When the -o option was used, pcregrep was setting a return code of 1, even
678        when matches were found, and --line-buffered was not being honoured.
679    
680    12. Added an optional parentheses number to the -o and --only-matching options
681        of pcregrep.
682    
683    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
684        can match an empty string. The code to do it in pcretest and pcredemo
685        needed fixing:
686    
687        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
688            only one byte after an empty string match just before CRLF (this case
689            just got forgotten; "any" and "anycrlf" were OK).
690    
691        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
692            mode when an empty string match preceded an ASCII character followed by
693            a non-ASCII character. (The code for advancing by one character rather
694            than one byte was nonsense.)
695    
696        (c) The pcredemo.c sample program did not have any code at all to handle
697            the cases when CRLF is a valid newline sequence.
698    
699    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
700        as a starting offset was within the subject string. There is now a new
701        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
702        negative or greater than the length of the string. In order to test this,
703        pcretest is extended to allow the setting of negative starting offsets.
704    
705    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
706        starting offset points to the beginning of a UTF-8 character was
707        unnecessarily clumsy. I tidied it up.
708    
709    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
710        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
711    
712    17. Nobody had reported that the --include_dir option, which was added in
713        release 7.7 should have been called --include-dir (hyphen, not underscore)
714        for compatibility with GNU grep. I have changed it to --include-dir, but
715        left --include_dir as an undocumented synonym, and the same for
716        --exclude-dir, though that is not available in GNU grep, at least as of
717        release 2.5.4.
718    
719    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
720        characters from a string of bytes) have been redefined so as not to use
721        loops, in order to improve performance in some environments. At the same
722        time, I abstracted some of the common code into auxiliary macros to save
723        repetition (this should not affect the compiled code).
724    
725    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
726        compile-time error is now given if \c is not followed by an ASCII
727        character, that is, a byte less than 128. (In EBCDIC mode, the code is
728        different, and any byte value is allowed.)
729    
730    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
731        START_OPTIMIZE option, which is now allowed at compile time - but just
732        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
733        to pcregrep and other applications that have no direct access to PCRE
734        options. The new /Y option in pcretest sets this option when calling
735        pcre_compile().
736    
737    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
738        back references. Groups containing recursive back references were forced to
739        be atomic by that change, but in the case of named groups, the amount of
740        memory required was incorrectly computed, leading to "Failed: internal
741        error: code overflow". This has been fixed.
742    
743    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
744        pcretest.c, to avoid build problems in some Borland environments.
745    
746    
747    Version 8.10 25-Jun-2010
748    ------------------------
749    
750    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
751        THEN.
752    
753    2.  (*ACCEPT) was not working when inside an atomic group.
754    
755    3.  Inside a character class, \B is treated as a literal by default, but
756        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
757        causes the error). The code is unchanged, but I tidied the documentation.
758    
759    4.  Inside a character class, PCRE always treated \R and \X as literals,
760        whereas Perl faults them if its -w option is set. I have changed PCRE so
761        that it faults them when PCRE_EXTRA is set.
762    
763    5.  Added support for \N, which always matches any character other than
764        newline. (It is the same as "." when PCRE_DOTALL is not set.)
765    
766    6.  When compiling pcregrep with newer versions of gcc which may have
767        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
768        declared with attribute warn_unused_result" were given. Just casting the
769        result to (void) does not stop the warnings; a more elaborate fudge is
770        needed. I've used a macro to implement this.
771    
772    7.  Minor change to pcretest.c to avoid a compiler warning.
773    
774    8.  Added four artifical Unicode properties to help with an option to make
775        \s etc use properties (see next item). The new properties are: Xan
776        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
777    
778    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
779        use Unicode properties. (*UCP) at the start of a pattern can be used to set
780        this option. Modified pcretest to add /W to test this facility. Added
781        REG_UCP to make it available via the POSIX interface.
782    
783    10. Added --line-buffered to pcregrep.
784    
785    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
786        studied, and the match started with a letter with a code point greater than
787        127 whose first byte was different to the first byte of the other case of
788        the letter, the other case of this starting letter was not recognized
789        (#976).
790    
791    12. If a pattern that was studied started with a repeated Unicode property
792        test, for example, \p{Nd}+, there was the theoretical possibility of
793        setting up an incorrect bitmap of starting bytes, but fortunately it could
794        not have actually happened in practice until change 8 above was made (it
795        added property types that matched character-matching opcodes).
796    
797    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
798        possible starting bytes for non-anchored patterns.
799    
800    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
801        \R, and also a number of cases that involve Unicode properties, both
802        explicit and implicit when PCRE_UCP is set.
803    
804    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
805        input, it could crash or give wrong results if characters with values
806        greater than 0xc0 were present in the subject string. (Detail: it assumed
807        UTF-8 input when processing these items.)
808    
809    16. Added a lot of (int) casts to avoid compiler warnings in systems where
810        size_t is 64-bit (#991).
811    
812    17. Added a check for running out of memory when PCRE is compiled with
813        --disable-stack-for-recursion (#990).
814    
815    18. If the last data line in a file for pcretest does not have a newline on
816        the end, a newline was missing in the output.
817    
818    19. The default pcre_chartables.c file recognizes only ASCII characters (values
819        less than 128) in its various bitmaps. However, there is a facility for
820        generating tables according to the current locale when PCRE is compiled. It
821        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
822        characters, are recognized by isspace() and therefore were getting set in
823        these tables, and indeed these tables seem to approximate to ISO 8859. This
824        caused a problem in UTF-8 mode when pcre_study() was used to create a list
825        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
826        which of course cannot start UTF-8 characters. I have changed the code so
827        that only real ASCII characters (less than 128) and the correct starting
828        bytes for UTF-8 encodings are set for characters greater than 127 when in
829        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
830        altogether.)
831    
832    20. Added the /T option to pcretest so as to be able to run tests with non-
833        standard character tables, thus making it possible to include the tests
834        used for 19 above in the standard set of tests.
835    
836    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
837        reference to a subpattern the other side of a comment that contains an
838        opening parenthesis caused either an internal compiling error, or a
839        reference to the wrong subpattern.
840    
841    
842    Version 8.02 19-Mar-2010
843    ------------------------
844    
845    1.  The Unicode data tables have been updated to Unicode 5.2.0.
846    
847    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
848        configured.
849    
850    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
851        original author of that file, following a query about its status.
852    
853    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
854        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
855    
856    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
857        quantifier applied to a forward-referencing subroutine call, could compile
858        incorrect code or give the error "internal error: previously-checked
859        referenced subpattern not found".
860    
861    6.  Both MS Visual Studio and Symbian OS have problems with initializing
862        variables to point to external functions. For these systems, therefore,
863        pcre_malloc etc. are now initialized to local functions that call the
864        relevant global functions.
865    
866    7.  There were two entries missing in the vectors called coptable and poptable
867        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
868        I've fixed the data, and added a kludgy way of testing at compile time that
869        the lengths are correct (equal to the number of opcodes).
870    
871    8.  Following on from 7, I added a similar kludge to check the length of the
872        eint vector in pcreposix.c.
873    
874    9.  Error texts for pcre_compile() are held as one long string to avoid too
875        much relocation at load time. To find a text, the string is searched,
876        counting zeros. There was no check for running off the end of the string,
877        which could happen if a new error number was added without updating the
878        string.
879    
880    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
881    
882    11. \K was not working if it appeared in an atomic group or in a group that
883        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
884        \K is "not well defined" if used in an assertion. PCRE now accepts it if
885        the assertion is positive, but not if it is negative.
886    
887    12. Change 11 fortuitously reduced the size of the stack frame used in the
888        "match()" function of pcre_exec.c by one pointer. Forthcoming
889        implementation of support for (*MARK) will need an extra pointer on the
890        stack; I have reserved it now, so that the stack frame size does not
891        decrease.
892    
893    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
894        item in branch that calls a recursion is a subroutine call - as in the
895        second branch in the above example - was incorrectly given the compile-
896        time error "recursive call could loop indefinitely" because pcre_compile()
897        was not correctly checking the subroutine for matching a non-empty string.
898    
899    14. The checks for overrunning compiling workspace could trigger after an
900        overrun had occurred. This is a "should never occur" error, but it can be
901        triggered by pathological patterns such as hundreds of nested parentheses.
902        The checks now trigger 100 bytes before the end of the workspace.
903    
904    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
905    
906    
907    Version 8.01 19-Jan-2010
908    ------------------------
909    
910    1.  If a pattern contained a conditional subpattern with only one branch (in
911        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
912        computed the wrong minimum data length (which is of course zero for such
913        subpatterns). This could cause incorrect "no match" results.
914    
915    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
916        the pattern is reset in the first branch, pcre_compile() failed with
917        "internal error: code overflow at offset...". This happened only when
918        the reset was to the original external option setting. (An optimization
919        abstracts leading options settings into an external setting, which was the
920        cause of this.)
921    
922    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
923        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
924        assertion pattern did not match (meaning that the assertion was true), it
925        was incorrectly treated as false if the SKIP had been reached during the
926        matching. This also applied to assertions used as conditions.
927    
928    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
929        assertion subpattern, including such a pattern used as a condition,
930        unpredictable results occurred, instead of the error return
931        PCRE_ERROR_DFA_UITEM.
932    
933    5.  The C++ GlobalReplace function was not working like Perl for the special
934        situation when an empty string is matched. It now does the fancy magic
935        stuff that is necessary.
936    
937    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
938        removed. (These were left over from very, very early versions of PCRE.)
939    
940    7.  Some cosmetic changes to the code to make life easier when compiling it
941        as part of something else:
942    
943        (a) Change DEBUG to PCRE_DEBUG.
944    
945        (b) In pcre_compile(), rename the member of the "branch_chain" structure
946            called "current" as "current_branch", to prevent a collision with the
947            Linux macro when compiled as a kernel module.
948    
949        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
950            prevent a collision with the Linux macro when compiled as a kernel
951            module.
952    
953    8.  In pcre_compile() there are some checks for integer overflows that used to
954        cast potentially large values to (double). This has been changed to that
955        when building, a check for int64_t is made, and if it is found, it is used
956        instead, thus avoiding the use of floating point arithmetic. (There is no
957        other use of FP in PCRE.) If int64_t is not found, the fallback is to
958        double.
959    
960    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
961        2005 (difference between two addresses compared to an unsigned value).
962    
963    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
964        custom one, because of the following reported problem in Windows:
965    
966          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
967              under Win32.
968          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
969              therefore missing the function definition.
970          - The compiler thus generates a "C" signature for the test function.
971          - The linker fails to find the "C" function.
972          - PCRE fails to configure if asked to do so against libbz2.
973    
974    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
975        messages were output:
976    
977          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
978          rerunning libtoolize, to keep the correct libtool macros in-tree.
979          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
980    
981        I have done both of these things.
982    
983    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
984        most of the time, it *can* run out if it is given a pattern that contains a
985        runaway infinite recursion. I updated the discussion in the pcrestack man
986        page.
987    
988    13. Now that we have gone to the x.xx style of version numbers, the minor
989        version may start with zero. Using 08 or 09 is a bad idea because users
990        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
991        interpreted as invalid octal numbers. I've updated the previous comment in
992        configure.ac, and also added a check that gives an error if 08 or 09 are
993        used.
994    
995    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
996        causing partial matching to fail when the end of the subject matched \W
997        in a UTF-8 pattern where \W was quantified with a minimum of 3.
998    
999    15. There were some discrepancies between the declarations in pcre_internal.h
1000        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1001        their definitions. The declarations used "const uschar *" and the
1002        definitions used USPTR. Even though USPTR is normally defined as "const
1003        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1004        reported that: "This difference in casting confuses some C++ compilers, for
1005        example, SunCC recognizes above declarations as different functions and
1006        generates broken code for hbpcre." I have changed the declarations to use
1007        USPTR.
1008    
1009    16. GNU libtool is named differently on some systems. The autogen.sh script now
1010        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1011        (FreeBSD).
1012    
1013    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1014        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1015        comment: "Figure out how to create a longlong from a string: strtoll and
1016        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1017        instance, but it only takes 2 args instead of 3!"
1018    
1019    18. A subtle bug concerned with back references has been fixed by a change of
1020        specification, with a corresponding code fix. A pattern such as
1021        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1022        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1023        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1024        same bug. Such groups have to be quantified to be useful, or contained
1025        inside another quantified group. (If there's no repetition, the reference
1026        can never match.) The problem arises because, having left the group and
1027        moved on to the rest of the pattern, a later failure that backtracks into
1028        the group uses the captured value from the final iteration of the group
1029        rather than the correct earlier one. I have fixed this in PCRE by forcing
1030        any group that contains a reference to itself to be an atomic group; that
1031        is, there cannot be any backtracking into it once it has completed. This is
1032        similar to recursive and subroutine calls.
1033    
1034    
1035    Version 8.00 19-Oct-09
1036  ----------------------  ----------------------
1037    
1038    1.  The table for translating pcre_compile() error codes into POSIX error codes
1039        was out-of-date, and there was no check on the pcre_compile() error code
1040        being within the table. This could lead to an OK return being given in
1041        error.
1042    
1043    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1044        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1045        in a Windows environment.
1046    
1047    3.  The pcregrep --count option prints the count for each file even when it is
1048        zero, as does GNU grep. However, pcregrep was also printing all files when
1049        --files-with-matches was added. Now, when both options are given, it prints
1050        counts only for those files that have at least one match. (GNU grep just
1051        prints the file name in this circumstance, but including the count seems
1052        more useful - otherwise, why use --count?) Also ensured that the
1053        combination -clh just lists non-zero counts, with no names.
1054    
1055    4.  The long form of the pcregrep -F option was incorrectly implemented as
1056        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1057        but it seems right to fix it, and I didn't think it was worth preserving
1058        the old behaviour.
1059    
1060    5.  The command line items --regex=pattern and --regexp=pattern were not
1061        recognized by pcregrep, which required --regex pattern or --regexp pattern
1062        (with a space rather than an '='). The man page documented the '=' forms,
1063        which are compatible with GNU grep; these now work.
1064    
1065    6.  No libpcreposix.pc file was created for pkg-config; there was just
1066        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1067    
1068    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1069        when UCP support is not needed, by modifying the Python script that
1070        generates it from Unicode data files. This should not matter if the module
1071        is correctly used as a library, but I received one complaint about 50K of
1072        unwanted data. My guess is that the person linked everything into his
1073        program rather than using a library. Anyway, it does no harm.
1074    
1075    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1076        was a minimum greater than 1 for a wide character in a possessive
1077        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1078        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1079        character. Chaos in the form of incorrect output or a compiling loop could
1080        result.
1081    
1082    9.  The restrictions on what a pattern can contain when partial matching is
1083        requested for pcre_exec() have been removed. All patterns can now be
1084        partially matched by this function. In addition, if there are at least two
1085        slots in the offset vector, the offset of the earliest inspected character
1086        for the match and the offset of the end of the subject are set in them when
1087        PCRE_ERROR_PARTIAL is returned.
1088    
1089    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1090        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1091        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1092        and may be more useful for multi-segment matching.
1093    
1094    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1095        used to be given if ever the end of the subject was reached; now it is
1096        given only if matching could not proceed because another character was
1097        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1098        string "Z", which now yields "no match" instead of "partial match". In the
1099        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1100        final character ended with (*FAIL).
1101    
1102    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1103        if the pattern had a "must contain" character that was already found in the
1104        earlier partial match, unless partial matching was again requested. For
1105        example, with the pattern /dog.(body)?/, the "must contain" character is
1106        "g". If the first part-match was for the string "dog", restarting with
1107        "sbody" failed. This bug has been fixed.
1108    
1109    13. The string returned by pcre_dfa_exec() after a partial match has been
1110        changed so that it starts at the first inspected character rather than the
1111        first character of the match. This makes a difference only if the pattern
1112        starts with a lookbehind assertion or \b or \B (\K is not supported by
1113        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1114        matching functions compatible, and I think it's the right thing to do.
1115    
1116    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1117        so that the demonstration program is easily available in environments where
1118        PCRE has not been installed from source.
1119    
1120    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1121        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1122        library.
1123    
1124    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1125        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1126        is not the first non-POSIX option to be added. Clearly some people find
1127        these options useful.
1128    
1129    17. If a caller to the POSIX matching function regexec() passes a non-zero
1130        value for nmatch with a NULL value for pmatch, the value of
1131        nmatch is forced to zero.
1132    
1133    18. RunGrepTest did not have a test for the availability of the -u option of
1134        the diff command, as RunTest does. It now checks in the same way as
1135        RunTest, and also checks for the -b option.
1136    
1137    19. If an odd number of negated classes containing just a single character
1138        interposed, within parentheses, between a forward reference to a named
1139        subpattern and the definition of the subpattern, compilation crashed with
1140        an internal error, complaining that it could not find the referenced
1141        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1142        [The bug was that it was starting one character too far in when skipping
1143        over the character class, thus treating the ] as data rather than
1144        terminating the class. This meant it could skip too much.]
1145    
1146    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1147        /g option in pcretest when the pattern contains \K, which makes it possible
1148        to have an empty string match not at the start, even when the pattern is
1149        anchored. Updated pcretest and pcredemo to use this option.
1150    
1151    21. If the maximum number of capturing subpatterns in a recursion was greater
1152        than the maximum at the outer level, the higher number was returned, but
1153        with unset values at the outer level. The correct (outer level) value is
1154        now given.
1155    
1156    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1157        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1158        make it do so. The string so far is captured, making this feature
1159        compatible with Perl.
1160    
1161    23. The tests have been re-organized, adding tests 11 and 12, to make it
1162        possible to check the Perl 5.10 features against Perl 5.10.
1163    
1164    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1165        pattern matches a fixed length string. PCRE did not allow this; now it
1166        does. Neither allows recursion.
1167    
1168    25. I finally figured out how to implement a request to provide the minimum
1169        length of subject string that was needed in order to match a given pattern.
1170        (It was back references and recursion that I had previously got hung up
1171        on.) This code has now been added to pcre_study(); it finds a lower bound
1172        to the length of subject needed. It is not necessarily the greatest lower
1173        bound, but using it to avoid searching strings that are too short does give
1174        some useful speed-ups. The value is available to calling programs via
1175        pcre_fullinfo().
1176    
1177    26. While implementing 25, I discovered to my embarrassment that pcretest had
1178        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1179        study optimizations had never been tested with that matching function.
1180        Oops. What is worse, even when it was passed study data, there was a bug in
1181        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1182        were also very few tests of studied patterns with pcre_dfa_exec().
1183    
1184    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1185        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1186        on the other side of the coin, they are no longer allowed to have different
1187        names, because these cannot be distinguished in PCRE, and this has caused
1188        confusion. (This is a difference from Perl.)
1189    
1190    28. When duplicate subpattern names are present (necessarily with different
1191        numbers, as required by 27 above), and a test is made by name in a
1192        conditional pattern, either for a subpattern having been matched, or for
1193        recursion in such a pattern, all the associated numbered subpatterns are
1194        tested, and the overall condition is true if the condition is true for any
1195        one of them. This is the way Perl works, and is also more like the way
1196        testing by number works.
1197    
1198    
1199    Version 7.9 11-Apr-09
1200    ---------------------
1201    
1202    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1203        (pcretest), all targets were linked against these libraries. This included
1204        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1205        libraries. This caused unwanted dependencies to be created. This problem
1206        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1207        pcretest is linked with readline.
1208    
1209    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1210        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1211        moved outside it again, because FALSE and TRUE are already defined in AIX,
1212        but BOOL is not.
1213    
1214    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1215        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1216    
1217    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1218        hyphens) following file names and line numbers when outputting matching
1219        lines. This is not true; no spaces are inserted. I have also clarified the
1220        wording for the --colour (or --color) option.
1221    
1222    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1223        was not coloured; this is different to GNU grep, so I have changed it to be
1224        the same.
1225    
1226    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1227        each matching line was coloured. Now it goes on to look for further matches
1228        of any of the test patterns, which is the same behaviour as GNU grep.
1229    
1230    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1231        doesn't make sense to accept an empty string match in pcregrep, so I have
1232        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1233        seems to be how GNU grep behaves.
1234    
1235    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1236        start or after a newline", because the conditional assertion was not being
1237        correctly handled. The rule now is that both the assertion and what follows
1238        in the first alternative must satisfy the test.
1239    
1240    9.  If auto-callout was enabled in a pattern with a conditional group whose
1241        condition was an assertion, PCRE could crash during matching, both with
1242        pcre_exec() and pcre_dfa_exec().
1243    
1244    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1245        used for matching.
1246    
1247    11. Unicode property support in character classes was not working for
1248        characters (bytes) greater than 127 when not in UTF-8 mode.
1249    
1250    12. Added the -M command line option to pcretest.
1251    
1252    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1253    
1254    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1255    
1256    16. Added comments and documentation about mis-use of no_arg in the C++
1257        wrapper.
1258    
1259    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1260        from Martin Jerabek that uses macro names for all relevant character and
1261        string constants.
1262    
1263    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1264        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1265        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1266        these, but not everybody uses configure.
1267    
1268    19. A conditional group that had only one branch was not being correctly
1269        recognized as an item that could match an empty string. This meant that an
1270        enclosing group might also not be so recognized, causing infinite looping
1271        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1272        with the subject "ab", where knowledge that the repeated group can match
1273        nothing is needed in order to break the loop.
1274    
1275    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1276        exec(), but without supplying a callout function, matching went wrong.
1277    
1278    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1279        leak if the size of the offset vector was greater than 30. When the vector
1280        is smaller, the saved offsets during recursion go onto a local stack
1281        vector, but for larger vectors malloc() is used. It was failing to free
1282        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1283        error, in fact).
1284    
1285    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1286        heapframe that is used only when UTF-8 support is enabled. This caused no
1287        problem, but was untidy.
1288    
1289    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1290        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1291        included within another project.
1292    
1293    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1294        slightly modified by me:
1295    
1296          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1297              not building pcregrep.
1298    
1299          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1300              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1301    
1302    25. Forward references, both numeric and by name, in patterns that made use of
1303        duplicate group numbers, could behave incorrectly or give incorrect errors,
1304        because when scanning forward to find the reference group, PCRE was not
1305        taking into account the duplicate group numbers. A pattern such as
1306        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1307    
1308    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1309        the feature of a custom pointer more persuasive (as requested by a user).
1310    
1311    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1312        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1313        pre-defined.
1314    
1315    28. Added support for (*UTF8) at the start of a pattern.
1316    
1317    29. Arrange for flags added by the "release type" setting in CMake to be shown
1318        in the configuration summary.
1319    
1320    
1321    Version 7.8 05-Sep-08
1322    ---------------------
1323    
1324    1.  Replaced UCP searching code with optimized version as implemented for Ad
1325        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1326        stage table and inline lookup instead of a function, giving speed ups of 2
1327        to 5 times on some simple patterns that I tested. Permission was given to
1328        distribute the MultiStage2.py script that generates the tables (it's not in
1329        the tarball, but is in the Subversion repository).
1330    
1331    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1332        scripts.
1333    
1334    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1335        a group with a zero qualifier. The result of the study could be incorrect,
1336        or the function might crash, depending on the pattern.
1337    
1338    4.  Caseless matching was not working for non-ASCII characters in back
1339        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1340        It now works when Unicode Property Support is available.
1341    
1342    5.  In pcretest, an escape such as \x{de} in the data was always generating
1343        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1344        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1345        truncation.
1346    
1347    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1348    
1349    7.  Added two (int) casts to pcregrep when printing the difference of two
1350        pointers, in case they are 64-bit values.
1351    
1352    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1353        test 2 if it fails.
1354    
1355    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1356        and a #define of that name to empty if it is not externally set. This is to
1357        allow users of MSVC to set it if necessary.
1358    
1359    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1360        the convenience functions in the pcre_get.c source file.
1361    
1362    11. An option change at the start of a pattern that had top-level alternatives
1363        could cause overwriting and/or a crash. This command provoked a crash in
1364        some environments:
1365    
1366          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1367    
1368        This potential security problem was recorded as CVE-2008-2371.
1369    
1370    12. For a pattern where the match had to start at the beginning or immediately
1371        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1372        pcre_dfa_exec() could read past the end of the passed subject if there was
1373        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1374        pcretest so that it places the subject at the end of its malloc-ed buffer.
1375    
1376    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1377        exec() might read past the end of the data buffer in UTF-8 mode.
1378    
1379    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1380        the data contained the byte 0x85 as part of a UTF-8 character within its
1381        first line. This applied both to normal and DFA matching.
1382    
1383    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1384        /^[^d]*?$/8 failed to match "abc".
1385    
1386    16. Added a missing copyright notice to pcrecpp_internal.h.
1387    
1388    17. Make it more clear in the documentation that values returned from
1389        pcre_exec() in ovector are byte offsets, not character counts.
1390    
1391    18. Tidied a few places to stop certain compilers from issuing warnings.
1392    
1393    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1394        supplied by Stefan Weber. I made a further small update for 7.8 because
1395        there is a change of source arrangements: the pcre_searchfuncs.c module is
1396        replaced by pcre_ucd.c.
1397    
1398    
1399    Version 7.7 07-May-08
1400    ---------------------
1401    
1402    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1403        a string to a long long, pretend we don't even have a long long." This is
1404        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1405    
1406    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1407        pre-7.6 versions, which defined a global no_arg variable instead of putting
1408        it in the RE class. (See also #8 below.)
1409    
1410    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1411        Lopes.
1412    
1413    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1414    
1415        (1) The include/exclude patterns were being applied to the whole pathnames
1416            of files, instead of just to the final components.
1417    
1418        (2) If there was more than one level of directory, the subdirectories were
1419            skipped unless they satisfied the include/exclude conditions. This is
1420            inconsistent with GNU grep (and could even be seen as contrary to the
1421            pcregrep specification - which I improved to make it absolutely clear).
1422            The action now is always to scan all levels of directory, and just
1423            apply the include/exclude patterns to regular files.
1424    
1425    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1426        --exclude_dir in the tests to avoid scanning .svn directories.
1427    
1428    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1429        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1430        doesn't support NULs in patterns.
1431    
1432    7.  Added some missing "const"s to declarations of static tables in
1433        pcre_compile.c and pcre_dfa_exec.c.
1434    
1435    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1436        caused by fix #2  above. (Subsequently also a second patch to fix the
1437        first patch. And a third patch - this was a messy problem.)
1438    
1439    9.  Applied Craig's patch to remove the use of push_back().
1440    
1441    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1442        matching function regexec().
1443    
1444    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1445        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1446        references. PCRE supports relative numbers with this syntax (I don't think
1447        Oniguruma does).
1448    
1449    12. Previously, a group with a zero repeat such as (...){0} was completely
1450        omitted from the compiled regex. However, this means that if the group
1451        was called as a subroutine from elsewhere in the pattern, things went wrong
1452        (an internal error was given). Such groups are now left in the compiled
1453        pattern, with a new opcode that causes them to be skipped at execution
1454        time.
1455    
1456    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1457        to the way PCRE behaves:
1458    
1459        (a) A lone ] character is dis-allowed (Perl treats it as data).
1460    
1461        (b) A back reference to an unmatched subpattern matches an empty string
1462            (Perl fails the current match path).
1463    
1464        (c) A data ] in a character class must be notated as \] because if the
1465            first data character in a class is ], it defines an empty class. (In
1466            Perl it is not possible to have an empty class.) The empty class []
1467            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1468            The negative empty class [^] matches any one character, independently
1469            of the DOTALL setting.
1470    
1471    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1472        non-existent subpattern following a character class starting with ']' and
1473        containing () gave an internal compiling error instead of "reference to
1474        non-existent subpattern". Fortunately, when the pattern did exist, the
1475        compiled code was correct. (When scanning forwards to check for the
1476        existencd of the subpattern, it was treating the data ']' as terminating
1477        the class, so got the count wrong. When actually compiling, the reference
1478        was subsequently set up correctly.)
1479    
1480    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1481        it was being rejected as not supported by pcre_dfa_exec(), even though
1482        other assertions are supported. I have made pcre_dfa_exec() support
1483        (*FAIL).
1484    
1485    16. The implementation of 13c above involved the invention of a new opcode,
1486        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1487        cannot be changed at match time, I realized I could make a small
1488        improvement to matching performance by compiling OP_ALLANY instead of
1489        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1490        on the OP_ANY path.
1491    
1492    17. Compiling pcretest on Windows with readline support failed without the
1493        following two fixes: (1) Make the unistd.h include conditional on
1494        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1495    
1496    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1497        ncurses library to be included for pcretest when ReadLine support is
1498        requested, but also to allow for it to be overridden. This patch came from
1499        Daniel Bergström.
1500    
1501    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1502        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1503        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1504        spotting this.
1505    
1506    
1507    Version 7.6 28-Jan-08
1508    ---------------------
1509    
1510    1.  A character class containing a very large number of characters with
1511        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1512        overflow.
1513    
1514    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1515        HAVE_LONG_LONG is not defined.
1516    
1517    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1518        bring them up to date and include new features. This patch includes:
1519    
1520        - Fixed PH's badly added libz and libbz2 support.
1521        - Fixed a problem with static linking.
1522        - Added pcredemo. [But later removed - see 7 below.]
1523        - Fixed dftables problem and added an option.
1524        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1525            HAVE_LONG_LONG.
1526        - Added readline support for pcretest.
1527        - Added an listing of the option settings after cmake has run.
1528    
1529    4.  A user submitted a patch to Makefile that makes it easy to create
1530        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1531        Makefile.am that cause it to include this special target, without
1532        affecting anything else. Note that the same mingw target plus all
1533        the other distribution libraries and programs are now supported
1534        when configuring with CMake (see 6 below) instead of with
1535        Configure/Make.
1536    
1537    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1538        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1539        exported in the Windows port". It has not yet been confirmed that the patch
1540        solves the problem, but it does no harm.
1541    
1542    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1543        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1544        with CMake, and also correct the comment about stack recursion.
1545    
1546    7.  Remove the automatic building of pcredemo from the ./configure system and
1547        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1548        of a program that users should build themselves after PCRE is installed, so
1549        building it automatically is not really right. What is more, it gave
1550        trouble in some build environments.
1551    
1552    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1553    
1554    
1555    Version 7.5 10-Jan-08
1556    ---------------------
1557    
1558    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1559        values in parens when parsing an RE using the C++ wrapper."
1560    
1561    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1562        Characters greater than 255 were excluded from the class instead of being
1563        included.
1564    
1565    3.  The same bug as (2) above applied to negated POSIX classes such as
1566        [:^space:].
1567    
1568    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1569        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1570        I have changed it.
1571    
1572    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1573        first named subpattern) and a construct such as (?&a) would reference the
1574        first named subpattern whose name started with "a" (in other words, the
1575        length check was missing). Both these problems are fixed. "Subpattern name
1576        expected" is now given for (?&) (a zero-length name), and this patch also
1577        makes it give the same error for \k'' (previously it complained that that
1578        was a reference to a non-existent subpattern).
1579    
1580    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1581        this is right because (?- can be followed by option settings as well as by
1582        digits. I have, however, made the messages clearer.
1583    
1584    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1585        than the number used in the conditional) now cause a compile-time error.
1586        This is actually not compatible with Perl, which accepts such patterns, but
1587        treats the conditional as always being FALSE (as PCRE used to), but it
1588        seems to me that giving a diagnostic is better.
1589    
1590    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1591        and messages.
1592    
1593    9.  Fix two occurrences of "backslash" in comments that should have been
1594        "backspace".
1595    
1596    10. Remove two redundant lines of code that can never be obeyed (their function
1597        was moved elsewhere).
1598    
1599    11. The program that makes PCRE's Unicode character property table had a bug
1600        which caused it to generate incorrect table entries for sequences of
1601        characters that have the same character type, but are in different scripts.
1602        It amalgamated them into a single range, with the script of the first of
1603        them. In other words, some characters were in the wrong script. There were
1604        thirteen such cases, affecting characters in the following ranges:
1605    
1606          U+002b0 - U+002c1
1607          U+0060c - U+0060d
1608          U+0061e - U+00612
1609          U+0064b - U+0065e
1610          U+0074d - U+0076d
1611          U+01800 - U+01805
1612          U+01d00 - U+01d77
1613          U+01d9b - U+01dbf
1614          U+0200b - U+0200f
1615          U+030fc - U+030fe
1616          U+03260 - U+0327f
1617          U+0fb46 - U+0fbb1
1618          U+10450 - U+1049d
1619    
1620    12. The -o option (show only the matching part of a line) for pcregrep was not
1621        compatible with GNU grep in that, if there was more than one match in a
1622        line, it showed only the first of them. It now behaves in the same way as
1623        GNU grep.
1624    
1625    13. If the -o and -v options were combined for pcregrep, it printed a blank
1626        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1627        does the same. The return code can be used to tell if there were any
1628        non-matching lines.
1629    
1630    14. Added --file-offsets and --line-offsets to pcregrep.
1631    
1632    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1633        infinitely looping recursion. The bug was that positive lookaheads were not
1634        being skipped when checking for a possible empty match (negative lookaheads
1635        and both kinds of lookbehind were skipped).
1636    
1637    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1638        inclusion of <windows.h> to before rather than after the definition of
1639        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1640    
1641    17. Specifying a possessive quantifier with a specific limit for a Unicode
1642        character property caused pcre_compile() to compile bad code, which led at
1643        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1644        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1645        caused the error; without that there was no problem.
1646    
1647    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1648    
1649    19. Added --enable-pcretest-libreadline.
1650    
1651    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1652        RE::GlobalReplace(). As a result, the number of replacements returned was
1653        double what it should be. I removed one of the increments, but Craig sent a
1654        later patch that removed the other one (the right fix) and added unit tests
1655        that check the return values (which was not done before).
1656    
1657    21. Several CMake things:
1658    
1659        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1660            the names libpcre and libpcreposix, not just pcre and pcreposix.
1661    
1662        (2) The above change means that pcretest and pcregrep are now correctly
1663            linked with the newly-built libraries, not previously installed ones.
1664    
1665        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1666    
1667    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1668        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1669        UTF-8 newline character). The key issue is that the pattern starts .*;
1670        this means that the match must be either at the beginning, or after a
1671        newline. The bug was in the code for advancing after a failed match and
1672        checking that the new position followed a newline. It was not taking
1673        account of UTF-8 characters correctly.
1674    
1675    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1676        character classes. PCRE was not treating the sequence [:...:] as a
1677        character class unless the ... were all letters. Perl, however, seems to
1678        allow any characters between [: and :], though of course it rejects as
1679        unknown any "names" that contain non-letters, because all the known class
1680        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1681        for example, whereas PCRE did not - it did not recognize a POSIX character
1682        class. This seemed a bit dangerous, so the code has been changed to be
1683        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1684        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1685        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1686        Perl does, and where it didn't before.
1687    
1688    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1689        Windows environments %n is disabled by default.
1690    
1691    
1692    Version 7.4 21-Sep-07
1693    ---------------------
1694    
1695    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1696        means that a class such as [\s] counted as "explicit reference to CR or
1697        LF". That isn't really right - the whole point of the change was to try to
1698        help when there was an actual mention of one of the two characters. So now
1699        the change happens only if \r or \n (or a literal CR or LF) character is
1700        encountered.
1701    
1702    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1703        of both had grown to the point where there were only 3 bits left.
1704        Fortunately, there was spare space in the data structure, and so I have
1705        moved the internal flags into a new 16-bit field to free up more option
1706        bits.
1707    
1708    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1709        but did not set the internal JCHANGED flag - either of these is enough to
1710        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1711        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1712        start sets both bits.
1713    
1714    4.  Added options (at build time, compile time, exec time) to change \R from
1715        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1716    
1717    5.  doc/pcresyntax.html was missing from the distribution.
1718    
1719    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1720        compatibility, even though it is no longer used.
1721    
1722    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1723        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1724        windows.h file is present (where different names are used). [This was
1725        reversed later after testing - see 16 below.]
1726    
1727    8.  Changed all #include <config.h> to #include "config.h". There were also
1728        some further <pcre.h> cases that I changed to "pcre.h".
1729    
1730    9.  When pcregrep was used with the --colour option, it missed the line ending
1731        sequence off the lines that it output.
1732    
1733    10. It was pointed out to me that arrays of string pointers cause lots of
1734        relocations when a shared library is dynamically loaded. A technique of
1735        using a single long string with a table of offsets can drastically reduce
1736        these. I have refactored PCRE in four places to do this. The result is
1737        dramatic:
1738    
1739          Originally:                          290
1740          After changing UCP table:            187
1741          After changing error message table:   43
1742          After changing table of "verbs"       36
1743          After changing table of Posix names   22
1744    
1745        Thanks to the folks working on Gregex for glib for this insight.
1746    
1747    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1748        unicode-properties was also set.
1749    
1750    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1751    
1752    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1753        checked only for CRLF.
1754    
1755    14. Added casts to pcretest.c to avoid compiler warnings.
1756    
1757    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1758    
1759    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1760        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1761        entirely. This removes changes made in 7 above.
1762    
1763    17. The CMake files have been updated, and there is now more information about
1764        building with CMake in the NON-UNIX-USE document.
1765    
1766    
1767    Version 7.3 28-Aug-07
1768    ---------------------
1769    
1770     1. In the rejigging of the build system that eventually resulted in 7.1, the
1771        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1772        brackets there is not right, since it causes compilers to look for an
1773        installed pcre.h, not the version that is in the source that is being
1774        compiled (which of course may be different). I have changed it back to:
1775    
1776          #include "pcre.h"
1777    
1778        I have a vague recollection that the change was concerned with compiling in
1779        different directories, but in the new build system, that is taken care of
1780        by the VPATH setting the Makefile.
1781    
1782     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1783        when the subject happened to end in the byte 0x85 (e.g. if the last
1784        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1785        characters but of course it shouldn't be taken as a newline when it is part
1786        of another character. The bug was that, for an unlimited repeat of . in
1787        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1788        characters when looking for a newline.
1789    
1790     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1791    
1792     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1793        in debug output.
1794    
1795     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1796        long printing in the pcrecpp unittest when running under MinGW.
1797    
1798     6. ESC_K was left out of the EBCDIC table.
1799    
1800     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1801        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1802        limit also applies to "virtual nesting" when a pattern is recursive, and in
1803        this case 1000 isn't so big. I have been able to remove this limit at the
1804        expense of backing off one optimization in certain circumstances. Normally,
1805        when pcre_exec() would call its internal match() function recursively and
1806        immediately return the result unconditionally, it uses a "tail recursion"
1807        feature to save stack. However, when a subpattern that can match an empty
1808        string has an unlimited repetition quantifier, it no longer makes this
1809        optimization. That gives it a stack frame in which to save the data for
1810        checking that an empty string has been matched. Previously this was taken
1811        from the 1000-entry workspace that had been reserved. So now there is no
1812        explicit limit, but more stack is used.
1813    
1814     8. Applied Daniel's patches to solve problems with the import/export magic
1815        syntax that is required for Windows, and which was going wrong for the
1816        pcreposix and pcrecpp parts of the library. These were overlooked when this
1817        problem was solved for the main library.
1818    
1819     9. There were some crude static tests to avoid integer overflow when computing
1820        the size of patterns that contain repeated groups with explicit upper
1821        limits. As the maximum quantifier is 65535, the maximum group length was
1822        set at 30,000 so that the product of these two numbers did not overflow a
1823        32-bit integer. However, it turns out that people want to use groups that
1824        are longer than 30,000 bytes (though not repeat them that many times).
1825        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1826        made it possible to implement the integer overflow checks in a much more
1827        dynamic way, which I have now done. The artificial limitation on group
1828        length has been removed - we now have only the limit on the total length of
1829        the compiled pattern, which depends on the LINK_SIZE setting.
1830    
1831    10. Fixed a bug in the documentation for get/copy named substring when
1832        duplicate names are permitted. If none of the named substrings are set, the
1833        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1834        empty string.
1835    
1836    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1837        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1838        because the ] is interpreted as the first data character and the
1839        terminating ] is not found. PCRE has been made compatible with Perl in this
1840        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1841        cause memory overwriting.
1842    
1843    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1844        string has been matched (to stop an infinite loop). It was not recognizing
1845        a conditional subpattern that could match an empty string if that
1846        subpattern was within another subpattern. For example, it looped when
1847        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1848        condition was not nested. This bug has been fixed.
1849    
1850    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1851        past the start of the subject in the presence of bytes with the top bit
1852        set, for example "\x8aBCD".
1853    
1854    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1855        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1856    
1857    14. Optimized (?!) to (*FAIL).
1858    
1859    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1860        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1861        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1862        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1863        does: it's just the validity check that is more restrictive.
1864    
1865    16. Inserted checks for integer overflows during escape sequence (backslash)
1866        processing, and also fixed erroneous offset values for syntax errors during
1867        backslash processing.
1868    
1869    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1870        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1871    
1872    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1873        caused an overrun.
1874    
1875    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1876        something other than just ASCII characters) inside a group that had an
1877        unlimited repeat caused a loop at compile time (while checking to see
1878        whether the group could match an empty string).
1879    
1880    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1881        [\P{Any}] did so. (Error in the code for printing property names.)
1882    
1883    21. An orphan \E inside a character class could cause a crash.
1884    
1885    22. A repeated capturing bracket such as (A)? could cause a wild memory
1886        reference during compilation.
1887    
1888    23. There are several functions in pcre_compile() that scan along a compiled
1889        expression for various reasons (e.g. to see if it's fixed length for look
1890        behind). There were bugs in these functions when a repeated \p or \P was
1891        present in the pattern. These operators have additional parameters compared
1892        with \d, etc, and these were not being taken into account when moving along
1893        the compiled data. Specifically:
1894    
1895        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1896            length.
1897    
1898        (b) An item such as \pL+ within a repeated group could cause crashes or
1899            loops.
1900    
1901        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1902            "reference to non-existent subpattern" error.
1903    
1904        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1905    
1906    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1907        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1908    
1909    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1910    
1911    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1912        character were causing crashes (broken optimization).
1913    
1914    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1915        \p or \P) caused a compile-time loop.
1916    
1917    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1918        break. For example, the unstudied pattern [\r\n]A does not match the string
1919        "\r\nA" because change 7.0/46 below moves the current point on by two
1920        characters after failing to match at the start. However, the pattern \nA
1921        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1922        the same is true. There doesn't seem any very clean way out of this, but
1923        what I have chosen to do makes the common cases work: PCRE now takes note
1924        of whether there can be an explicit match for \r or \n anywhere in the
1925        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1926        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1927        pattern has explicit CR or LF references.
1928    
1929    29. Added (*CR) etc for changing newline setting at start of pattern.
1930    
1931    
1932    Version 7.2 19-Jun-07
1933    ---------------------
1934    
1935     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1936        which is apparently normally available under Windows.
1937    
1938     2. Re-jig the pcregrep tests with different newline settings in an attempt
1939        to make them independent of the local environment's newline setting.
1940    
1941     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1942    
1943     4. Some of the "internals" tests were previously cut out when the link size
1944        was not 2, because the output contained actual offsets. The recent new
1945        "Z" feature of pcretest means that these can be cut out, making the tests
1946        usable with all link sizes.
1947    
1948     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1949        stack recursion. This gives a massive performance boost under BSD, but just
1950        a small improvement under Linux. However, it saves one field in the frame
1951        in all cases.
1952    
1953     6. Added more features from the forthcoming Perl 5.10:
1954    
1955        (a) (?-n) (where n is a string of digits) is a relative subroutine or
1956            recursion call. It refers to the nth most recently opened parentheses.
1957    
1958        (b) (?+n) is also a relative subroutine call; it refers to the nth next
1959            to be opened parentheses.
1960    
1961        (c) Conditions that refer to capturing parentheses can be specified
1962            relatively, for example, (?(-2)... or (?(+3)...
1963    
1964        (d) \K resets the start of the current match so that everything before
1965            is not part of it.
1966    
1967        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1968    
1969        (f) \g{name} is another synonym - part of Perl 5.10's unification of
1970            reference syntax.
1971    
1972        (g) (?| introduces a group in which the numbering of parentheses in each
1973            alternative starts with the same number.
1974    
1975        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1976    
1977     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1978        PCRE_INFO_JCHANGED.
1979    
1980     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
1981        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1982        for detecting groups that can match an empty string.
1983    
1984     9. A pattern with a very large number of alternatives (more than several
1985        hundred) was running out of internal workspace during the pre-compile
1986        phase, where pcre_compile() figures out how much memory will be needed. A
1987        bit of new cunning has reduced the workspace needed for groups with
1988        alternatives. The 1000-alternative test pattern now uses 12 bytes of
1989        workspace instead of running out of the 4096 that are available.
1990    
1991    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1992    
1993    11. Applied patch from Google to remove an optimization that didn't quite work.
1994        The report of the bug said:
1995    
1996          pcrecpp::RE("a*").FullMatch("aaa") matches, while
1997          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1998          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1999    
2000    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2001        it matched the wrong number of bytes.
2002    
2003    
2004    Version 7.1 24-Apr-07
2005    ---------------------
2006    
2007     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2008        that is more "standard", making use of automake and other Autotools. There
2009        is some re-arrangement of the files and adjustment of comments consequent
2010        on this.
2011    
2012     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2013        for recursive directory scanning broke on some systems because the files
2014        are not scanned in any specific order and on different systems the order
2015        was different. A call to "sort" has been inserted into RunGrepTest for the
2016        approprate test as a short-term fix. In the longer term there may be an
2017        alternative.
2018    
2019     3. I had an email from Eric Raymond about problems translating some of PCRE's
2020        man pages to HTML (despite the fact that I distribute HTML pages, some
2021        people do their own conversions for various reasons). The problems
2022        concerned the use of low-level troff macros .br and .in. I have therefore
2023        removed all such uses from the man pages (some were redundant, some could
2024        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2025        HTML has been updated to handle .nf/.fi and to complain if it encounters
2026        .br or .in.
2027    
2028     4. Updated comments in configure.ac that get placed in config.h.in and also
2029        arranged for config.h to be included in the distribution, with the name
2030        config.h.generic, for the benefit of those who have to compile without
2031        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2032    
2033     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2034        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2035        makevp.bat for the current PCRE, using the additional files
2036        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2037    
2038     6. A Windows user reported a minor discrepancy with test 2, which turned out
2039        to be caused by a trailing space on an input line that had got lost in his
2040        copy. The trailing space was an accident, so I've just removed it.
2041    
2042     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2043        that is needed.
2044    
2045     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2046        as "const" (a) because they are and (b) because it helps the PHP
2047        maintainers who have recently made a script to detect big data structures
2048        in the php code that should be moved to the .rodata section. I remembered
2049        to update Builducptable as well, so it won't revert if ucptable.h is ever
2050        re-created.
2051    
2052     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2053        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2054        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2055        support is not required. This saves 1.5-2K of code, which is important in
2056        some applications.
2057    
2058        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2059        so as not to refer to the tables, even though these functions will never be
2060        called when UTF-8 support is disabled. Otherwise there are problems with a
2061        shared library.
2062    
2063    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2064    
2065        (a) It was defining its arguments as char * instead of void *.
2066    
2067        (b) It was assuming that all moves were upwards in memory; this was true
2068            a long time ago when I wrote it, but is no longer the case.
2069    
2070        The emulated memove() is provided for those environments that have neither
2071        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2072        is clearly not the case, as these two bugs were recently reported.
2073    
2074    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2075        and Detrail to create the HTML documentation, the .txt form of the man
2076        pages, and it removes trailing spaces from listed files. It also creates
2077        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2078        case, it wraps all the #defines with #ifndefs. This script should be run
2079        before "make dist".
2080    
2081    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2082        with Unicode property support.
2083    
2084        (a) For a maximizing quantifier, if the two different cases of the
2085            character were of different lengths in their UTF-8 codings (there are
2086            some cases like this - I found 11), and the matching function had to
2087            back up over a mixture of the two cases, it incorrectly assumed they
2088            were both the same length.
2089    
2090        (b) When PCRE was configured to use the heap rather than the stack for
2091            recursion during matching, it was not correctly preserving the data for
2092            the other case of a UTF-8 character when checking ahead for a match
2093            while processing a minimizing repeat. If the check also involved
2094            matching a wide character, but failed, corruption could cause an
2095            erroneous result when trying to check for a repeat of the original
2096            character.
2097    
2098    13. Some tidying changes to the testing mechanism:
2099    
2100        (a) The RunTest script now detects the internal link size and whether there
2101            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2102            values substituted by "configure". (The RunGrepTest script already did
2103            this for UTF-8.) The configure.ac script no longer substitutes the
2104            relevant variables.
2105    
2106        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2107            with length and offset values. This means that the output is different
2108            for different internal link sizes. Test 2 is skipped for link sizes
2109            other than 2 because of this, bypassing the problem. Unfortunately,
2110            there was also a test in test 3 (the locale tests) that used /B and
2111            failed for link sizes other than 2. Rather than cut the whole test out,
2112            I have added a new /Z option to pcretest that replaces the length and
2113            offset values with spaces. This is now used to make test 3 independent
2114            of link size. (Test 2 will be tidied up later.)
2115    
2116    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2117        segmentation fault instead of returning the appropriate error message.
2118    
2119    15. In multiline mode when the newline sequence was set to "any", the pattern
2120        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2121        This doesn't seem right; it now treats the CRLF combination as the line
2122        ending, and so does not match in that case. It's only a pattern such as ^$
2123        that would hit this one: something like ^ABC$ would have failed after \r
2124        and then tried again after \r\n.
2125    
2126    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2127        in an attempt to make files that differ only in their line terminators
2128        compare equal. This works on Linux.
2129    
2130    17. Under certain error circumstances pcregrep might try to free random memory
2131        as it exited. This is now fixed, thanks to valgrind.
2132    
2133    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2134        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2135        was because its rules for how to advance for /g after matching an empty
2136        string at the end of a line did not allow for this case. They now check for
2137        it specially.
2138    
2139    20. pcretest is supposed to handle patterns and data of any length, by
2140        extending its buffers when necessary. It was getting this wrong when the
2141        buffer for a data line had to be extended.
2142    
2143    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2144        CRLF as a newline sequence.
2145    
2146    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2147        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2148        I have nevertheless tidied it up.
2149    
2150    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2151    
2152    24. Added a man page for pcre-config.
2153    
2154    
2155    Version 7.0 19-Dec-06
2156    ---------------------
2157    
2158     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2159        moving to gcc 4.1.1.
2160    
2161     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2162        sys/time.h, which is documented as needed for this function. It doesn't
2163        seem to matter on Linux, but it showed up on some releases of OS X.
2164    
2165     3. It seems that there are systems where bytes whose values are greater than
2166        127 match isprint() in the "C" locale. The "C" locale should be the
2167        default when a C program starts up. In most systems, only ASCII printing
2168        characters match isprint(). This difference caused the output from pcretest
2169        to vary, making some of the tests fail. I have changed pcretest so that:
2170    
2171        (a) When it is outputting text in the compiled version of a pattern, bytes
2172            other than 32-126 are always shown as hex escapes.
2173    
2174        (b) When it is outputting text that is a matched part of a subject string,
2175            it does the same, unless a different locale has been set for the match
2176            (using the /L modifier). In this case, it uses isprint() to decide.
2177    
2178     4. Fixed a major bug that caused incorrect computation of the amount of memory
2179        required for a compiled pattern when options that changed within the
2180        pattern affected the logic of the preliminary scan that determines the
2181        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2182        that the computed length was too small. The symptoms of this bug were
2183        either the PCRE error "internal error: code overflow" from pcre_compile(),
2184        or a glibc crash with a message such as "pcretest: free(): invalid next
2185        size (fast)". Examples of patterns that provoked this bug (shown in
2186        pcretest format) are:
2187    
2188          /(?-x: )/x
2189          /(?x)(?-x: \s*#\s*)/
2190          /((?i)[\x{c0}])/8
2191          /(?i:[\x{c0}])/8
2192    
2193        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2194        is now done differently.
2195    
2196     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2197        wrapper classes; (b) implement a new function in the C++ scanner that is
2198        more efficient than the old way of doing things because it avoids levels of
2199        recursion in the regex matching; (c) add a paragraph to the documentation
2200        for the FullMatch() function.
2201    
2202     6. The escape sequence \n was being treated as whatever was defined as
2203        "newline". Not only was this contrary to the documentation, which states
2204        that \n is character 10 (hex 0A), but it also went horribly wrong when
2205        "newline" was defined as CRLF. This has been fixed.
2206    
2207     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2208        was being set to -1 for the "end of line" case (supposedly a value that no
2209        character can have). Though this value is never used (the check for end of
2210        line is "zero bytes in current character"), it caused compiler complaints.
2211        I've changed it to 0xffffffff.
2212    
2213     8. In pcre_version.c, the version string was being built by a sequence of
2214        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2215        string (as it is for production releases) called a macro with an empty
2216        argument. The C standard says the result of this is undefined. The gcc
2217        compiler treats it as an empty string (which was what was wanted) but it is
2218        reported that Visual C gives an error. The source has been hacked around to
2219        avoid this problem.
2220    
2221     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2222        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2223        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2224        of them did).
2225    
2226    10. Originally, pcretest opened its input and output without "b"; then I was
2227        told that "b" was needed in some environments, so it was added for release
2228        5.0 to both the input and output. (It makes no difference on Unix-like
2229        systems.) Later I was told that it is wrong for the input on Windows. I've
2230        now abstracted the modes into two macros, to make it easier to fiddle with
2231        them, and removed "b" from the input mode under Windows.
2232    
2233    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2234    
2235    12. Added -help and --help to pcretest as an official way of being reminded
2236        of the options.
2237    
2238    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2239        and pcrecpp.cc because they annoy compilers at high warning levels.
2240    
2241    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2242    
2243    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2244        scripts are not C programs :-) and which was not noticed because it works
2245        on Linux.
2246    
2247    16. pcretest is supposed to handle any length of pattern and data line (as one
2248        line or as a continued sequence of lines) by extending its input buffer if
2249        necessary. This feature was broken for very long pattern lines, leading to
2250        a string of junk being passed to pcre_compile() if the pattern was longer
2251        than about 50K.
2252    
2253    17. I have done a major re-factoring of the way pcre_compile() computes the
2254        amount of memory needed for a compiled pattern. Previously, there was code
2255        that made a preliminary scan of the pattern in order to do this. That was
2256        OK when PCRE was new, but as the facilities have expanded, it has become
2257        harder and harder to keep it in step with the real compile phase, and there
2258        have been a number of bugs (see for example, 4 above). I have now found a
2259        cunning way of running the real compile function in a "fake" mode that
2260        enables it to compute how much memory it would need, while actually only
2261        ever using a few hundred bytes of working memory and without too many
2262        tests of the mode. This should make future maintenance and development
2263        easier. A side effect of this work is that the limit of 200 on the nesting
2264        depth of parentheses has been removed (though this was never a serious
2265        limitation, I suspect). However, there is a downside: pcre_compile() now
2266        runs more slowly than before (30% or more, depending on the pattern). I
2267        hope this isn't a big issue. There is no effect on runtime performance.
2268    
2269    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2270        newline (only possible for the last line of a file) and it was a
2271        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2272    
2273    19. Added additional timing features to pcretest. (1) The -tm option now times
2274        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2275        separate command line item, by a number that specifies the number of
2276        repeats to use when timing. The default is 50000; this gives better
2277        precision, but takes uncomfortably long for very large patterns.
2278    
2279    20. Extended pcre_study() to be more clever in cases where a branch of a
2280        subpattern has no definite first character. For example, (a*|b*)[cd] would
2281        previously give no result from pcre_study(). Now it recognizes that the
2282        first character must be a, b, c, or d.
2283    
2284    21. There was an incorrect error "recursive call could loop indefinitely" if
2285        a subpattern (or the entire pattern) that was being tested for matching an
2286        empty string contained only one non-empty item after a nested subpattern.
2287        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2288        incorrectly, because the \d was being skipped in the check.
2289    
2290    22. The pcretest program now has a new pattern option /B and a command line
2291        option -b, which is equivalent to adding /B to every pattern. This causes
2292        it to show the compiled bytecode, without the additional information that
2293        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2294        is the same as /B/I).
2295    
2296    23. A new optimization is now able automatically to treat some sequences such
2297        as a*b as a*+b. More specifically, if something simple (such as a character
2298        or a simple class like \d) has an unlimited quantifier, and is followed by
2299        something that cannot possibly match the quantified thing, the quantifier
2300        is automatically "possessified".
2301    
2302    24. A recursive reference to a subpattern whose number was greater than 39
2303        went wrong under certain circumstances in UTF-8 mode. This bug could also
2304        have affected the operation of pcre_study().
2305    
2306    25. Realized that a little bit of performance could be had by replacing
2307        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2308    
2309    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2310    
2311    27. Possessive quantifiers such as a++ were previously implemented by turning
2312        them into atomic groups such as ($>a+). Now they have their own opcodes,
2313        which improves performance. This includes the automatically created ones
2314        from 23 above.
2315    
2316    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2317        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2318        the first matched character to be a colon. This applied both to named and
2319        numbered groups.
2320    
2321    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2322    
2323    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2324        building PCRE on Windows easier, so I have included it in the distribution.
2325    
2326    31. There is now a check in pcretest against a ridiculously large number being
2327        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2328        loop, the loop is abandoned.
2329    
2330    32. Forward references to subpatterns in conditions such as (?(2)...) where
2331        subpattern 2 is defined later cause pcre_compile() to search forwards in
2332        the pattern for the relevant set of parentheses. This search went wrong
2333        when there were unescaped parentheses in a character class, parentheses
2334        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2335    
2336    33. "Subroutine" calls and backreferences were previously restricted to
2337        referencing subpatterns earlier in the regex. This restriction has now
2338        been removed.
2339    
2340    34. Added a number of extra features that are going to be in Perl 5.10. On the
2341        whole, these are just syntactic alternatives for features that PCRE had
2342        previously implemented using the Python syntax or my own invention. The
2343        other formats are all retained for compatibility.
2344    
2345        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2346            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2347            also .NET compatible.
2348    
2349        (b) A recursion or subroutine call to a named group can now be defined as
2350            (?&name) as well as (?P>name).
2351    
2352        (c) A backreference to a named group can now be defined as \k<name> or
2353            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2354            5.10, are also .NET compatible.
2355    
2356        (d) A conditional reference to a named group can now use the syntax
2357            (?(<name>) or (?('name') as well as (?(name).
2358    
2359        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2360            groups (named and numbered) that are never evaluated inline, but can be
2361            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2362            is always false. There may be only one alternative in such a group.
2363    
2364        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2365            as the simple (?(R). The condition is true only if the most recent
2366            recursion is that of the given number or name. It does not search out
2367            through the entire recursion stack.
2368    
2369        (g) The escape \gN or \g{N} has been added, where N is a positive or
2370            negative number, specifying an absolute or relative reference.
2371    
2372    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2373        some "unreachable code" warnings.
2374    
2375    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2376        things, this adds five new scripts.
2377    
2378    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2379        There were also incompatibilities regarding the handling of \Q..\E inside
2380        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2381        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2382    
2383    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2384        matches an empty string, and forcibly breaks the loop. There were bugs in
2385        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2386        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2387        separate and independent bugs (that affected different cases) have been
2388        fixed.
2389    
2390    39. Refactored the code to abolish the use of different opcodes for small
2391        capturing bracket numbers. This is a tidy that I avoided doing when I
2392        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2393        The new approach is not only tidier, it makes it possible to reduce the
2394        memory needed to fix the previous bug (38).
2395    
2396    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2397        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2398        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2399        mode.
2400    
2401    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2402        report.
2403    
2404    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2405        copy construction and assignment in the C++ wrapper.
2406    
2407    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2408        couple of bugs that could have given wrong results in the "--newline=crlf"
2409        case.
2410    
2411    44. Added a number of casts and did some reorganization of signed/unsigned int
2412        variables following suggestions from Dair Grant. Also renamed the variable
2413        "this" as "item" because it is a C++ keyword.
2414    
2415    45. Arranged for dftables to add
2416    
2417          #include "pcre_internal.h"
2418    
2419        to pcre_chartables.c because without it, gcc 4.x may remove the array
2420        definition from the final binary if PCRE is built into a static library and
2421        dead code stripping is activated.
2422    
2423    46. For an unanchored pattern, if a match attempt fails at the start of a
2424        newline sequence, and the newline setting is CRLF or ANY, and the next two
2425        characters are CRLF, advance by two characters instead of one.
2426    
2427    
2428    Version 6.7 04-Jul-06
2429    ---------------------
2430    
2431     1. In order to handle tests when input lines are enormously long, pcretest has
2432        been re-factored so that it automatically extends its buffers when
2433        necessary. The code is crude, but this _is_ just a test program. The
2434        default size has been increased from 32K to 50K.
2435    
2436     2. The code in pcre_study() was using the value of the re argument before
2437        testing it for NULL. (Of course, in any sensible call of the function, it
2438        won't be NULL.)
2439    
2440     3. The memmove() emulation function in pcre_internal.h, which is used on
2441        systems that lack both memmove() and bcopy() - that is, hardly ever -
2442        was missing a "static" storage class specifier.
2443    
2444     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2445        containing an extended class (one that cannot be represented by a bitmap
2446        because it contains high-valued characters or Unicode property items, e.g.
2447        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2448        pattern, but PCRE should not loop if you do not (it no longer does).
2449        [Detail: two cases were found: (a) a repeated subpattern containing an
2450        extended class; (b) a recursive reference to a subpattern that followed a
2451        previous extended class. It wasn't skipping over the extended class
2452        correctly when UTF-8 mode was not set.]
2453    
2454     5. A negated single-character class was not being recognized as fixed-length
2455        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2456        compile error "lookbehind assertion is not fixed length".
2457    
2458     6. The RunPerlTest auxiliary script was showing an unexpected difference
2459        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2460        write a Perl script that can interpret lines of an input file either as
2461        byte characters or as UTF-8, which is what "perltest" was being required to
2462        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2463        can't do is switch easily at run time between having the "use utf8;" pragma
2464        or not. In the end, I fudged it by using the RunPerlTest script to insert
2465        "use utf8;" explicitly for the UTF-8 tests.
2466    
2467     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2468        the end of the subject string, contrary to the documentation and to what
2469        Perl does. This was true of both matching functions. Now it matches only at
2470        the start of the subject and immediately after *internal* newlines.
2471    
2472     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2473        a pointer to an int instead of a pointer to an unsigned long int. This
2474        caused problems on 64-bit systems.
2475    
2476     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2477        instance of the 'standard' template library not being so standard".
2478    
2479    10. There was no check on the number of named subpatterns nor the maximum
2480        length of a subpattern name. The product of these values is used to compute
2481        the size of the memory block for a compiled pattern. By supplying a very
2482        long subpattern name and a large number of named subpatterns, the size
2483        computation could be caused to overflow. This is now prevented by limiting
2484        the length of names to 32 characters, and the number of named subpatterns
2485        to 10,000.
2486    
2487    11. Subpatterns that are repeated with specific counts have to be replicated in
2488        the compiled pattern. The size of memory for this was computed from the
2489        length of the subpattern and the repeat count. The latter is limited to
2490        65535, but there was no limit on the former, meaning that integer overflow
2491        could in principle occur. The compiled length of a repeated subpattern is
2492        now limited to 30,000 bytes in order to prevent this.
2493    
2494    12. Added the optional facility to have named substrings with the same name.
2495    
2496    13. Added the ability to use a named substring as a condition, using the
2497        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2498        are numbers (not recommended). Forward references are permitted.
2499    
2500    14. Added forward references in named backreferences (if you see what I mean).
2501    
2502    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2503        pattern could run off the end of the subject. For example, the pattern
2504        "(?s)(.{1,5})"8 did this with the subject "ab".
2505    
2506    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2507        PCRE_CASELESS was set when matching characters that were quantified with ?
2508        or *.
2509    
2510    17. A character class other than a single negated character that had a minimum
2511        but no maximum quantifier - for example [ab]{6,} - was not handled
2512        correctly by pce_dfa_exec(). It would match only one character.
2513    
2514    18. A valid (though odd) pattern that looked like a POSIX character
2515        class but used an invalid character after [ (for example [[,abc,]]) caused
2516        pcre_compile() to give the error "Failed: internal error: code overflow" or
2517        in some cases to crash with a glibc free() error. This could even happen if
2518        the pattern terminated after [[ but there just happened to be a sequence of
2519        letters, a binary zero, and a closing ] in the memory that followed.
2520    
2521    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2522        over the years. Originally (before any Unicode support), just the bottom 8
2523        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2524        output from "man perlunicode" includes this:
2525    
2526          The regular expression compiler produces polymorphic opcodes.  That
2527          is, the pattern adapts to the data and automatically switches to
2528          the Unicode character scheme when presented with Unicode data--or
2529          instead uses a traditional byte scheme when presented with byte
2530          data.
2531    
2532        Sadly, a wide octal escape does not cause a switch, and in a string with
2533        no other multibyte characters, these octal escapes are treated as before.
2534        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2535        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2536        Unicode string.
2537    
2538        I have not perpetrated such confusion in PCRE. Up till now, it took just
2539        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2540        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2541        translate to the appropriate multibyte character.
2542    
2543    29. Applied some refactoring to reduce the number of warnings from Microsoft
2544        and Borland compilers. This has included removing the fudge introduced
2545        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2546        a warning about an unused variable.
2547    
2548    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2549        characters since release 4.0, because Perl (from release 5.004) does not.
2550        [Or at least, is documented not to: some releases seem to be in conflict
2551        with the documentation.] However, when a pattern was studied with
2552        pcre_study() and all its branches started with \s, PCRE still included VT
2553        as a possible starting character. Of course, this did no harm; it just
2554        caused an unnecessary match attempt.
2555    
2556    22. Removed a now-redundant internal flag bit that recorded the fact that case
2557        dependency changed within the pattern. This was once needed for "required
2558        byte" processing, but is no longer used. This recovers a now-scarce options
2559        bit. Also moved the least significant internal flag bit to the most-
2560        significant bit of the word, which was not previously used (hangover from
2561        the days when it was an int rather than a uint) to free up another bit for
2562        the future.
2563    
2564    23. Added support for CRLF line endings as well as CR and LF. As well as the
2565        default being selectable at build time, it can now be changed at runtime
2566        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2567        specify that it is scanning data with non-default line endings.
2568    
2569    24. Changed the definition of CXXLINK to make it agree with the definition of
2570        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2571    
2572    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2573        recursions. This makes a big different to stack usage for some patterns.
2574    
2575    26. If a subpattern containing a named recursion or subroutine reference such
2576        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2577        the space required for the compiled pattern went wrong and gave too small a
2578        value. Depending on the environment, this could lead to "Failed: internal
2579        error: code overflow at offset 49" or "glibc detected double free or
2580        corruption" errors.
2581    
2582    27. Applied patches from Google (a) to support the new newline modes and (b) to
2583        advance over multibyte UTF-8 characters in GlobalReplace.
2584    
2585    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2586        difference for some implementation of PCRE in some Windows version.
2587    
2588    29. Added some extra testing facilities to pcretest:
2589    
2590        \q<number>   in a data line sets the "match limit" value
2591        \Q<number>   in a data line sets the "match recursion limt" value
2592        -S <number>  sets the stack size, where <number> is in megabytes
2593    
2594        The -S option isn't available for Windows.
2595    
2596    
2597    Version 6.6 06-Feb-06
2598    ---------------------
2599    
2600     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2601        in pcreposix.h. I have copied the definition from pcre.h.
2602    
2603     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2604        because pcre.h is no longer a built file.
2605    
2606     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2607        not normally included in the compiled code.
2608    
2609    
2610    Version 6.5 01-Feb-06
2611    ---------------------
2612    
2613     1. When using the partial match feature with pcre_dfa_exec(), it was not
2614        anchoring the second and subsequent partial matches at the new starting
2615        point. This could lead to incorrect results. For example, with the pattern
2616        /1234/, partially matching against "123" and then "a4" gave a match.
2617    
2618     2. Changes to pcregrep:
2619    
2620        (a) All non-match returns from pcre_exec() were being treated as failures
2621            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2622            error message is output. Some extra information is given for the
2623            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2624            probably the only errors that are likely to be caused by users (by
2625            specifying a regex that has nested indefinite repeats, for instance).
2626            If there are more than 20 of these errors, pcregrep is abandoned.
2627    
2628        (b) A binary zero was treated as data while matching, but terminated the
2629            output line if it was written out. This has been fixed: binary zeroes
2630            are now no different to any other data bytes.
2631    
2632        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2633            used to set a locale for matching. The --locale=xxxx long option has
2634            been added (no short equivalent) to specify a locale explicitly on the
2635            pcregrep command, overriding the environment variables.
2636    
2637        (d) When -B was used with -n, some line numbers in the output were one less
2638            than they should have been.
2639    
2640        (e) Added the -o (--only-matching) option.
2641    
2642        (f) If -A or -C was used with -c (count only), some lines of context were
2643            accidentally printed for the final match.
2644    
2645        (g) Added the -H (--with-filename) option.
2646    
2647        (h) The combination of options -rh failed to suppress file names for files
2648            that were found from directory arguments.
2649    
2650        (i) Added the -D (--devices) and -d (--directories) options.
2651    
2652        (j) Added the -F (--fixed-strings) option.
2653    
2654        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2655    
2656        (l) Added the --colo(u)r option.
2657    
2658        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2659            is not present by default.
2660    
2661     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2662        items such as (?R) or (?1), when the recursion could match a number of
2663        alternatives. If it matched one of the alternatives, but subsequently,
2664        outside the recursion, there was a failure, the code tried to back up into
2665        the recursion. However, because of the way PCRE is implemented, this is not
2666        possible, and the result was an incorrect result from the match.
2667    
2668        In order to prevent this happening, the specification of recursion has
2669        been changed so that all such subpatterns are automatically treated as
2670        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2671    
2672     4. I had overlooked the fact that, in some locales, there are characters for
2673        which isalpha() is true but neither isupper() nor islower() are true. In
2674        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2675        and ordfeminine) are like this. This affected the treatment of \w and \W
2676        when they appeared in character classes, but not when they appeared outside
2677        a character class. The bit map for "word" characters is now created
2678        separately from the results of isalnum() instead of just taking it from the
2679        upper, lower, and digit maps. (Plus the underscore character, of course.)
2680    
2681     5. The above bug also affected the handling of POSIX character classes such as
2682        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2683        permanent tables. Instead, the bit maps for such a class were previously
2684        created as the appropriate unions of the upper, lower, and digit bitmaps.
2685        Now they are created by subtraction from the [[:word:]] class, which has
2686        its own bitmap.
2687    
2688     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2689        It is created by subtracting the vertical space characters (\x09, \x0a,
2690        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2691        subtraction was done in the overall bitmap for a character class, meaning
2692        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2693        be recognized. This bug has been fixed.
2694    
2695     7. Patches from the folks at Google:
2696    
2697          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2698          real life, but is still worth protecting against".
2699    
2700          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2701          regular expressions".
2702    
2703          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2704          have it.
2705    
2706          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2707          "configure" and the latter not, in order to fix a problem somebody had
2708          with compiling the Arg class on HP-UX.
2709    
2710          (e) Improve the error-handling of the C++ wrapper a little bit.
2711    
2712          (f) New tests for checking recursion limiting.
2713    
2714     8. The pcre_memmove() function, which is used only if the environment does not
2715        have a standard memmove() function (and is therefore rarely compiled),
2716        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2717        returning a result (though PCRE never actually uses the result).
2718    
2719     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2720        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2721        returned instead of calling malloc() with an overflowing number that would
2722        most likely cause subsequent chaos.
2723    
2724    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2725    
2726    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2727        with this option is matched, the nmatch and pmatch options of regexec() are
2728        ignored.
2729    
2730    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2731        provided in case anyone wants to the the POSIX interface with UTF-8
2732        strings.
2733    
2734    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2735        C++ linking (needed for some HP-UX environments).
2736    
2737    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2738        (unused parameter) and in the pcre_printint() function (omitted "default"
2739        switch label when the default is to do nothing).
2740    
2741    15. Added some code to make it possible, when PCRE is compiled as a C++
2742        library, to replace subject pointers for pcre_exec() with a smart pointer
2743        class, thus making it possible to process discontinuous strings.
2744    
2745    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2746        much the same function. They were added by different people who were trying
2747        to make PCRE easy to compile on non-Unix systems. It has been suggested
2748        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2749        for compiling on Windows systems. I have therefore replaced it with
2750        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2751        defaults to "extern" for C or "extern C" for C++, which works fine on
2752        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2753        SCOPE with something explicit in config.h. In addition:
2754    
2755        (a) pcreposix.h still had just "extern" instead of either of these macros;
2756            I have replaced it with PCRE_DATA_SCOPE.
2757    
2758        (b) Functions such as _pcre_xclass(), which are internal to the library,
2759            but external in the C sense, all had PCRE_EXPORT in their definitions.
2760            This is apparently wrong for the Windows case, so I have removed it.
2761            (It makes no difference on Unix-like systems.)
2762    
2763    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2764        of recursive calls to match(). This is different to MATCH_LIMIT because
2765        that limits the total number of calls to match(), not all of which increase
2766        the depth of recursion. Limiting the recursion depth limits the amount of
2767        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2768        when PCRE is compiled, and changed at run time. A patch from Google adds
2769        this functionality to the C++ interface.
2770    
2771    18. Changes to the handling of Unicode character properties:
2772    
2773        (a) Updated the table to Unicode 4.1.0.
2774    
2775        (b) Recognize characters that are not in the table as "Cn" (undefined).
2776    
2777        (c) I revised the way the table is implemented to a much improved format
2778            which includes recognition of ranges. It now supports the ranges that
2779            are defined in UnicodeData.txt, and it also amalgamates other
2780            characters into ranges. This has reduced the number of entries in the
2781            table from around 16,000 to around 3,000, thus reducing its size
2782            considerably. I realized I did not need to use a tree structure after
2783            all - a binary chop search is just as efficient. Having reduced the
2784            number of entries, I extended their size from 6 bytes to 8 bytes to
2785            allow for more data.
2786    
2787        (d) Added support for Unicode script names via properties such as \p{Han}.
2788    
2789    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2790        matching that character.
2791    
2792    20. When matching a repeated Unicode property with a minimum greater than zero,
2793        (for example \pL{2,}), PCRE could look past the end of the subject if it
2794        reached it while seeking the minimum number of characters. This could
2795        happen only if some of the characters were more than one byte long, because
2796        there is a check for at least the minimum number of bytes.
2797    
2798    21. Refactored the implementation of \p and \P so as to be more general, to
2799        allow for more different types of property in future. This has changed the
2800        compiled form incompatibly. Anybody with saved compiled patterns that use
2801        \p or \P will have to recompile them.
2802    
2803    22. Added "Any" and "L&" to the supported property types.
2804    
2805    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2806        but give a compile time error if the value is greater than 0xff.
2807    
2808    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2809        accidentally not being installed or uninstalled.
2810    
2811    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2812        made were to insert the current release number. This seemed silly, because
2813        it made things harder for people building PCRE on systems that don't run
2814        "configure". I have turned pcre.h into a distributed file, no longer built
2815        by "configure", with the version identification directly included. There is
2816        no longer a pcre.h.in file.
2817    
2818        However, this change necessitated a change to the pcre-config script as
2819        well. It is built from pcre-config.in, and one of the substitutions was the
2820        release number. I have updated configure.ac so that ./configure now finds
2821        the release number by grepping pcre.h.
2822    
2823    26. Added the ability to run the tests under valgrind.
2824    
2825    
2826    Version 6.4 05-Sep-05
2827    ---------------------
2828    
2829     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2830        "--" to be printed when multiple files were scanned, even when none of the
2831        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2832        consider it to be a bug, and have restored the previous behaviour.
2833    
2834     2. A couple of code tidies to get rid of compiler warnings.
2835    
2836     3. The pcretest program used to cheat by referring to symbols in the library
2837        whose names begin with _pcre_. These are internal symbols that are not
2838        really supposed to be visible externally, and in some environments it is
2839        possible to suppress them. The cheating is now confined to including
2840        certain files from the library's source, which is a bit cleaner.
2841    
2842     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2843        file's purpose clearer.
2844    
2845     5. Reorganized pcre_ucp_findchar().
2846    
2847    
2848    Version 6.3 15-Aug-05
2849    ---------------------
2850    
2851     1. The file libpcre.pc.in did not have general read permission in the tarball.
2852    
2853     2. There were some problems when building without C++ support:
2854    
2855        (a) If C++ support was not built, "make install" and "make test" still
2856            tried to test it.
2857    
2858        (b) There were problems when the value of CXX was explicitly set. Some
2859            changes have been made to try to fix these, and ...
2860    
2861        (c) --disable-cpp can now be used to explicitly disable C++ support.
2862    
2863        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2864            backslash in a target when C++ was disabled. This confuses some
2865            versions of "make", apparently. Using an intermediate variable solves
2866            this. (Same for CPP_LOBJ.)
2867    
2868     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2869        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2870        necessary on certain architectures.
2871    
2872     4. Added a setting of -export-symbols-regex to the link command to remove
2873        those symbols that are exported in the C sense, but actually are local
2874        within the library, and not documented. Their names all begin with
2875        "_pcre_". This is not a perfect job, because (a) we have to except some
2876        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2877        available (and never for static libraries). I have made a note to try to
2878        find a way round (a) in the future.
2879    
2880    
2881    Version 6.2 01-Aug-05
2882    ---------------------
2883    
2884     1. There was no test for integer overflow of quantifier values. A construction
2885        such as {1111111111111111} would give undefined results. What is worse, if
2886        a minimum quantifier for a parenthesized subpattern overflowed and became
2887        negative, the calculation of the memory size went wrong. This could have
2888        led to memory overwriting.
2889    
2890     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2891    
2892     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2893        operating environments where this matters.
2894    
2895     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2896        PCRE options from within the C++ wrapper.
2897    
2898     5. Named capturing subpatterns were not being correctly counted when a pattern
2899        was compiled. This caused two problems: (a) If there were more than 100
2900        such subpatterns, the calculation of the memory needed for the whole
2901        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2902        back references of the form \12, where the number was greater than 9, were
2903        not recognized as back references, even though there were sufficient
2904        previous subpatterns.
2905    
2906     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2907        versions of gcc, e.g. 2.95.4.
2908    
2909    
2910    Version 6.1 21-Jun-05
2911    ---------------------
2912    
2913     1. There was one reference to the variable "posix" in pcretest.c that was not
2914        surrounded by "#if !defined NOPOSIX".
2915    
2916     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2917        the cross-check on the old pcre_info() function, for the benefit of the
2918        cut-down version of PCRE that is currently imported into Exim.
2919    
2920     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2921        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2922        patterns that start (?i) but I don't think that matters. The use of (?i) is
2923        just an example; this all applies to the other options as well.
2924    
2925     4. Since libtool seems to echo the compile commands it is issuing, the output
2926        from "make" can be reduced a bit by putting "@" in front of each libtool
2927        compile command.
2928    
2929     5. Patch from the folks at Google for configure.in to be a bit more thorough
2930        in checking for a suitable C++ installation before trying to compile the
2931        C++ stuff. This should fix a reported problem when a compiler was present,
2932        but no suitable headers.
2933    
2934     6. The man pages all had just "PCRE" as their title. I have changed them to
2935        be the relevant file name. I have also arranged that these names are
2936        retained in the file doc/pcre.txt, which is a concatenation in text format
2937        of all the man pages except the little individual ones for each function.
2938    
2939     7. The NON-UNIX-USE file had not been updated for the different set of source
2940        files that come with release 6. I also added a few comments about the C++
2941        wrapper.
2942    
2943    
2944    Version 6.0 07-Jun-05
2945    ---------------------
2946    
2947     1. Some minor internal re-organization to help with my DFA experiments.
2948    
2949     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2950        didn't matter for the library itself when fully configured, but did matter
2951        when compiling without UCP support, or within Exim, where the ucp files are
2952        not imported.
2953    
2954     3. Refactoring of the library code to split up the various functions into
2955        different source modules. The addition of the new DFA matching code (see
2956        below) to a single monolithic source would have made it really too
2957        unwieldy, quite apart from causing all the code to be include in a
2958        statically linked application, when only some functions are used. This is
2959        relevant even without the DFA addition now that patterns can be compiled in
2960        one application and matched in another.
2961    
2962        The downside of splitting up is that there have to be some external
2963        functions and data tables that are used internally in different modules of
2964        the library but which are not part of the API. These have all had their
2965        names changed to start with "_pcre_" so that they are unlikely to clash
2966        with other external names.
2967    
2968     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2969        a different (DFA) algorithm. Although it is slower than the original
2970        function, it does have some advantages for certain types of matching
2971        problem.
2972    
2973     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2974        including restarting after a partial match.
2975    
2976     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2977        defined when compiling for Windows was sent to me. I have put it into the
2978        code, though I have no means of testing or verifying it.
2979    
2980     7. Added the pcre_refcount() auxiliary function.
2981    
2982     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2983        match before or at the first newline in the subject string. In pcretest,
2984        the /f option on a pattern can be used to set this.
2985    
2986     9. A repeated \w when used in UTF-8 mode with characters greater than 256
2987        would behave wrongly. This has been present in PCRE since release 4.0.
2988    
2989    10. A number of changes to the pcregrep command:
2990    
2991        (a) Refactored how -x works; insert ^(...)$ instead of setting
2992            PCRE_ANCHORED and checking the length, in preparation for adding
2993            something similar for -w.
2994    
2995        (b) Added the -w (match as a word) option.
2996    
2997        (c) Refactored the way lines are read and buffered so as to have more
2998            than one at a time available.
2999    
3000        (d) Implemented a pcregrep test script.
3001    
3002        (e) Added the -M (multiline match) option. This allows patterns to match
3003            over several lines of the subject. The buffering ensures that at least
3004            8K, or the rest of the document (whichever is the shorter) is available
3005            for matching (and similarly the previous 8K for lookbehind assertions).
3006    
3007        (f) Changed the --help output so that it now says
3008    
3009              -w, --word-regex(p)
3010    
3011            instead of two lines, one with "regex" and the other with "regexp"
3012            because that confused at least one person since the short forms are the
3013            same. (This required a bit of code, as the output is generated
3014            automatically from a table. It wasn't just a text change.)
3015    
3016        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3017            option but starts with a hyphen. Could be a pattern or a path name
3018            starting with a hyphen, for instance.
3019    
3020        (h) "-" can be given as a file name to represent stdin.
3021    
3022        (i) When file names are being printed, "(standard input)" is used for
3023            the standard input, for compatibility with GNU grep. Previously
3024            "<stdin>" was used.
3025    
3026        (j) The option --label=xxx can be used to supply a name to be used for
3027            stdin when file names are being printed. There is no short form.
3028    
3029        (k) Re-factored the options decoding logic because we are going to add
3030            two more options that take data. Such options can now be given in four
3031            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3032    
3033        (l) Added the -A, -B, and -C options for requesting that lines of context
3034            around matches be printed.
3035    
3036        (m) Added the -L option to print the names of files that do not contain
3037            any matching lines, that is, the complement of -l.
3038    
3039        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3040            continue to scan other files.
3041    
3042        (o) The -s option was incorrectly implemented. For compatibility with other
3043            greps, it now suppresses the error message for a non-existent or non-
3044            accessible file (but not the return code). There is a new option called
3045            -q that suppresses the output of matching lines, which was what -s was
3046            previously doing.
3047    
3048        (p) Added --include and --exclude options to specify files for inclusion
3049            and exclusion when recursing.
3050    
3051    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3052        Hopefully, it now does.
3053    
3054    12. Missing cast in pcre_study().
3055    
3056    13. Added an "uninstall" target to the makefile.
3057    
3058    14. Replaced "extern" in the function prototypes in Makefile.in with
3059        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3060        world, but is set differently for Windows.
3061    
3062    15. Added a second compiling function called pcre_compile2(). The only
3063        difference is that it has an extra argument, which is a pointer to an
3064        integer error code. When there is a compile-time failure, this is set
3065        non-zero, in addition to the error test pointer being set to point to an
3066        error message. The new argument may be NULL if no error number is required
3067        (but then you may as well call pcre_compile(), which is now just a
3068        wrapper). This facility is provided because some applications need a
3069        numeric error indication, but it has also enabled me to tidy up the way
3070        compile-time errors are handled in the POSIX wrapper.
3071    
3072    16. Added VPATH=.libs to the makefile; this should help when building with one
3073        prefix path and installing with another. (Or so I'm told by someone who
3074        knows more about this stuff than I do.)
3075    
3076    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3077        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3078        match everything, including newlines. This is not POSIX-compatible, but
3079        somebody wanted the feature. From pcretest it can be activated by using
3080        both the P and the s flags.
3081    
3082    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3083    
3084    19. libpcre.pc was being incorrectly installed as executable.
3085    
3086    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3087        it now also looks for '\r' so that it will work unmodified on Windows.
3088    
3089    21. Added Google's contributed C++ wrapper to the distribution.
3090    
3091    22. Added some untidy missing memory free() calls in pcretest, to keep
3092        Electric Fence happy when testing.
3093    
3094    
3095    
3096    Version 5.0 13-Sep-04
3097    ---------------------
3098    
3099     1. Internal change: literal characters are no longer packed up into items
3100        containing multiple characters in a single byte-string. Each character
3101        is now matched using a separate opcode. However, there may be more than one
3102        byte in the character in UTF-8 mode.
3103    
3104     2. The pcre_callout_block structure has two new fields: pattern_position and
3105        next_item_length. These contain the offset in the pattern to the next match
3106        item, and its length, respectively.
3107    
3108     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3109        insertion of callouts before each pattern item. Added the /C option to
3110        pcretest to make use of this.
3111    
3112     4. On the advice of a Windows user, the lines
3113    
3114          #if defined(_WIN32) || defined(WIN32)
3115          _setmode( _fileno( stdout ), 0x8000 );
3116          #endif  /* defined(_WIN32) || defined(WIN32) */
3117    
3118        have been added to the source of pcretest. This apparently does useful
3119        magic in relation to line terminators.
3120    
3121     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3122        for the benefit of those environments where the "b" makes a difference.
3123    
3124     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3125        to know about it. I have put a hack into configure.in that adds in code
3126        to set GCC=yes if CC=icc. This seems to end up at a point in the
3127        generated configure script that is early enough to affect the setting of
3128        compiler options, which is what is needed, but I have no means of testing
3129        whether it really works. (The user who reported this had patched the
3130        generated configure script, which of course I cannot do.)
3131    
3132        LATER: After change 22 below (new libtool files), the configure script
3133        seems to know about icc (and also ecc). Therefore, I have commented out
3134        this hack in configure.in.
3135    
3136     7. Added support for pkg-config (2 patches were sent in).
3137    
3138     8. Negated POSIX character classes that used a combination of internal tables
3139        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3140        [[:^ascii]]. Typically, they would match almost any characters. The other
3141        POSIX classes were not broken in this way.
3142    
3143     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3144        to find the match, as PCRE was deluded into thinking that the match had to
3145        start at the start point or following a newline. The same bug applied to
3146        patterns with negative forward assertions or any backward assertions
3147        preceding ".*" at the start, unless the pattern required a fixed first
3148        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3149    
3150    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3151        starting at the last subject character, bytes beyond the end of the subject
3152        string were read.
3153    
3154    11. Renamed the variable "class" as "classbits" to make life easier for C++
3155        users. (Previously there was a macro definition, but it apparently wasn't
3156        enough.)
3157    
3158    12. Added the new field "tables" to the extra data so that tables can be passed
3159        in at exec time, or the internal tables can be re-selected. This allows
3160        a compiled regex to be saved and re-used at a later time by a different
3161        program that might have everything at different addresses.
3162    
3163    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3164        -R library as well as a -L library.
3165    
3166    14. The debugging options of pcretest (-d on the command line or D on a
3167        pattern) showed incorrect output for anything following an extended class
3168        that contained multibyte characters and which was followed by a quantifier.
3169    
3170    15. Added optional support for general category Unicode character properties
3171        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3172        support. It adds about 90K to the size of the library. The meanings of the
3173        inbuilt class escapes such as \d and \s have NOT been changed.
3174    
3175    16. Updated pcredemo.c to include calls to free() to release the memory for the
3176        compiled pattern.
3177    
3178    17. The generated file chartables.c was being created in the source directory
3179        instead of in the building directory. This caused the build to fail if the
3180        source directory was different from the building directory, and was
3181        read-only.
3182    
3183    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3184        file. No doubt somebody will tell me if they don't make sense... Also added
3185        Dan Mooney's comments about building on OpenVMS.
3186    
3187    19. Added support for partial matching via the PCRE_PARTIAL option for
3188        pcre_exec() and the \P data escape in pcretest.
3189    
3190    20. Extended pcretest with 3 new pattern features:
3191    
3192        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3193              write the compiled pattern to the file whose name is "rest-of-line".
3194              This is a straight binary dump of the data, with the saved pointer to
3195              the character tables forced to be NULL. The study data, if any, is
3196              written too. After writing, pcretest reads a new pattern.
3197    
3198        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3199              compiled pattern from the given file. There must not be any
3200              occurrences of "<" in the file name (pretty unlikely); if there are,
3201              pcretest will instead treat the initial "<" as a pattern delimiter.
3202              After reading in the pattern, pcretest goes on to read data lines as
3203              usual.
3204    
3205        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3206              and 16-bit fields in a compiled pattern, to simulate a pattern that
3207              was compiled on a host of opposite endianness.
3208    
3209    21. The pcre-exec() function can now cope with patterns that were compiled on
3210        hosts of opposite endianness, with this restriction:
3211    
3212          As for any compiled expression that is saved and used later, the tables
3213          pointer field cannot be preserved; the extra_data field in the arguments
3214          to pcre_exec() should be used to pass in a tables address if a value
3215          other than the default internal tables were used at compile time.
3216    
3217    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3218        now diagnosed as an error. Previously, most of the time, a negative number
3219        would have been treated as zero, but if in addition "ovector" was passed as
3220        NULL, a crash could occur.
3221    
3222    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3223        new versions from the libtool 1.5 distribution (the last one is a copy of
3224        a file called libtool.m4). This seems to have fixed the need to patch
3225        "configure" to support Darwin 1.3 (which I used to do). However, I still
3226        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3227        workstation).
3228    
3229    24. Changed the PCRE licence to be the more standard "BSD" licence.
3230    
3231    
3232    Version 4.5 01-Dec-03
3233    ---------------------
3234    
3235     1. There has been some re-arrangement of the code for the match() function so
3236        that it can be compiled in a version that does not call itself recursively.
3237        Instead, it keeps those local variables that need separate instances for
3238        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3239        needs to "recurse". Keeping track of where control must go is done by means
3240        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3241        hide most of the details from the main code, and operates only if
3242        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3243        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3244        operating.
3245    
3246        To make it easier for callers to provide specially tailored get/free
3247        functions for this usage, two new functions, pcre_stack_malloc, and
3248        pcre_stack_free, are used. They are always called in strict stacking order,
3249        and the size of block requested is always the same.
3250    
3251        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3252        PCRE has been compiled to use the stack or the heap for recursion. The
3253        -C option of pcretest uses this to show which version is compiled.
3254    
3255        A new data escape \S, is added to pcretest; it causes the amounts of store
3256        obtained and freed by both kinds of malloc/free at match time to be added
3257        to the output.
3258    
3259     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3260        what's available on my current Linux desktop machine.
3261    
3262     3. When matching a UTF-8 string, the test for a valid string at the start has
3263        been extended. If start_offset is not zero, PCRE now checks that it points
3264        to a byte that is the start of a UTF-8 character. If not, it returns
3265        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3266        this is necessary because there may be backward assertions in the pattern.
3267        When matching the same subject several times, it may save resources to use
3268        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3269    
3270     4. The code for checking the validity of UTF-8 strings has been tightened so
3271        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3272        containing "overlong sequences".
3273    
3274     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3275        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3276        should have been "&", but it just so happened that all the cases this let
3277        through by mistake were picked up later in the function.
3278    
3279     6. I had used a variable called "isblank" - this is a C99 function, causing
3280        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3281    
3282     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3283        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3284        so the version is shown; (c) stop "make test" echoing "./RunTest".
3285    
3286     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3287    
3288     9. The prototype for memmove() for systems that don't have it was using
3289        size_t, but the inclusion of the header that defines size_t was later. I've
3290        moved the #includes for the C headers earlier to avoid this.
3291    
3292    10. Added some adjustments to the code to make it easier to compiler on certain
3293        special systems:
3294    
3295          (a) Some "const" qualifiers were missing.
3296          (b) Added the macro EXPORT before all exported functions; by default this
3297              is defined to be empty.
3298          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3299              that it reads its output file name as an argument instead of writing
3300              to the standard output and assuming this can be redirected.
3301    
3302    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3303        class containing characters with values greater than 255, PCRE compilation
3304        went into a loop.
3305    
3306    12. A recursive reference to a subpattern that was within another subpattern
3307        that had a minimum quantifier of zero caused PCRE to crash. For example,
3308        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3309        recursion. If the recursively-called subpattern itself had a zero repeat,
3310        that was OK.
3311    
3312    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3313        buffer into which it was copied (for escape processing) was still set at
3314        1024, so long lines caused crashes.
3315    
3316    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3317        "internal error: code overflow...". This applied to any character class
3318        that was followed by a possessive quantifier.
3319    
3320    15. Modified the Makefile to add libpcre.la as a prerequisite for
3321        libpcreposix.la because I was told this is needed for a parallel build to
3322        work.
3323    
3324    16. If a pattern that contained .* following optional items at the start was
3325        studied, the wrong optimizing data was generated, leading to matching
3326        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3327        matching string must start with a or b or c. The correct conclusion for
3328        this pattern is that a match can start with any character.
3329    
3330    
3331    Version 4.4 13-Aug-03
3332    ---------------------
3333    
3334     1. In UTF-8 mode, a character class containing characters with values between
3335        127 and 255 was not handled correctly if the compiled pattern was studied.
3336        In fixing this, I have also improved the studying algorithm for such
3337        classes (slightly).
3338    
3339     2. Three internal functions had redundant arguments passed to them. Removal
3340        might give a very teeny performance improvement.
3341    
3342     3. Documentation bug: the value of the capture_top field in a callout is *one
3343        more than* the number of the hightest numbered captured substring.
3344    
3345     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3346        in incorrectly linking with a previously installed version. They now link
3347        explicitly with libpcre.la.
3348    
3349     5. configure.in no longer needs to recognize Cygwin specially.
3350    
3351     6. A problem in pcre.in for Windows platforms is fixed.
3352    
3353     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3354        pcretest, it used to include the size of the study block as part of its
3355        output. Unfortunately, the structure contains a field that has a different
3356        size on different hardware architectures. This meant that the tests that
3357        showed this size failed. As the block is currently always of a fixed size,
3358        this information isn't actually particularly useful in pcretest output, so
3359        I have just removed it.
3360    
3361     8. Three pre-processor statements accidentally did not start in column 1.
3362        Sadly, there are *still* compilers around that complain, even though
3363        standard C has not required this for well over a decade. Sigh.
3364    
3365     9. In pcretest, the code for checking callouts passed small integers in the
3366        callout_data field, which is a void * field. However, some picky compilers
3367        complained about the casts involved for this on 64-bit systems. Now
3368        pcretest passes the address of the small integer instead, which should get
3369        rid of the warnings.
3370    
3371    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3372        both compile and run time, and gives an error if an invalid UTF-8 sequence
3373        is found. There is a option for disabling this check in cases where the
3374        string is known to be correct and/or the maximum performance is wanted.
3375    
3376    11. In response to a bug report, I changed one line in Makefile.in from
3377    
3378            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3379        to
3380            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3381    
3382        to look similar to other lines, but I have no way of telling whether this
3383        is the right thing to do, as I do not use Windows. No doubt I'll get told
3384        if it's wrong...
3385    
3386    
3387    Version 4.3 21-May-03
3388    ---------------------
3389    
3390    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3391       Makefile.
3392    
3393    2. Some refactoring to improve the quality of the code:
3394    
3395       (i)   The utf8_table... variables are now declared "const".
3396    
3397       (ii)  The code for \cx, which used the "case flipping" table to upper case
3398             lower case letters, now just substracts 32. This is ASCII-specific,
3399             but the whole concept of \cx is ASCII-specific, so it seems
3400             reasonable.
3401    
3402       (iii) PCRE was using its character types table to recognize decimal and
3403             hexadecimal digits in the pattern. This is silly, because it handles
3404             only 0-9, a-f, and A-F, but the character types table is locale-
3405             specific, which means strange things might happen. A private
3406             table is now used for this - though it costs 256 bytes, a table is
3407             much faster than multiple explicit tests. Of course, the standard
3408             character types table is still used for matching digits in subject
3409             strings against \d.
3410    
3411       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3412             ending in _t are). So I've renamed it as ESC_tee.
3413    
3414    3. The first argument for regexec() in the POSIX wrapper should have been
3415       defined as "const".
3416    
3417    4. Changed pcretest to use malloc() for its buffers so that they can be
3418       Electric Fenced for debugging.
3419    
3420    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3421       to read one or more bytes before the start of the subject string. Often this
3422       had no effect on PCRE's behaviour, but in some circumstances it could
3423       provoke a segmentation fault.
3424    
3425    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3426       to try to read one or more bytes before the start of the subject string.
3427    
3428    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3429       UTF-8 support could misbehave in various ways if the subject string
3430       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3431       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3432       back over UTF-8 characters.)
3433    
3434    
3435    Version 4.2 14-Apr-03
3436    ---------------------
3437    
3438    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3439    
3440    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3441         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3442         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3443         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3444         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3445           and BUILD_EXEEXT
3446         Note: automatic setting of the BUILD variables is not yet working
3447         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3448           compile-time but not at link-time
3449         [LINK]: use for linking executables only
3450         make different versions for Windows and non-Windows
3451         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3452           libraries
3453         [LINK_FOR_BUILD]: new variable
3454         [OBJEXT]: use throughout
3455         [EXEEXT]: use throughout
3456         <winshared>: new target
3457         <wininstall>: new target
3458         <dftables.o>: use native compiler
3459         <dftables>: use native linker
3460         <install>: handle Windows platform correctly
3461         <clean>: ditto
3462         <check>: ditto
3463         copy DLL to top builddir before testing
3464    
3465       As part of these changes, -no-undefined was removed again. This was reported
3466       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3467       in any case.
3468    
3469    3. Some tidies to get rid of compiler warnings:
3470    
3471       . In the match_data structure, match_limit was an unsigned long int, whereas
3472         match_call_count was an int. I've made them both unsigned long ints.
3473    
3474       . In pcretest the fact that a const uschar * doesn't automatically cast to
3475         a void * provoked a warning.
3476    
3477       . Turning on some more compiler warnings threw up some "shadow" variables
3478         and a few more missing casts.
3479    
3480    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3481       option, a class that contained a single character with a value between 128
3482       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3483    
3484    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3485       option, a class that contained several characters, but with at least one
3486       whose value was between 128 and 255 caused PCRE to crash.
3487    
3488    
3489    Version 4.1 12-Mar-03
3490    ---------------------
3491    
3492    1. Compiling with gcc -pedantic found a couple of places where casts were
3493    needed, and a string in dftables.c that was longer than standard compilers are
3494    required to support.
3495    
3496    2. Compiling with Sun's compiler found a few more places where the code could
3497    be tidied up in order to avoid warnings.
3498    
3499    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3500    first of these names is deprecated in the latest Autoconf in favour of the name
3501    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3502    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3503    analogy I have changed it to CFLAGS_FOR_BUILD.
3504    
3505    4. Added -no-undefined to the linking command in the Makefile, because this is
3506    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3507    linking step for the pcreposix library.
3508    
3509    5. PCRE was failing to diagnose the case of two named groups with the same
3510    name.
3511    
3512    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3513    literal character that is needed in the subject for a match, and scans along to
3514    ensure that it is present before embarking on the full matching process. This
3515    saves time in cases of nested unlimited repeats that are never going to match.
3516    Problem: the scan can take a lot of time if the subject is very long (e.g.
3517    megabytes), thus penalizing straightforward matches. It is now done only if the
3518    amount of subject to be scanned is less than 1000 bytes.
3519    
3520    7. A lesser problem with the same optimization is that it was recording the
3521    first character of an anchored pattern as "needed", thus provoking a search
3522    right along the subject, even when the first match of the pattern was going to
3523    fail. The "needed" character is now not set for anchored patterns, unless it
3524    follows something in the pattern that is of non-fixed length. Thus, it still
3525    fulfils its original purpose of finding quick non-matches in cases of nested
3526    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3527    
3528    
3529    Version 4.0 17-Feb-03
3530    ---------------------
3531    
3532    1. If a comment in an extended regex that started immediately after a meta-item
3533    extended to the end of string, PCRE compiled incorrect data. This could lead to
3534    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3535    
3536    2. Moved to autoconf 2.53 and libtool 1.4.2.
3537    
3538    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3539    the special perltest8 script is no longer needed - all the tests can be run
3540    from a single perltest script.