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

Diff of /code/trunk/ChangeLog

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

revision 37 by nigel, Sat Feb 24 21:39:09 2007 UTC revision 962 by ph10, Fri Apr 20 17:28:23 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.31
5    -----------------------------
6    
7    1.  Fixing a wrong JIT test case and some compiler warnings.
8    
9    2.  Removed a bashism from the RunTest script.
10    
11    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
12        to unsigned type, result still unsigned" that was given by an MS compiler
13        on encountering the code "-sizeof(xxx)".
14    
15    4.  Partial matching support is added to the JIT compiler.
16    
17    5.  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.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
37        running pcretest, the text "(JIT)" added to the output whenever JIT is
38        actually used to run the match.
39    
40    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
41        or /S+[+] with a digit between 1 and 7.
42    
43    8.  OP_NOT now supports any UTF character not just single-byte ones.
44    
45    9.  (*MARK) control verb is now supported by the JIT compiler.
46    
47    10. The command "./RunTest list" lists the available tests without actually
48        running any of them. (Because I keep forgetting what they all are.)
49    
50    11. Add PCRE_INFO_MAXLOOKBEHIND.
51    
52    12. Applied a (slightly modified) user-supplied patch that improves performance
53        when the heap is used for recursion (compiled with --disable-stack-for-
54        recursion). Instead of malloc and free for each heap frame each time a
55        logical recursion happens, frames are retained on a chain and re-used where
56        possible. This sometimes gives as much as 30% improvement.
57    
58    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
59        call.
60    
61    14. As documented, (*COMMIT) is now confined to within a positive assertion.
62    
63    15. It is now possible to link pcretest with libedit as an alternative to
64        libreadline.
65    
66    16. (*COMMIT) control verb is now supported by the JIT compiler.
67    
68    17. The Unicode data tables have been updated to Unicode 6.1.0.
69    
70    18. Added --file-list option to pcregrep.
71    
72    19. Added binary file support to pcregrep, including the -a, --binary-files,
73        -I, and --text options.
74    
75    20. The madvise function is renamed for posix_madvise for QNX compatibility
76        reasons. Fixed by Giuseppe D'Angelo.
77    
78    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
79    
80    22. Changed the option for creating symbolic links for 16-bit man pages from
81        -s to -sf so that re-installing does not cause issues.
82    
83    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
84    
85    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
86        matches in certain environments (the workspace was not being correctly
87        retained). Also added to pcre_dfa_exec() a simple plausibility check on
88        some of the workspace data at the beginning of a restart.
89    
90    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
91        was not doing so when it should - probably a typo introduced by SVN 528
92        (change 8.10/14).
93    
94    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
95        \w+ when the character tables indicated that \x{c4} was a word character.
96        There were several related cases, all because the tests for doing a table
97        lookup were testing for characters less than 127 instead of 255.
98    
99    
100    Version 8.30 04-February-2012
101    -----------------------------
102    
103    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
104        name is defined in ctype.h.
105    
106    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
107        only in quite long subpatterns.
108    
109    3.  Removed the function pcre_info(), which has been obsolete and deprecated
110        since it was replaced by pcre_fullinfo() in February 2000.
111    
112    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
113        match a (*MARK), and the match failed at the start of the subject, a
114        reference to memory before the start of the subject could occur. This bug
115        was introduced by fix 17 of release 8.21.
116    
117    5.  A reference to an unset group with zero minimum repetition was giving
118        totally wrong answers (in non-JavaScript-compatibility mode). For example,
119        /(another)?(\1?)test/ matched against "hello world test". This bug was
120        introduced in release 8.13.
121    
122    6.  Add support for 16-bit character strings (a large amount of work involving
123        many changes and refactorings).
124    
125    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
126        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
127        from a file.
128    
129    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
130        rounding is not applied in this particular case).
131    
132    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
133        if they appear, or are escaped, in patterns.
134    
135    10. Get rid of a number of -Wunused-but-set-variable warnings.
136    
137    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
138        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
139        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
140        also returns the mark "x". This bug applied to capturing parentheses,
141        non-capturing parentheses, and atomic parentheses. It also applied to some
142        assertions.
143    
144    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
145        information out of configure.ac instead of relying on pcre.h.generic, which
146        is not stored in the repository.
147    
148    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
149        -lreadline.
150    
151    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
152    
153    15. Applied Graycode's patch to put the top-level frame on the stack rather
154        than the heap when not using the stack for recursion. This gives a
155        performance improvement in many cases when recursion is not deep.
156    
157    16. Experimental code added to "pcretest -C" to output the stack frame size.
158    
159    
160    Version 8.21 12-Dec-2011
161    ------------------------
162    
163    1.  Updating the JIT compiler.
164    
165    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
166        are added as well.
167    
168    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
169        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
170        calling _pcre_jit_exec. Some extra comments are added.
171    
172    4.  (*MARK) settings inside atomic groups that do not contain any capturing
173        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
174        was introduced by change 18 for 8.20.
175    
176    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
177        ECMA-262 standard.
178    
179    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
180        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
181        This bug was probably introduced by change 9 of 8.13.
182    
183    7.  While fixing 6 above, I noticed that a number of other items were being
184        incorrectly rejected as "not fixed length". This arose partly because newer
185        opcodes had not been added to the fixed-length checking code. I have (a)
186        corrected the bug and added tests for these items, and (b) arranged for an
187        error to occur if an unknown opcode is encountered while checking for fixed
188        length instead of just assuming "not fixed length". The items that were
189        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
190        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
191        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
192    
193    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
194        being incorrectly compiled and would have given unpredicatble results.
195    
196    9.  A possessively repeated subpattern with minimum repeat count greater than
197        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
198        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
199        the first (A) could occur when it should not.
200    
201    10. Add a cast and remove a redundant test from the code.
202    
203    11. JIT should use pcre_malloc/pcre_free for allocation.
204    
205    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
206        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
207        is anything other than /usr, -L is still shown).
208    
209    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
210    
211    14. Perl does not support \N without a following name in a [] class; PCRE now
212        also gives an error.
213    
214    15. If a forward reference was repeated with an upper limit of around 2000,
215        it caused the error "internal error: overran compiling workspace". The
216        maximum number of forward references (including repeats) was limited by the
217        internal workspace, and dependent on the LINK_SIZE. The code has been
218        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
219        the default depends on LINK_SIZE. There is a new upper limit (for safety)
220        of around 200,000 forward references. While doing this, I also speeded up
221        the filling in of repeated forward references.
222    
223    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
224        incorrectly expecting the subject to contain another "a" after the start.
225    
226    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
227        in the match, the SKIP should be ignored. This was not happening; instead
228        the SKIP was being treated as NOMATCH. For patterns such as
229        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
230        tested.
231    
232    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
233        now much more compatible with Perl, in particular in cases where the result
234        is a non-match for a non-anchored pattern. For example, if
235        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
236        "m", where previously it did not return a name. A side effect of this
237        change is that for partial matches, the last encountered mark name is
238        returned, as for non matches. A number of tests that were previously not
239        Perl-compatible have been moved into the Perl-compatible test files. The
240        refactoring has had the pleasing side effect of removing one argument from
241        the match() function, thus reducing its stack requirements.
242    
243    19. If the /S+ option was used in pcretest to study a pattern using JIT,
244        subsequent uses of /S (without +) incorrectly behaved like /S+.
245    
246    21. Retrieve executable code size support for the JIT compiler and fixing
247        some warnings.
248    
249    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
250        not work when the shorter character appeared right at the end of the
251        subject string.
252    
253    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
254        systems.
255    
256    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
257        output it when the /M option is used in pcretest.
258    
259    25. The CheckMan script was not being included in the distribution. Also, added
260        an explicit "perl" to run Perl scripts from the PrepareRelease script
261        because this is reportedly needed in Windows.
262    
263    26. If study data was being save in a file and studying had not found a set of
264        "starts with" bytes for the pattern, the data written to the file (though
265        never used) was taken from uninitialized memory and so caused valgrind to
266        complain.
267    
268    27. Updated RunTest.bat as provided by Sheri Pierce.
269    
270    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
271    
272    29. Computation of memory usage for the table of capturing group names was
273        giving an unnecessarily large value.
274    
275    
276    Version 8.20 21-Oct-2011
277    ------------------------
278    
279    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
280        a POSIX class. After further experiments with Perl, which convinced me that
281        Perl has bugs and confusions, a closing square bracket is no longer allowed
282        in a POSIX name. This bug also affected patterns with classes that started
283        with full stops.
284    
285    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
286        captured substring, but while checking the failing first alternative,
287        substring 1 is temporarily captured. If the output vector supplied to
288        pcre_exec() was not big enough for this capture, the yield of the function
289        was still zero ("insufficient space for captured substrings"). This cannot
290        be totally fixed without adding another stack variable, which seems a lot
291        of expense for a edge case. However, I have improved the situation in cases
292        such as /(a)(b)x|abc/ matched against "abc", where the return code
293        indicates that fewer than the maximum number of slots in the ovector have
294        been set.
295    
296    3.  Related to (2) above: when there are more back references in a pattern than
297        slots in the output vector, pcre_exec() uses temporary memory during
298        matching, and copies in the captures as far as possible afterwards. It was
299        using the entire output vector, but this conflicts with the specification
300        that only 2/3 is used for passing back captured substrings. Now it uses
301        only the first 2/3, for compatibility. This is, of course, another edge
302        case.
303    
304    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
305        main code base, and can be used by building with --enable-jit. When this is
306        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
307        runtime --no-jit option is given.
308    
309    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
310        ovector, the return from the function was zero, implying that there were
311        other matches that did not fit. The correct "exactly full" value is now
312        returned.
313    
314    6.  If a subpattern that was called recursively or as a subroutine contained
315        (*PRUNE) or any other control that caused it to give a non-standard return,
316        invalid errors such as "Error -26 (nested recursion at the same subject
317        position)" or even infinite loops could occur.
318    
319    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
320        computing the minimum length on reaching *ACCEPT, and so ended up with the
321        wrong value of 1 rather than 0. Further investigation indicates that
322        computing a minimum subject length in the presence of *ACCEPT is difficult
323        (think back references, subroutine calls), and so I have changed the code
324        so that no minimum is registered for a pattern that contains *ACCEPT.
325    
326    8.  If (*THEN) was present in the first (true) branch of a conditional group,
327        it was not handled as intended. [But see 16 below.]
328    
329    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
330        Sheri Pierce.
331    
332    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
333        the first byte in a match must be "a".
334    
335    11. Change 17 for 8.13 increased the recursion depth for patterns like
336        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
337        pattern contains any instances of (*THEN). If it does not, the old
338        optimizations are restored. It would be nice to do this on a per-group
339        basis, but at the moment that is not feasible.
340    
341    12. In some environments, the output of pcretest -C is CRLF terminated. This
342        broke RunTest's code that checks for the link size. A single white space
343        character after the value is now allowed for.
344    
345    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
346        For "fr", it uses the Windows-specific input and output files.
347    
348    14. If (*THEN) appeared in a group that was called recursively or as a
349        subroutine, it did not work as intended. [But see next item.]
350    
351    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
352        pattern fragments (but not containing any | characters). If A and B are
353        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
354        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
355        D. In other words, Perl considers parentheses that do not contain any |
356        characters to be part of a surrounding alternative, whereas PCRE was
357        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
358        differently. PCRE now behaves in the same way as Perl, except in the case
359        of subroutine/recursion calls such as (?1) which have in any case always
360        been different (but PCRE had them first :-).
361    
362    16. Related to 15 above: Perl does not treat the | in a conditional group as
363        creating alternatives. Such a group is treated in the same way as an
364        ordinary group without any | characters when processing (*THEN). PCRE has
365        been changed to match Perl's behaviour.
366    
367    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
368        RunGrepTest script failed.
369    
370    18. Change 22 for version 13 caused atomic groups to use more stack. This is
371        inevitable for groups that contain captures, but it can lead to a lot of
372        stack use in large patterns. The old behaviour has been restored for atomic
373        groups that do not contain any capturing parentheses.
374    
375    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
376        suppress the check for a minimum subject length at run time. (If it was
377        given to pcre_exec() or pcre_dfa_exec() it did work.)
378    
379    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
380        fail to work when decoding hex characters in data strings in EBCDIC
381        environments.
382    
383    21. It appears that in at least one Mac OS environment, the isxdigit() function
384        is implemented as a macro that evaluates to its argument more than once,
385        contravening the C 90 Standard (I haven't checked a later standard). There
386        was an instance in pcretest which caused it to go wrong when processing
387        \x{...} escapes in subject strings. The has been rewritten to avoid using
388        things like p++ in the argument of isxdigit().
389    
390    
391    Version 8.13 16-Aug-2011
392    ------------------------
393    
394    1.  The Unicode data tables have been updated to Unicode 6.0.0.
395    
396    2.  Two minor typos in pcre_internal.h have been fixed.
397    
398    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
399        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
400        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
401    
402    4.  There were a number of related bugs in the code for matching backrefences
403        caselessly in UTF-8 mode when codes for the characters concerned were
404        different numbers of bytes. For example, U+023A and U+2C65 are an upper
405        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
406        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
407        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
408        2-byte code at the end of the subject (it thought there wasn't enough data
409        left).
410    
411    5.  Comprehensive information about what went wrong is now returned by
412        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
413        as the output vector has at least 2 elements. The offset of the start of
414        the failing character and a reason code are placed in the vector.
415    
416    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
417        now returned is for the first byte of the failing character, instead of the
418        last byte inspected. This is an incompatible change, but I hope it is small
419        enough not to be a problem. It makes the returned offset consistent with
420        pcre_exec() and pcre_dfa_exec().
421    
422    7.  pcretest now gives a text phrase as well as the error number when
423        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
424        failure, the offset and reason code are output.
425    
426    8.  When \R was used with a maximizing quantifier it failed to skip backwards
427        over a \r\n pair if the subsequent match failed. Instead, it just skipped
428        back over a single character (\n). This seems wrong (because it treated the
429        two characters as a single entity when going forwards), conflicts with the
430        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
431        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
432        has been changed.
433    
434    9.  Some internal refactoring has changed the processing so that the handling
435        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
436        time (the PCRE_DOTALL option was changed this way some time ago: version
437        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
438        which was always a bit of a fudge. It also means that there is one less
439        argument for the match() function, which reduces its stack requirements
440        slightly. This change also fixes an incompatibility with Perl: the pattern
441        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
442    
443    10. More internal refactoring has drastically reduced the number of recursive
444        calls to match() for possessively repeated groups such as (abc)++ when
445        using pcre_exec().
446    
447    11. While implementing 10, a number of bugs in the handling of groups were
448        discovered and fixed:
449    
450        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
451        (a|)*(?1) gave a compile-time internal error.
452        ((a|)+)+  did not notice that the outer group could match an empty string.
453        (^a|^)+   was not marked as anchored.
454        (.*a|.*)+ was not marked as matching at start or after a newline.
455    
456    12. Yet more internal refactoring has removed another argument from the match()
457        function. Special calls to this function are now indicated by setting a
458        value in a variable in the "match data" data block.
459    
460    13. Be more explicit in pcre_study() instead of relying on "default" for
461        opcodes that mean there is no starting character; this means that when new
462        ones are added and accidentally left out of pcre_study(), testing should
463        pick them up.
464    
465    14. The -s option of pcretest has been documented for ages as being an old
466        synonym of -m (show memory usage). I have changed it to mean "force study
467        for every regex", that is, assume /S for every regex. This is similar to -i
468        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
469        using it. It makes it easier to run collections of tests with and without
470        study enabled, and thereby test pcre_study() more easily. All the standard
471        tests are now run with and without -s (but some patterns can be marked as
472        "never study" - see 20 below).
473    
474    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
475        restoration of the capturing data to the outer values was not happening
476        correctly.
477    
478    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
479        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
480        pattern had matched an empty string, and so incorrectly returned a no
481        match.
482    
483    17. There was optimizing code for the last branch of non-capturing parentheses,
484        and also for the obeyed branch of a conditional subexpression, which used
485        tail recursion to cut down on stack usage. Unfortunately, now that there is
486        the possibility of (*THEN) occurring in these branches, tail recursion is
487        no longer possible because the return has to be checked for (*THEN). These
488        two optimizations have therefore been removed. [But see 8.20/11 above.]
489    
490    18. If a pattern containing \R was studied, it was assumed that \R always
491        matched two bytes, thus causing the minimum subject length to be
492        incorrectly computed because \R can also match just one byte.
493    
494    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
495        was incorrectly computed.
496    
497    20. If /S is present twice on a test pattern in pcretest input, it now
498        *disables* studying, thereby overriding the use of -s on the command line
499        (see 14 above). This is necessary for one or two tests to keep the output
500        identical in both cases.
501    
502    21. When (*ACCEPT) was used in an assertion that matched an empty string and
503        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
504    
505    22. When an atomic group that contained a capturing parenthesis was
506        successfully matched, but the branch in which it appeared failed, the
507        capturing was not being forgotten if a higher numbered group was later
508        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
509        group 1 to "a", when in fact it should be unset. This applied to multi-
510        branched capturing and non-capturing groups, repeated or not, and also to
511        positive assertions (capturing in negative assertions does not happen
512        in PCRE) and also to nested atomic groups.
513    
514    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
515        subject after a captured substring, to make it easier to tell which of a
516        number of identical substrings has been captured.
517    
518    24. The way atomic groups are processed by pcre_exec() has been changed so that
519        if they are repeated, backtracking one repetition now resets captured
520        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
521        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
522        "aaa". Previously, it would have been "a". As part of this code
523        refactoring, the way recursive calls are handled has also been changed.
524    
525    25. If an assertion condition captured any substrings, they were not passed
526        back unless some other capturing happened later. For example, if
527        (?(?=(a))a) was matched against "a", no capturing was returned.
528    
529    26. When studying a pattern that contained subroutine calls or assertions,
530        the code for finding the minimum length of a possible match was handling
531        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
532        group 1 called group 2 while simultaneously a separate group 2 called group
533        1). A stack overflow occurred in this case. I have fixed this by limiting
534        the recursion depth to 10.
535    
536    27. Updated RunTest.bat in the distribution to the version supplied by Tom
537        Fortmann. This supports explicit test numbers on the command line, and has
538        argument validation and error reporting.
539    
540    28. An instance of \X with an unlimited repeat could fail if at any point the
541        first character it looked at was a mark character.
542    
543    29. Some minor code refactoring concerning Unicode properties and scripts
544        should reduce the stack requirement of match() slightly.
545    
546    30. Added the '=' option to pcretest to check the setting of unused capturing
547        slots at the end of the pattern, which are documented as being -1, but are
548        not included in the return count.
549    
550    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
551        compiled something random. Now it gives a compile-time error (as does
552        Perl).
553    
554    32. A *MARK encountered during the processing of a positive assertion is now
555        recorded and passed back (compatible with Perl).
556    
557    33. If --only-matching or --colour was set on a pcregrep call whose pattern
558        had alternative anchored branches, the search for a second match in a line
559        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
560        matched the line "0102" twice. The same bug affected patterns that started
561        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
562        twice.
563    
564    34. Previously, PCRE did not allow quantification of assertions. However, Perl
565        does, and because of capturing effects, quantifying parenthesized
566        assertions may at times be useful. Quantifiers are now allowed for
567        parenthesized assertions.
568    
569    35. A minor code tidy in pcre_compile() when checking options for \R usage.
570    
571    36. \g was being checked for fancy things in a character class, when it should
572        just be a literal "g".
573    
574    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
575        appearance of a nested POSIX class supersedes an apparent external class.
576        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
577        unescaped square brackets may also appear as part of class names. For
578        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
579        more like Perl. (But see 8.20/1 above.)
580    
581    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
582        was because it thought it was \N{name}, which is not supported).
583    
584    39. Add minix to OS list not supporting the -S option in pcretest.
585    
586    40. PCRE tries to detect cases of infinite recursion at compile time, but it
587        cannot analyze patterns in sufficient detail to catch mutual recursions
588        such as ((?1))((?2)). There is now a runtime test that gives an error if a
589        subgroup is called recursively as a subpattern for a second time at the
590        same position in the subject string. In previous releases this might have
591        been caught by the recursion limit, or it might have run out of stack.
592    
593    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
594        happen only once. PCRE was, however incorrectly giving a compile time error
595        "recursive call could loop indefinitely" because it cannot analyze the
596        pattern in sufficient detail. The compile time test no longer happens when
597        PCRE is compiling a conditional subpattern, but actual runaway loops are
598        now caught at runtime (see 40 above).
599    
600    42. It seems that Perl allows any characters other than a closing parenthesis
601        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
602        has been changed to be the same.
603    
604    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
605        as not to get warnings when autogen.sh is called. Also changed
606        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
607    
608    44. To help people who use pcregrep to scan files containing exceedingly long
609        lines, the following changes have been made:
610    
611        (a) The default value of the buffer size parameter has been increased from
612            8K to 20K. (The actual buffer used is three times this size.)
613    
614        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
615            PCRE is built.
616    
617        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
618            to be set at run time.
619    
620        (d) Numerical values in pcregrep options can be followed by K or M, for
621            example --buffer-size=50K.
622    
623        (e) If a line being scanned overflows pcregrep's buffer, an error is now
624            given and the return code is set to 2.
625    
626    45. Add a pointer to the latest mark to the callout data block.
627    
628    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
629        partial match of an empty string instead of no match. This was specific to
630        the use of ".".
631    
632    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
633        complete match instead of a partial match. This bug was dependent on both
634        the PCRE_UTF8 and PCRE_DOTALL options being set.
635    
636    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
637        starting byte set, because \b was not being ignored.
638    
639    
640    Version 8.12 15-Jan-2011
641    ------------------------
642    
643    1.  Fixed some typos in the markup of the man pages, and wrote a script that
644        checks for such things as part of the documentation building process.
645    
646    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
647        --match-limit and --recursion-limit options (added for 8.11). In
648        particular, this made one of the standard tests fail. (The integer value
649        went into the wrong half of a long int.)
650    
651    3.  If the --colour option was given to pcregrep with -v (invert match), it
652        did strange things, either producing crazy output, or crashing. It should,
653        of course, ignore a request for colour when reporting lines that do not
654        match.
655    
656    4.  Another pcregrep bug caused similar problems if --colour was specified with
657        -M (multiline) and the pattern match finished with a line ending.
658    
659    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
660        matched in multiline mode, the following line was shown as part of the
661        match. This seems wrong, so I have changed it.
662    
663    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
664        the check for further matches in the same line (so they could be coloured)
665        to overrun the end of the current line. If another match was found, it was
666        incorrectly shown (and then shown again when found in the next line).
667    
668    7.  If pcregrep was compiled under Windows, there was a reference to the
669        function pcregrep_exit() before it was defined. I am assuming this was
670        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
671        reported by a user. I've moved the definition above the reference.
672    
673    
674    Version 8.11 10-Dec-2010
675    ------------------------
676    
677    1.  (*THEN) was not working properly if there were untried alternatives prior
678        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
679        backtracked to try for "b" instead of moving to the next alternative branch
680        at the same level (in this case, to look for "c"). The Perl documentation
681        is clear that when (*THEN) is backtracked onto, it goes to the "next
682        alternative in the innermost enclosing group".
683    
684    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
685        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
686        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
687        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
688        (*THEN).
689    
690    3.  If \s appeared in a character class, it removed the VT character from
691        the class, even if it had been included by some previous item, for example
692        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
693        of \s, but is part of the POSIX "space" class.)
694    
695    4.  A partial match never returns an empty string (because you can always
696        match an empty string at the end of the subject); however the checking for
697        an empty string was starting at the "start of match" point. This has been
698        changed to the "earliest inspected character" point, because the returned
699        data for a partial match starts at this character. This means that, for
700        example, /(?<=abc)def/ gives a partial match for the subject "abc"
701        (previously it gave "no match").
702    
703    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
704        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
705        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
706        has an implication that the given string is incomplete (because a partial
707        match is preferred over a full match). For this reason, these items now
708        give a partial match in this situation. [Aside: previously, the one case
709        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
710        match rather than a full match, which was wrong by the old rules, but is
711        now correct.]
712    
713    6.  There was a bug in the handling of #-introduced comments, recognized when
714        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
715        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
716        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
717        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
718        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
719        places in pcre_compile().
720    
721    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
722        comments when looking ahead for named forward references to subpatterns,
723        the only newline sequence it recognized was NL. It now handles newlines
724        according to the set newline convention.
725    
726    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
727        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
728        cater for a lack of strerror(). These oversights have been fixed.
729    
730    9.  Added --match-limit and --recursion-limit to pcregrep.
731    
732    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
733    
734    11. When the -o option was used, pcregrep was setting a return code of 1, even
735        when matches were found, and --line-buffered was not being honoured.
736    
737    12. Added an optional parentheses number to the -o and --only-matching options
738        of pcregrep.
739    
740    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
741        can match an empty string. The code to do it in pcretest and pcredemo
742        needed fixing:
743    
744        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
745            only one byte after an empty string match just before CRLF (this case
746            just got forgotten; "any" and "anycrlf" were OK).
747    
748        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
749            mode when an empty string match preceded an ASCII character followed by
750            a non-ASCII character. (The code for advancing by one character rather
751            than one byte was nonsense.)
752    
753        (c) The pcredemo.c sample program did not have any code at all to handle
754            the cases when CRLF is a valid newline sequence.
755    
756    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
757        as a starting offset was within the subject string. There is now a new
758        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
759        negative or greater than the length of the string. In order to test this,
760        pcretest is extended to allow the setting of negative starting offsets.
761    
762    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
763        starting offset points to the beginning of a UTF-8 character was
764        unnecessarily clumsy. I tidied it up.
765    
766    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
767        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
768    
769    17. Nobody had reported that the --include_dir option, which was added in
770        release 7.7 should have been called --include-dir (hyphen, not underscore)
771        for compatibility with GNU grep. I have changed it to --include-dir, but
772        left --include_dir as an undocumented synonym, and the same for
773        --exclude-dir, though that is not available in GNU grep, at least as of
774        release 2.5.4.
775    
776    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
777        characters from a string of bytes) have been redefined so as not to use
778        loops, in order to improve performance in some environments. At the same
779        time, I abstracted some of the common code into auxiliary macros to save
780        repetition (this should not affect the compiled code).
781    
782    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
783        compile-time error is now given if \c is not followed by an ASCII
784        character, that is, a byte less than 128. (In EBCDIC mode, the code is
785        different, and any byte value is allowed.)
786    
787    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
788        START_OPTIMIZE option, which is now allowed at compile time - but just
789        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
790        to pcregrep and other applications that have no direct access to PCRE
791        options. The new /Y option in pcretest sets this option when calling
792        pcre_compile().
793    
794    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
795        back references. Groups containing recursive back references were forced to
796        be atomic by that change, but in the case of named groups, the amount of
797        memory required was incorrectly computed, leading to "Failed: internal
798        error: code overflow". This has been fixed.
799    
800    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
801        pcretest.c, to avoid build problems in some Borland environments.
802    
803    
804    Version 8.10 25-Jun-2010
805    ------------------------
806    
807    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
808        THEN.
809    
810    2.  (*ACCEPT) was not working when inside an atomic group.
811    
812    3.  Inside a character class, \B is treated as a literal by default, but
813        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
814        causes the error). The code is unchanged, but I tidied the documentation.
815    
816    4.  Inside a character class, PCRE always treated \R and \X as literals,
817        whereas Perl faults them if its -w option is set. I have changed PCRE so
818        that it faults them when PCRE_EXTRA is set.
819    
820    5.  Added support for \N, which always matches any character other than
821        newline. (It is the same as "." when PCRE_DOTALL is not set.)
822    
823    6.  When compiling pcregrep with newer versions of gcc which may have
824        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
825        declared with attribute warn_unused_result" were given. Just casting the
826        result to (void) does not stop the warnings; a more elaborate fudge is
827        needed. I've used a macro to implement this.
828    
829    7.  Minor change to pcretest.c to avoid a compiler warning.
830    
831    8.  Added four artifical Unicode properties to help with an option to make
832        \s etc use properties (see next item). The new properties are: Xan
833        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
834    
835    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
836        use Unicode properties. (*UCP) at the start of a pattern can be used to set
837        this option. Modified pcretest to add /W to test this facility. Added
838        REG_UCP to make it available via the POSIX interface.
839    
840    10. Added --line-buffered to pcregrep.
841    
842    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
843        studied, and the match started with a letter with a code point greater than
844        127 whose first byte was different to the first byte of the other case of
845        the letter, the other case of this starting letter was not recognized
846        (#976).
847    
848    12. If a pattern that was studied started with a repeated Unicode property
849        test, for example, \p{Nd}+, there was the theoretical possibility of
850        setting up an incorrect bitmap of starting bytes, but fortunately it could
851        not have actually happened in practice until change 8 above was made (it
852        added property types that matched character-matching opcodes).
853    
854    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
855        possible starting bytes for non-anchored patterns.
856    
857    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
858        \R, and also a number of cases that involve Unicode properties, both
859        explicit and implicit when PCRE_UCP is set.
860    
861    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
862        input, it could crash or give wrong results if characters with values
863        greater than 0xc0 were present in the subject string. (Detail: it assumed
864        UTF-8 input when processing these items.)
865    
866    16. Added a lot of (int) casts to avoid compiler warnings in systems where
867        size_t is 64-bit (#991).
868    
869    17. Added a check for running out of memory when PCRE is compiled with
870        --disable-stack-for-recursion (#990).
871    
872    18. If the last data line in a file for pcretest does not have a newline on
873        the end, a newline was missing in the output.
874    
875    19. The default pcre_chartables.c file recognizes only ASCII characters (values
876        less than 128) in its various bitmaps. However, there is a facility for
877        generating tables according to the current locale when PCRE is compiled. It
878        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
879        characters, are recognized by isspace() and therefore were getting set in
880        these tables, and indeed these tables seem to approximate to ISO 8859. This
881        caused a problem in UTF-8 mode when pcre_study() was used to create a list
882        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
883        which of course cannot start UTF-8 characters. I have changed the code so
884        that only real ASCII characters (less than 128) and the correct starting
885        bytes for UTF-8 encodings are set for characters greater than 127 when in
886        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
887        altogether.)
888    
889    20. Added the /T option to pcretest so as to be able to run tests with non-
890        standard character tables, thus making it possible to include the tests
891        used for 19 above in the standard set of tests.
892    
893    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
894        reference to a subpattern the other side of a comment that contains an
895        opening parenthesis caused either an internal compiling error, or a
896        reference to the wrong subpattern.
897    
898    
899    Version 8.02 19-Mar-2010
900    ------------------------
901    
902    1.  The Unicode data tables have been updated to Unicode 5.2.0.
903    
904    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
905        configured.
906    
907    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
908        original author of that file, following a query about its status.
909    
910    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
911        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
912    
913    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
914        quantifier applied to a forward-referencing subroutine call, could compile
915        incorrect code or give the error "internal error: previously-checked
916        referenced subpattern not found".
917    
918    6.  Both MS Visual Studio and Symbian OS have problems with initializing
919        variables to point to external functions. For these systems, therefore,
920        pcre_malloc etc. are now initialized to local functions that call the
921        relevant global functions.
922    
923    7.  There were two entries missing in the vectors called coptable and poptable
924        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
925        I've fixed the data, and added a kludgy way of testing at compile time that
926        the lengths are correct (equal to the number of opcodes).
927    
928    8.  Following on from 7, I added a similar kludge to check the length of the
929        eint vector in pcreposix.c.
930    
931    9.  Error texts for pcre_compile() are held as one long string to avoid too
932        much relocation at load time. To find a text, the string is searched,
933        counting zeros. There was no check for running off the end of the string,
934        which could happen if a new error number was added without updating the
935        string.
936    
937    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
938    
939    11. \K was not working if it appeared in an atomic group or in a group that
940        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
941        \K is "not well defined" if used in an assertion. PCRE now accepts it if
942        the assertion is positive, but not if it is negative.
943    
944    12. Change 11 fortuitously reduced the size of the stack frame used in the
945        "match()" function of pcre_exec.c by one pointer. Forthcoming
946        implementation of support for (*MARK) will need an extra pointer on the
947        stack; I have reserved it now, so that the stack frame size does not
948        decrease.
949    
950    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
951        item in branch that calls a recursion is a subroutine call - as in the
952        second branch in the above example - was incorrectly given the compile-
953        time error "recursive call could loop indefinitely" because pcre_compile()
954        was not correctly checking the subroutine for matching a non-empty string.
955    
956    14. The checks for overrunning compiling workspace could trigger after an
957        overrun had occurred. This is a "should never occur" error, but it can be
958        triggered by pathological patterns such as hundreds of nested parentheses.
959        The checks now trigger 100 bytes before the end of the workspace.
960    
961    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
962    
963    
964    Version 8.01 19-Jan-2010
965    ------------------------
966    
967    1.  If a pattern contained a conditional subpattern with only one branch (in
968        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
969        computed the wrong minimum data length (which is of course zero for such
970        subpatterns). This could cause incorrect "no match" results.
971    
972    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
973        the pattern is reset in the first branch, pcre_compile() failed with
974        "internal error: code overflow at offset...". This happened only when
975        the reset was to the original external option setting. (An optimization
976        abstracts leading options settings into an external setting, which was the
977        cause of this.)
978    
979    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
980        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
981        assertion pattern did not match (meaning that the assertion was true), it
982        was incorrectly treated as false if the SKIP had been reached during the
983        matching. This also applied to assertions used as conditions.
984    
985    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
986        assertion subpattern, including such a pattern used as a condition,
987        unpredictable results occurred, instead of the error return
988        PCRE_ERROR_DFA_UITEM.
989    
990    5.  The C++ GlobalReplace function was not working like Perl for the special
991        situation when an empty string is matched. It now does the fancy magic
992        stuff that is necessary.
993    
994    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
995        removed. (These were left over from very, very early versions of PCRE.)
996    
997    7.  Some cosmetic changes to the code to make life easier when compiling it
998        as part of something else:
999    
1000        (a) Change DEBUG to PCRE_DEBUG.
1001    
1002        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1003            called "current" as "current_branch", to prevent a collision with the
1004            Linux macro when compiled as a kernel module.
1005    
1006        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1007            prevent a collision with the Linux macro when compiled as a kernel
1008            module.
1009    
1010    8.  In pcre_compile() there are some checks for integer overflows that used to
1011        cast potentially large values to (double). This has been changed to that
1012        when building, a check for int64_t is made, and if it is found, it is used
1013        instead, thus avoiding the use of floating point arithmetic. (There is no
1014        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1015        double.
1016    
1017    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1018        2005 (difference between two addresses compared to an unsigned value).
1019    
1020    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1021        custom one, because of the following reported problem in Windows:
1022    
1023          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1024              under Win32.
1025          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1026              therefore missing the function definition.
1027          - The compiler thus generates a "C" signature for the test function.
1028          - The linker fails to find the "C" function.
1029          - PCRE fails to configure if asked to do so against libbz2.
1030    
1031    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1032        messages were output:
1033    
1034          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1035          rerunning libtoolize, to keep the correct libtool macros in-tree.
1036          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1037    
1038        I have done both of these things.
1039    
1040    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1041        most of the time, it *can* run out if it is given a pattern that contains a
1042        runaway infinite recursion. I updated the discussion in the pcrestack man
1043        page.
1044    
1045    13. Now that we have gone to the x.xx style of version numbers, the minor
1046        version may start with zero. Using 08 or 09 is a bad idea because users
1047        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1048        interpreted as invalid octal numbers. I've updated the previous comment in
1049        configure.ac, and also added a check that gives an error if 08 or 09 are
1050        used.
1051    
1052    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1053        causing partial matching to fail when the end of the subject matched \W
1054        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1055    
1056    15. There were some discrepancies between the declarations in pcre_internal.h
1057        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1058        their definitions. The declarations used "const uschar *" and the
1059        definitions used USPTR. Even though USPTR is normally defined as "const
1060        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1061        reported that: "This difference in casting confuses some C++ compilers, for
1062        example, SunCC recognizes above declarations as different functions and
1063        generates broken code for hbpcre." I have changed the declarations to use
1064        USPTR.
1065    
1066    16. GNU libtool is named differently on some systems. The autogen.sh script now
1067        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1068        (FreeBSD).
1069    
1070    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1071        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1072        comment: "Figure out how to create a longlong from a string: strtoll and
1073        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1074        instance, but it only takes 2 args instead of 3!"
1075    
1076    18. A subtle bug concerned with back references has been fixed by a change of
1077        specification, with a corresponding code fix. A pattern such as
1078        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1079        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1080        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1081        same bug. Such groups have to be quantified to be useful, or contained
1082        inside another quantified group. (If there's no repetition, the reference
1083        can never match.) The problem arises because, having left the group and
1084        moved on to the rest of the pattern, a later failure that backtracks into
1085        the group uses the captured value from the final iteration of the group
1086        rather than the correct earlier one. I have fixed this in PCRE by forcing
1087        any group that contains a reference to itself to be an atomic group; that
1088        is, there cannot be any backtracking into it once it has completed. This is
1089        similar to recursive and subroutine calls.
1090    
1091    
1092    Version 8.00 19-Oct-09
1093    ----------------------
1094    
1095    1.  The table for translating pcre_compile() error codes into POSIX error codes
1096        was out-of-date, and there was no check on the pcre_compile() error code
1097        being within the table. This could lead to an OK return being given in
1098        error.
1099    
1100    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1101        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1102        in a Windows environment.
1103    
1104    3.  The pcregrep --count option prints the count for each file even when it is
1105        zero, as does GNU grep. However, pcregrep was also printing all files when
1106        --files-with-matches was added. Now, when both options are given, it prints
1107        counts only for those files that have at least one match. (GNU grep just
1108        prints the file name in this circumstance, but including the count seems
1109        more useful - otherwise, why use --count?) Also ensured that the
1110        combination -clh just lists non-zero counts, with no names.
1111    
1112    4.  The long form of the pcregrep -F option was incorrectly implemented as
1113        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1114        but it seems right to fix it, and I didn't think it was worth preserving
1115        the old behaviour.
1116    
1117    5.  The command line items --regex=pattern and --regexp=pattern were not
1118        recognized by pcregrep, which required --regex pattern or --regexp pattern
1119        (with a space rather than an '='). The man page documented the '=' forms,
1120        which are compatible with GNU grep; these now work.
1121    
1122    6.  No libpcreposix.pc file was created for pkg-config; there was just
1123        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1124    
1125    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1126        when UCP support is not needed, by modifying the Python script that
1127        generates it from Unicode data files. This should not matter if the module
1128        is correctly used as a library, but I received one complaint about 50K of
1129        unwanted data. My guess is that the person linked everything into his
1130        program rather than using a library. Anyway, it does no harm.
1131    
1132    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1133        was a minimum greater than 1 for a wide character in a possessive
1134        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1135        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1136        character. Chaos in the form of incorrect output or a compiling loop could
1137        result.
1138    
1139    9.  The restrictions on what a pattern can contain when partial matching is
1140        requested for pcre_exec() have been removed. All patterns can now be
1141        partially matched by this function. In addition, if there are at least two
1142        slots in the offset vector, the offset of the earliest inspected character
1143        for the match and the offset of the end of the subject are set in them when
1144        PCRE_ERROR_PARTIAL is returned.
1145    
1146    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1147        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1148        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1149        and may be more useful for multi-segment matching.
1150    
1151    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1152        used to be given if ever the end of the subject was reached; now it is
1153        given only if matching could not proceed because another character was
1154        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1155        string "Z", which now yields "no match" instead of "partial match". In the
1156        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1157        final character ended with (*FAIL).
1158    
1159    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1160        if the pattern had a "must contain" character that was already found in the
1161        earlier partial match, unless partial matching was again requested. For
1162        example, with the pattern /dog.(body)?/, the "must contain" character is
1163        "g". If the first part-match was for the string "dog", restarting with
1164        "sbody" failed. This bug has been fixed.
1165    
1166    13. The string returned by pcre_dfa_exec() after a partial match has been
1167        changed so that it starts at the first inspected character rather than the
1168        first character of the match. This makes a difference only if the pattern
1169        starts with a lookbehind assertion or \b or \B (\K is not supported by
1170        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1171        matching functions compatible, and I think it's the right thing to do.
1172    
1173    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1174        so that the demonstration program is easily available in environments where
1175        PCRE has not been installed from source.
1176    
1177    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1178        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1179        library.
1180    
1181    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1182        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1183        is not the first non-POSIX option to be added. Clearly some people find
1184        these options useful.
1185    
1186    17. If a caller to the POSIX matching function regexec() passes a non-zero
1187        value for nmatch with a NULL value for pmatch, the value of
1188        nmatch is forced to zero.
1189    
1190    18. RunGrepTest did not have a test for the availability of the -u option of
1191        the diff command, as RunTest does. It now checks in the same way as
1192        RunTest, and also checks for the -b option.
1193    
1194    19. If an odd number of negated classes containing just a single character
1195        interposed, within parentheses, between a forward reference to a named
1196        subpattern and the definition of the subpattern, compilation crashed with
1197        an internal error, complaining that it could not find the referenced
1198        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1199        [The bug was that it was starting one character too far in when skipping
1200        over the character class, thus treating the ] as data rather than
1201        terminating the class. This meant it could skip too much.]
1202    
1203    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1204        /g option in pcretest when the pattern contains \K, which makes it possible
1205        to have an empty string match not at the start, even when the pattern is
1206        anchored. Updated pcretest and pcredemo to use this option.
1207    
1208    21. If the maximum number of capturing subpatterns in a recursion was greater
1209        than the maximum at the outer level, the higher number was returned, but
1210        with unset values at the outer level. The correct (outer level) value is
1211        now given.
1212    
1213    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1214        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1215        make it do so. The string so far is captured, making this feature
1216        compatible with Perl.
1217    
1218    23. The tests have been re-organized, adding tests 11 and 12, to make it
1219        possible to check the Perl 5.10 features against Perl 5.10.
1220    
1221    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1222        pattern matches a fixed length string. PCRE did not allow this; now it
1223        does. Neither allows recursion.
1224    
1225    25. I finally figured out how to implement a request to provide the minimum
1226        length of subject string that was needed in order to match a given pattern.
1227        (It was back references and recursion that I had previously got hung up
1228        on.) This code has now been added to pcre_study(); it finds a lower bound
1229        to the length of subject needed. It is not necessarily the greatest lower
1230        bound, but using it to avoid searching strings that are too short does give
1231        some useful speed-ups. The value is available to calling programs via
1232        pcre_fullinfo().
1233    
1234    26. While implementing 25, I discovered to my embarrassment that pcretest had
1235        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1236        study optimizations had never been tested with that matching function.
1237        Oops. What is worse, even when it was passed study data, there was a bug in
1238        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1239        were also very few tests of studied patterns with pcre_dfa_exec().
1240    
1241    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1242        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1243        on the other side of the coin, they are no longer allowed to have different
1244        names, because these cannot be distinguished in PCRE, and this has caused
1245        confusion. (This is a difference from Perl.)
1246    
1247    28. When duplicate subpattern names are present (necessarily with different
1248        numbers, as required by 27 above), and a test is made by name in a
1249        conditional pattern, either for a subpattern having been matched, or for
1250        recursion in such a pattern, all the associated numbered subpatterns are
1251        tested, and the overall condition is true if the condition is true for any
1252        one of them. This is the way Perl works, and is also more like the way
1253        testing by number works.
1254    
1255    
1256    Version 7.9 11-Apr-09
1257    ---------------------
1258    
1259    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1260        (pcretest), all targets were linked against these libraries. This included
1261        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1262        libraries. This caused unwanted dependencies to be created. This problem
1263        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1264        pcretest is linked with readline.
1265    
1266    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1267        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1268        moved outside it again, because FALSE and TRUE are already defined in AIX,
1269        but BOOL is not.
1270    
1271    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1272        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1273    
1274    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1275        hyphens) following file names and line numbers when outputting matching
1276        lines. This is not true; no spaces are inserted. I have also clarified the
1277        wording for the --colour (or --color) option.
1278    
1279    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1280        was not coloured; this is different to GNU grep, so I have changed it to be
1281        the same.
1282    
1283    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1284        each matching line was coloured. Now it goes on to look for further matches
1285        of any of the test patterns, which is the same behaviour as GNU grep.
1286    
1287    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1288        doesn't make sense to accept an empty string match in pcregrep, so I have
1289        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1290        seems to be how GNU grep behaves.
1291    
1292    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1293        start or after a newline", because the conditional assertion was not being
1294        correctly handled. The rule now is that both the assertion and what follows
1295        in the first alternative must satisfy the test.
1296    
1297    9.  If auto-callout was enabled in a pattern with a conditional group whose
1298        condition was an assertion, PCRE could crash during matching, both with
1299        pcre_exec() and pcre_dfa_exec().
1300    
1301    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1302        used for matching.
1303    
1304    11. Unicode property support in character classes was not working for
1305        characters (bytes) greater than 127 when not in UTF-8 mode.
1306    
1307    12. Added the -M command line option to pcretest.
1308    
1309    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1310    
1311    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1312    
1313    16. Added comments and documentation about mis-use of no_arg in the C++
1314        wrapper.
1315    
1316    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1317        from Martin Jerabek that uses macro names for all relevant character and
1318        string constants.
1319    
1320    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1321        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1322        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1323        these, but not everybody uses configure.
1324    
1325    19. A conditional group that had only one branch was not being correctly
1326        recognized as an item that could match an empty string. This meant that an
1327        enclosing group might also not be so recognized, causing infinite looping
1328        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1329        with the subject "ab", where knowledge that the repeated group can match
1330        nothing is needed in order to break the loop.
1331    
1332    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1333        exec(), but without supplying a callout function, matching went wrong.
1334    
1335    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1336        leak if the size of the offset vector was greater than 30. When the vector
1337        is smaller, the saved offsets during recursion go onto a local stack
1338        vector, but for larger vectors malloc() is used. It was failing to free
1339        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1340        error, in fact).
1341    
1342    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1343        heapframe that is used only when UTF-8 support is enabled. This caused no
1344        problem, but was untidy.
1345    
1346    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1347        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1348        included within another project.
1349    
1350    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1351        slightly modified by me:
1352    
1353          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1354              not building pcregrep.
1355    
1356          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1357              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1358    
1359    25. Forward references, both numeric and by name, in patterns that made use of
1360        duplicate group numbers, could behave incorrectly or give incorrect errors,
1361        because when scanning forward to find the reference group, PCRE was not
1362        taking into account the duplicate group numbers. A pattern such as
1363        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1364    
1365    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1366        the feature of a custom pointer more persuasive (as requested by a user).
1367    
1368    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1369        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1370        pre-defined.
1371    
1372    28. Added support for (*UTF8) at the start of a pattern.
1373    
1374    29. Arrange for flags added by the "release type" setting in CMake to be shown
1375        in the configuration summary.
1376    
1377    
1378    Version 7.8 05-Sep-08
1379    ---------------------
1380    
1381    1.  Replaced UCP searching code with optimized version as implemented for Ad
1382        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1383        stage table and inline lookup instead of a function, giving speed ups of 2
1384        to 5 times on some simple patterns that I tested. Permission was given to
1385        distribute the MultiStage2.py script that generates the tables (it's not in
1386        the tarball, but is in the Subversion repository).
1387    
1388    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1389        scripts.
1390    
1391    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1392        a group with a zero qualifier. The result of the study could be incorrect,
1393        or the function might crash, depending on the pattern.
1394    
1395    4.  Caseless matching was not working for non-ASCII characters in back
1396        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1397        It now works when Unicode Property Support is available.
1398    
1399    5.  In pcretest, an escape such as \x{de} in the data was always generating
1400        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1401        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1402        truncation.
1403    
1404    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1405    
1406    7.  Added two (int) casts to pcregrep when printing the difference of two
1407        pointers, in case they are 64-bit values.
1408    
1409    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1410        test 2 if it fails.
1411    
1412    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1413        and a #define of that name to empty if it is not externally set. This is to
1414        allow users of MSVC to set it if necessary.
1415    
1416    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1417        the convenience functions in the pcre_get.c source file.
1418    
1419    11. An option change at the start of a pattern that had top-level alternatives
1420        could cause overwriting and/or a crash. This command provoked a crash in
1421        some environments:
1422    
1423          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1424    
1425        This potential security problem was recorded as CVE-2008-2371.
1426    
1427    12. For a pattern where the match had to start at the beginning or immediately
1428        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1429        pcre_dfa_exec() could read past the end of the passed subject if there was
1430        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1431        pcretest so that it places the subject at the end of its malloc-ed buffer.
1432    
1433    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1434        exec() might read past the end of the data buffer in UTF-8 mode.
1435    
1436    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1437        the data contained the byte 0x85 as part of a UTF-8 character within its
1438        first line. This applied both to normal and DFA matching.
1439    
1440    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1441        /^[^d]*?$/8 failed to match "abc".
1442    
1443    16. Added a missing copyright notice to pcrecpp_internal.h.
1444    
1445    17. Make it more clear in the documentation that values returned from
1446        pcre_exec() in ovector are byte offsets, not character counts.
1447    
1448    18. Tidied a few places to stop certain compilers from issuing warnings.
1449    
1450    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1451        supplied by Stefan Weber. I made a further small update for 7.8 because
1452        there is a change of source arrangements: the pcre_searchfuncs.c module is
1453        replaced by pcre_ucd.c.
1454    
1455    
1456    Version 7.7 07-May-08
1457    ---------------------
1458    
1459    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1460        a string to a long long, pretend we don't even have a long long." This is
1461        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1462    
1463    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1464        pre-7.6 versions, which defined a global no_arg variable instead of putting
1465        it in the RE class. (See also #8 below.)
1466    
1467    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1468        Lopes.
1469    
1470    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1471    
1472        (1) The include/exclude patterns were being applied to the whole pathnames
1473            of files, instead of just to the final components.
1474    
1475        (2) If there was more than one level of directory, the subdirectories were
1476            skipped unless they satisfied the include/exclude conditions. This is
1477            inconsistent with GNU grep (and could even be seen as contrary to the
1478            pcregrep specification - which I improved to make it absolutely clear).
1479            The action now is always to scan all levels of directory, and just
1480            apply the include/exclude patterns to regular files.
1481    
1482    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1483        --exclude_dir in the tests to avoid scanning .svn directories.
1484    
1485    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1486        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1487        doesn't support NULs in patterns.
1488    
1489    7.  Added some missing "const"s to declarations of static tables in
1490        pcre_compile.c and pcre_dfa_exec.c.
1491    
1492    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1493        caused by fix #2  above. (Subsequently also a second patch to fix the
1494        first patch. And a third patch - this was a messy problem.)
1495    
1496    9.  Applied Craig's patch to remove the use of push_back().
1497    
1498    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1499        matching function regexec().
1500    
1501    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1502        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1503        references. PCRE supports relative numbers with this syntax (I don't think
1504        Oniguruma does).
1505    
1506    12. Previously, a group with a zero repeat such as (...){0} was completely
1507        omitted from the compiled regex. However, this means that if the group
1508        was called as a subroutine from elsewhere in the pattern, things went wrong
1509        (an internal error was given). Such groups are now left in the compiled
1510        pattern, with a new opcode that causes them to be skipped at execution
1511        time.
1512    
1513    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1514        to the way PCRE behaves:
1515    
1516        (a) A lone ] character is dis-allowed (Perl treats it as data).
1517    
1518        (b) A back reference to an unmatched subpattern matches an empty string
1519            (Perl fails the current match path).
1520    
1521        (c) A data ] in a character class must be notated as \] because if the
1522            first data character in a class is ], it defines an empty class. (In
1523            Perl it is not possible to have an empty class.) The empty class []
1524            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1525            The negative empty class [^] matches any one character, independently
1526            of the DOTALL setting.
1527    
1528    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1529        non-existent subpattern following a character class starting with ']' and
1530        containing () gave an internal compiling error instead of "reference to
1531        non-existent subpattern". Fortunately, when the pattern did exist, the
1532        compiled code was correct. (When scanning forwards to check for the
1533        existencd of the subpattern, it was treating the data ']' as terminating
1534        the class, so got the count wrong. When actually compiling, the reference
1535        was subsequently set up correctly.)
1536    
1537    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1538        it was being rejected as not supported by pcre_dfa_exec(), even though
1539        other assertions are supported. I have made pcre_dfa_exec() support
1540        (*FAIL).
1541    
1542    16. The implementation of 13c above involved the invention of a new opcode,
1543        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1544        cannot be changed at match time, I realized I could make a small
1545        improvement to matching performance by compiling OP_ALLANY instead of
1546        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1547        on the OP_ANY path.
1548    
1549    17. Compiling pcretest on Windows with readline support failed without the
1550        following two fixes: (1) Make the unistd.h include conditional on
1551        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1552    
1553    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1554        ncurses library to be included for pcretest when ReadLine support is
1555        requested, but also to allow for it to be overridden. This patch came from
1556        Daniel Bergström.
1557    
1558    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1559        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1560        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1561        spotting this.
1562    
1563    
1564    Version 7.6 28-Jan-08
1565    ---------------------
1566    
1567    1.  A character class containing a very large number of characters with
1568        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1569        overflow.
1570    
1571    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1572        HAVE_LONG_LONG is not defined.
1573    
1574    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1575        bring them up to date and include new features. This patch includes:
1576    
1577        - Fixed PH's badly added libz and libbz2 support.
1578        - Fixed a problem with static linking.
1579        - Added pcredemo. [But later removed - see 7 below.]
1580        - Fixed dftables problem and added an option.
1581        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1582            HAVE_LONG_LONG.
1583        - Added readline support for pcretest.
1584        - Added an listing of the option settings after cmake has run.
1585    
1586    4.  A user submitted a patch to Makefile that makes it easy to create
1587        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1588        Makefile.am that cause it to include this special target, without
1589        affecting anything else. Note that the same mingw target plus all
1590        the other distribution libraries and programs are now supported
1591        when configuring with CMake (see 6 below) instead of with
1592        Configure/Make.
1593    
1594    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1595        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1596        exported in the Windows port". It has not yet been confirmed that the patch
1597        solves the problem, but it does no harm.
1598    
1599    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1600        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1601        with CMake, and also correct the comment about stack recursion.
1602    
1603    7.  Remove the automatic building of pcredemo from the ./configure system and
1604        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1605        of a program that users should build themselves after PCRE is installed, so
1606        building it automatically is not really right. What is more, it gave
1607        trouble in some build environments.
1608    
1609    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1610    
1611    
1612    Version 7.5 10-Jan-08
1613    ---------------------
1614    
1615    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1616        values in parens when parsing an RE using the C++ wrapper."
1617    
1618    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1619        Characters greater than 255 were excluded from the class instead of being
1620        included.
1621    
1622    3.  The same bug as (2) above applied to negated POSIX classes such as
1623        [:^space:].
1624    
1625    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1626        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1627        I have changed it.
1628    
1629    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1630        first named subpattern) and a construct such as (?&a) would reference the
1631        first named subpattern whose name started with "a" (in other words, the
1632        length check was missing). Both these problems are fixed. "Subpattern name
1633        expected" is now given for (?&) (a zero-length name), and this patch also
1634        makes it give the same error for \k'' (previously it complained that that
1635        was a reference to a non-existent subpattern).
1636    
1637    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1638        this is right because (?- can be followed by option settings as well as by
1639        digits. I have, however, made the messages clearer.
1640    
1641    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1642        than the number used in the conditional) now cause a compile-time error.
1643        This is actually not compatible with Perl, which accepts such patterns, but
1644        treats the conditional as always being FALSE (as PCRE used to), but it
1645        seems to me that giving a diagnostic is better.
1646    
1647    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1648        and messages.
1649    
1650    9.  Fix two occurrences of "backslash" in comments that should have been
1651        "backspace".
1652    
1653    10. Remove two redundant lines of code that can never be obeyed (their function
1654        was moved elsewhere).
1655    
1656    11. The program that makes PCRE's Unicode character property table had a bug
1657        which caused it to generate incorrect table entries for sequences of
1658        characters that have the same character type, but are in different scripts.
1659        It amalgamated them into a single range, with the script of the first of
1660        them. In other words, some characters were in the wrong script. There were
1661        thirteen such cases, affecting characters in the following ranges:
1662    
1663          U+002b0 - U+002c1
1664          U+0060c - U+0060d
1665          U+0061e - U+00612
1666          U+0064b - U+0065e
1667          U+0074d - U+0076d
1668          U+01800 - U+01805
1669          U+01d00 - U+01d77
1670          U+01d9b - U+01dbf
1671          U+0200b - U+0200f
1672          U+030fc - U+030fe
1673          U+03260 - U+0327f
1674          U+0fb46 - U+0fbb1
1675          U+10450 - U+1049d
1676    
1677    12. The -o option (show only the matching part of a line) for pcregrep was not
1678        compatible with GNU grep in that, if there was more than one match in a
1679        line, it showed only the first of them. It now behaves in the same way as
1680        GNU grep.
1681    
1682    13. If the -o and -v options were combined for pcregrep, it printed a blank
1683        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1684        does the same. The return code can be used to tell if there were any
1685        non-matching lines.
1686    
1687    14. Added --file-offsets and --line-offsets to pcregrep.
1688    
1689    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1690        infinitely looping recursion. The bug was that positive lookaheads were not
1691        being skipped when checking for a possible empty match (negative lookaheads
1692        and both kinds of lookbehind were skipped).
1693    
1694    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1695        inclusion of <windows.h> to before rather than after the definition of
1696        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1697    
1698    17. Specifying a possessive quantifier with a specific limit for a Unicode
1699        character property caused pcre_compile() to compile bad code, which led at
1700        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1701        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1702        caused the error; without that there was no problem.
1703    
1704    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1705    
1706    19. Added --enable-pcretest-libreadline.
1707    
1708    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1709        RE::GlobalReplace(). As a result, the number of replacements returned was
1710        double what it should be. I removed one of the increments, but Craig sent a
1711        later patch that removed the other one (the right fix) and added unit tests
1712        that check the return values (which was not done before).
1713    
1714    21. Several CMake things:
1715    
1716        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1717            the names libpcre and libpcreposix, not just pcre and pcreposix.
1718    
1719        (2) The above change means that pcretest and pcregrep are now correctly
1720            linked with the newly-built libraries, not previously installed ones.
1721    
1722        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1723    
1724    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1725        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1726        UTF-8 newline character). The key issue is that the pattern starts .*;
1727        this means that the match must be either at the beginning, or after a
1728        newline. The bug was in the code for advancing after a failed match and
1729        checking that the new position followed a newline. It was not taking
1730        account of UTF-8 characters correctly.
1731    
1732    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1733        character classes. PCRE was not treating the sequence [:...:] as a
1734        character class unless the ... were all letters. Perl, however, seems to
1735        allow any characters between [: and :], though of course it rejects as
1736        unknown any "names" that contain non-letters, because all the known class
1737        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1738        for example, whereas PCRE did not - it did not recognize a POSIX character
1739        class. This seemed a bit dangerous, so the code has been changed to be
1740        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1741        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1742        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1743        Perl does, and where it didn't before.
1744    
1745    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1746        Windows environments %n is disabled by default.
1747    
1748    
1749    Version 7.4 21-Sep-07
1750    ---------------------
1751    
1752    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1753        means that a class such as [\s] counted as "explicit reference to CR or
1754        LF". That isn't really right - the whole point of the change was to try to
1755        help when there was an actual mention of one of the two characters. So now
1756        the change happens only if \r or \n (or a literal CR or LF) character is
1757        encountered.
1758    
1759    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1760        of both had grown to the point where there were only 3 bits left.
1761        Fortunately, there was spare space in the data structure, and so I have
1762        moved the internal flags into a new 16-bit field to free up more option
1763        bits.
1764    
1765    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1766        but did not set the internal JCHANGED flag - either of these is enough to
1767        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1768        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1769        start sets both bits.
1770    
1771    4.  Added options (at build time, compile time, exec time) to change \R from
1772        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1773    
1774    5.  doc/pcresyntax.html was missing from the distribution.
1775    
1776    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1777        compatibility, even though it is no longer used.
1778    
1779    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1780        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1781        windows.h file is present (where different names are used). [This was
1782        reversed later after testing - see 16 below.]
1783    
1784    8.  Changed all #include <config.h> to #include "config.h". There were also
1785        some further <pcre.h> cases that I changed to "pcre.h".
1786    
1787    9.  When pcregrep was used with the --colour option, it missed the line ending
1788        sequence off the lines that it output.
1789    
1790    10. It was pointed out to me that arrays of string pointers cause lots of
1791        relocations when a shared library is dynamically loaded. A technique of
1792        using a single long string with a table of offsets can drastically reduce
1793        these. I have refactored PCRE in four places to do this. The result is
1794        dramatic:
1795    
1796          Originally:                          290
1797          After changing UCP table:            187
1798          After changing error message table:   43
1799          After changing table of "verbs"       36
1800          After changing table of Posix names   22
1801    
1802        Thanks to the folks working on Gregex for glib for this insight.
1803    
1804    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1805        unicode-properties was also set.
1806    
1807    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1808    
1809    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1810        checked only for CRLF.
1811    
1812    14. Added casts to pcretest.c to avoid compiler warnings.
1813    
1814    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1815    
1816    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1817        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1818        entirely. This removes changes made in 7 above.
1819    
1820    17. The CMake files have been updated, and there is now more information about
1821        building with CMake in the NON-UNIX-USE document.
1822    
1823    
1824    Version 7.3 28-Aug-07
1825    ---------------------
1826    
1827     1. In the rejigging of the build system that eventually resulted in 7.1, the
1828        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1829        brackets there is not right, since it causes compilers to look for an
1830        installed pcre.h, not the version that is in the source that is being
1831        compiled (which of course may be different). I have changed it back to:
1832    
1833          #include "pcre.h"
1834    
1835        I have a vague recollection that the change was concerned with compiling in
1836        different directories, but in the new build system, that is taken care of
1837        by the VPATH setting the Makefile.
1838    
1839     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1840        when the subject happened to end in the byte 0x85 (e.g. if the last
1841        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1842        characters but of course it shouldn't be taken as a newline when it is part
1843        of another character. The bug was that, for an unlimited repeat of . in
1844        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1845        characters when looking for a newline.
1846    
1847     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1848    
1849     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1850        in debug output.
1851    
1852     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1853        long printing in the pcrecpp unittest when running under MinGW.
1854    
1855     6. ESC_K was left out of the EBCDIC table.
1856    
1857     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1858        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1859        limit also applies to "virtual nesting" when a pattern is recursive, and in
1860        this case 1000 isn't so big. I have been able to remove this limit at the
1861        expense of backing off one optimization in certain circumstances. Normally,
1862        when pcre_exec() would call its internal match() function recursively and
1863        immediately return the result unconditionally, it uses a "tail recursion"
1864        feature to save stack. However, when a subpattern that can match an empty
1865        string has an unlimited repetition quantifier, it no longer makes this
1866        optimization. That gives it a stack frame in which to save the data for
1867        checking that an empty string has been matched. Previously this was taken
1868        from the 1000-entry workspace that had been reserved. So now there is no
1869        explicit limit, but more stack is used.
1870    
1871     8. Applied Daniel's patches to solve problems with the import/export magic
1872        syntax that is required for Windows, and which was going wrong for the
1873        pcreposix and pcrecpp parts of the library. These were overlooked when this
1874        problem was solved for the main library.
1875    
1876     9. There were some crude static tests to avoid integer overflow when computing
1877        the size of patterns that contain repeated groups with explicit upper
1878        limits. As the maximum quantifier is 65535, the maximum group length was
1879        set at 30,000 so that the product of these two numbers did not overflow a
1880        32-bit integer. However, it turns out that people want to use groups that
1881        are longer than 30,000 bytes (though not repeat them that many times).
1882        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1883        made it possible to implement the integer overflow checks in a much more
1884        dynamic way, which I have now done. The artificial limitation on group
1885        length has been removed - we now have only the limit on the total length of
1886        the compiled pattern, which depends on the LINK_SIZE setting.
1887    
1888    10. Fixed a bug in the documentation for get/copy named substring when
1889        duplicate names are permitted. If none of the named substrings are set, the
1890        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1891        empty string.
1892    
1893    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1894        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1895        because the ] is interpreted as the first data character and the
1896        terminating ] is not found. PCRE has been made compatible with Perl in this
1897        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1898        cause memory overwriting.
1899    
1900    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1901        string has been matched (to stop an infinite loop). It was not recognizing
1902        a conditional subpattern that could match an empty string if that
1903        subpattern was within another subpattern. For example, it looped when
1904        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1905        condition was not nested. This bug has been fixed.
1906    
1907    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1908        past the start of the subject in the presence of bytes with the top bit
1909        set, for example "\x8aBCD".
1910    
1911    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1912        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1913    
1914    14. Optimized (?!) to (*FAIL).
1915    
1916    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1917        This restricts code points to be within the range 0 to 0x10FFFF, excluding
1918        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1919        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1920        does: it's just the validity check that is more restrictive.
1921    
1922    16. Inserted checks for integer overflows during escape sequence (backslash)
1923        processing, and also fixed erroneous offset values for syntax errors during
1924        backslash processing.
1925    
1926    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1927        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1928    
1929    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1930        caused an overrun.
1931    
1932    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1933        something other than just ASCII characters) inside a group that had an
1934        unlimited repeat caused a loop at compile time (while checking to see
1935        whether the group could match an empty string).
1936    
1937    20. Debugging a pattern containing \p or \P could cause a crash. For example,
1938        [\P{Any}] did so. (Error in the code for printing property names.)
1939    
1940    21. An orphan \E inside a character class could cause a crash.
1941    
1942    22. A repeated capturing bracket such as (A)? could cause a wild memory
1943        reference during compilation.
1944    
1945    23. There are several functions in pcre_compile() that scan along a compiled
1946        expression for various reasons (e.g. to see if it's fixed length for look
1947        behind). There were bugs in these functions when a repeated \p or \P was
1948        present in the pattern. These operators have additional parameters compared
1949        with \d, etc, and these were not being taken into account when moving along
1950        the compiled data. Specifically:
1951    
1952        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1953            length.
1954    
1955        (b) An item such as \pL+ within a repeated group could cause crashes or
1956            loops.
1957    
1958        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1959            "reference to non-existent subpattern" error.
1960    
1961        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1962    
1963    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1964        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1965    
1966    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1967    
1968    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1969        character were causing crashes (broken optimization).
1970    
1971    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1972        \p or \P) caused a compile-time loop.
1973    
1974    28. More problems have arisen in unanchored patterns when CRLF is a valid line
1975        break. For example, the unstudied pattern [\r\n]A does not match the string
1976        "\r\nA" because change 7.0/46 below moves the current point on by two
1977        characters after failing to match at the start. However, the pattern \nA
1978        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1979        the same is true. There doesn't seem any very clean way out of this, but
1980        what I have chosen to do makes the common cases work: PCRE now takes note
1981        of whether there can be an explicit match for \r or \n anywhere in the
1982        pattern, and if so, 7.0/46 no longer applies. As part of this change,
1983        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1984        pattern has explicit CR or LF references.
1985    
1986    29. Added (*CR) etc for changing newline setting at start of pattern.
1987    
1988    
1989    Version 7.2 19-Jun-07
1990    ---------------------
1991    
1992     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1993        which is apparently normally available under Windows.
1994    
1995     2. Re-jig the pcregrep tests with different newline settings in an attempt
1996        to make them independent of the local environment's newline setting.
1997    
1998     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1999    
2000     4. Some of the "internals" tests were previously cut out when the link size
2001        was not 2, because the output contained actual offsets. The recent new
2002        "Z" feature of pcretest means that these can be cut out, making the tests
2003        usable with all link sizes.
2004    
2005     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2006        stack recursion. This gives a massive performance boost under BSD, but just
2007        a small improvement under Linux. However, it saves one field in the frame
2008        in all cases.
2009    
2010     6. Added more features from the forthcoming Perl 5.10:
2011    
2012        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2013            recursion call. It refers to the nth most recently opened parentheses.
2014    
2015        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2016            to be opened parentheses.
2017    
2018        (c) Conditions that refer to capturing parentheses can be specified
2019            relatively, for example, (?(-2)... or (?(+3)...
2020    
2021        (d) \K resets the start of the current match so that everything before
2022            is not part of it.
2023    
2024        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2025    
2026        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2027            reference syntax.
2028    
2029        (g) (?| introduces a group in which the numbering of parentheses in each
2030            alternative starts with the same number.
2031    
2032        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2033    
2034     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2035        PCRE_INFO_JCHANGED.
2036    
2037     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2038        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2039        for detecting groups that can match an empty string.
2040    
2041     9. A pattern with a very large number of alternatives (more than several
2042        hundred) was running out of internal workspace during the pre-compile
2043        phase, where pcre_compile() figures out how much memory will be needed. A
2044        bit of new cunning has reduced the workspace needed for groups with
2045        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2046        workspace instead of running out of the 4096 that are available.
2047    
2048    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2049    
2050    11. Applied patch from Google to remove an optimization that didn't quite work.
2051        The report of the bug said:
2052    
2053          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2054          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2055          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2056    
2057    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2058        it matched the wrong number of bytes.
2059    
2060    
2061    Version 7.1 24-Apr-07
2062    ---------------------
2063    
2064     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2065        that is more "standard", making use of automake and other Autotools. There
2066        is some re-arrangement of the files and adjustment of comments consequent
2067        on this.
2068    
2069     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2070        for recursive directory scanning broke on some systems because the files
2071        are not scanned in any specific order and on different systems the order
2072        was different. A call to "sort" has been inserted into RunGrepTest for the
2073        approprate test as a short-term fix. In the longer term there may be an
2074        alternative.
2075    
2076     3. I had an email from Eric Raymond about problems translating some of PCRE's
2077        man pages to HTML (despite the fact that I distribute HTML pages, some
2078        people do their own conversions for various reasons). The problems
2079        concerned the use of low-level troff macros .br and .in. I have therefore
2080        removed all such uses from the man pages (some were redundant, some could
2081        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2082        HTML has been updated to handle .nf/.fi and to complain if it encounters
2083        .br or .in.
2084    
2085     4. Updated comments in configure.ac that get placed in config.h.in and also
2086        arranged for config.h to be included in the distribution, with the name
2087        config.h.generic, for the benefit of those who have to compile without
2088        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2089    
2090     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2091        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2092        makevp.bat for the current PCRE, using the additional files
2093        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2094    
2095     6. A Windows user reported a minor discrepancy with test 2, which turned out
2096        to be caused by a trailing space on an input line that had got lost in his
2097        copy. The trailing space was an accident, so I've just removed it.
2098    
2099     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2100        that is needed.
2101    
2102     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2103        as "const" (a) because they are and (b) because it helps the PHP
2104        maintainers who have recently made a script to detect big data structures
2105        in the php code that should be moved to the .rodata section. I remembered
2106        to update Builducptable as well, so it won't revert if ucptable.h is ever
2107        re-created.
2108    
2109     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2110        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2111        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2112        support is not required. This saves 1.5-2K of code, which is important in
2113        some applications.
2114    
2115        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2116        so as not to refer to the tables, even though these functions will never be
2117        called when UTF-8 support is disabled. Otherwise there are problems with a
2118        shared library.
2119    
2120    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2121    
2122        (a) It was defining its arguments as char * instead of void *.
2123    
2124        (b) It was assuming that all moves were upwards in memory; this was true
2125            a long time ago when I wrote it, but is no longer the case.
2126    
2127        The emulated memove() is provided for those environments that have neither
2128        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2129        is clearly not the case, as these two bugs were recently reported.
2130    
2131    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2132        and Detrail to create the HTML documentation, the .txt form of the man
2133        pages, and it removes trailing spaces from listed files. It also creates
2134        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2135        case, it wraps all the #defines with #ifndefs. This script should be run
2136        before "make dist".
2137    
2138    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2139        with Unicode property support.
2140    
2141        (a) For a maximizing quantifier, if the two different cases of the
2142            character were of different lengths in their UTF-8 codings (there are
2143            some cases like this - I found 11), and the matching function had to
2144            back up over a mixture of the two cases, it incorrectly assumed they
2145            were both the same length.
2146    
2147        (b) When PCRE was configured to use the heap rather than the stack for
2148            recursion during matching, it was not correctly preserving the data for
2149            the other case of a UTF-8 character when checking ahead for a match
2150            while processing a minimizing repeat. If the check also involved
2151            matching a wide character, but failed, corruption could cause an
2152            erroneous result when trying to check for a repeat of the original
2153            character.
2154    
2155    13. Some tidying changes to the testing mechanism:
2156    
2157        (a) The RunTest script now detects the internal link size and whether there
2158            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2159            values substituted by "configure". (The RunGrepTest script already did
2160            this for UTF-8.) The configure.ac script no longer substitutes the
2161            relevant variables.
2162    
2163        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2164            with length and offset values. This means that the output is different
2165            for different internal link sizes. Test 2 is skipped for link sizes
2166            other than 2 because of this, bypassing the problem. Unfortunately,
2167            there was also a test in test 3 (the locale tests) that used /B and
2168            failed for link sizes other than 2. Rather than cut the whole test out,
2169            I have added a new /Z option to pcretest that replaces the length and
2170            offset values with spaces. This is now used to make test 3 independent
2171            of link size. (Test 2 will be tidied up later.)
2172    
2173    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2174        segmentation fault instead of returning the appropriate error message.
2175    
2176    15. In multiline mode when the newline sequence was set to "any", the pattern
2177        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2178        This doesn't seem right; it now treats the CRLF combination as the line
2179        ending, and so does not match in that case. It's only a pattern such as ^$
2180        that would hit this one: something like ^ABC$ would have failed after \r
2181        and then tried again after \r\n.
2182    
2183    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2184        in an attempt to make files that differ only in their line terminators
2185        compare equal. This works on Linux.
2186    
2187    17. Under certain error circumstances pcregrep might try to free random memory
2188        as it exited. This is now fixed, thanks to valgrind.
2189    
2190    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2191        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2192        was because its rules for how to advance for /g after matching an empty
2193        string at the end of a line did not allow for this case. They now check for
2194        it specially.
2195    
2196    20. pcretest is supposed to handle patterns and data of any length, by
2197        extending its buffers when necessary. It was getting this wrong when the
2198        buffer for a data line had to be extended.
2199    
2200    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2201        CRLF as a newline sequence.
2202    
2203    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2204        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2205        I have nevertheless tidied it up.
2206    
2207    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2208    
2209    24. Added a man page for pcre-config.
2210    
2211    
2212    Version 7.0 19-Dec-06
2213    ---------------------
2214    
2215     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2216        moving to gcc 4.1.1.
2217    
2218     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2219        sys/time.h, which is documented as needed for this function. It doesn't
2220        seem to matter on Linux, but it showed up on some releases of OS X.
2221    
2222     3. It seems that there are systems where bytes whose values are greater than
2223        127 match isprint() in the "C" locale. The "C" locale should be the
2224        default when a C program starts up. In most systems, only ASCII printing
2225        characters match isprint(). This difference caused the output from pcretest
2226        to vary, making some of the tests fail. I have changed pcretest so that:
2227    
2228        (a) When it is outputting text in the compiled version of a pattern, bytes
2229            other than 32-126 are always shown as hex escapes.
2230    
2231        (b) When it is outputting text that is a matched part of a subject string,
2232            it does the same, unless a different locale has been set for the match
2233            (using the /L modifier). In this case, it uses isprint() to decide.
2234    
2235     4. Fixed a major bug that caused incorrect computation of the amount of memory
2236        required for a compiled pattern when options that changed within the
2237        pattern affected the logic of the preliminary scan that determines the
2238        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2239        that the computed length was too small. The symptoms of this bug were
2240        either the PCRE error "internal error: code overflow" from pcre_compile(),
2241        or a glibc crash with a message such as "pcretest: free(): invalid next
2242        size (fast)". Examples of patterns that provoked this bug (shown in
2243        pcretest format) are:
2244    
2245          /(?-x: )/x
2246          /(?x)(?-x: \s*#\s*)/
2247          /((?i)[\x{c0}])/8
2248          /(?i:[\x{c0}])/8
2249    
2250        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2251        is now done differently.
2252    
2253     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2254        wrapper classes; (b) implement a new function in the C++ scanner that is
2255        more efficient than the old way of doing things because it avoids levels of
2256        recursion in the regex matching; (c) add a paragraph to the documentation
2257        for the FullMatch() function.
2258    
2259     6. The escape sequence \n was being treated as whatever was defined as
2260        "newline". Not only was this contrary to the documentation, which states
2261        that \n is character 10 (hex 0A), but it also went horribly wrong when
2262        "newline" was defined as CRLF. This has been fixed.
2263    
2264     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2265        was being set to -1 for the "end of line" case (supposedly a value that no
2266        character can have). Though this value is never used (the check for end of
2267        line is "zero bytes in current character"), it caused compiler complaints.
2268        I've changed it to 0xffffffff.
2269    
2270     8. In pcre_version.c, the version string was being built by a sequence of
2271        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2272        string (as it is for production releases) called a macro with an empty
2273        argument. The C standard says the result of this is undefined. The gcc
2274        compiler treats it as an empty string (which was what was wanted) but it is
2275        reported that Visual C gives an error. The source has been hacked around to
2276        avoid this problem.
2277    
2278     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2279        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2280        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2281        of them did).
2282    
2283    10. Originally, pcretest opened its input and output without "b"; then I was
2284        told that "b" was needed in some environments, so it was added for release
2285        5.0 to both the input and output. (It makes no difference on Unix-like
2286        systems.) Later I was told that it is wrong for the input on Windows. I've
2287        now abstracted the modes into two macros, to make it easier to fiddle with
2288        them, and removed "b" from the input mode under Windows.
2289    
2290    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2291    
2292    12. Added -help and --help to pcretest as an official way of being reminded
2293        of the options.
2294    
2295    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2296        and pcrecpp.cc because they annoy compilers at high warning levels.
2297    
2298    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2299    
2300    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2301        scripts are not C programs :-) and which was not noticed because it works
2302        on Linux.
2303    
2304    16. pcretest is supposed to handle any length of pattern and data line (as one
2305        line or as a continued sequence of lines) by extending its input buffer if
2306        necessary. This feature was broken for very long pattern lines, leading to
2307        a string of junk being passed to pcre_compile() if the pattern was longer
2308        than about 50K.
2309    
2310    17. I have done a major re-factoring of the way pcre_compile() computes the
2311        amount of memory needed for a compiled pattern. Previously, there was code
2312        that made a preliminary scan of the pattern in order to do this. That was
2313        OK when PCRE was new, but as the facilities have expanded, it has become
2314        harder and harder to keep it in step with the real compile phase, and there
2315        have been a number of bugs (see for example, 4 above). I have now found a
2316        cunning way of running the real compile function in a "fake" mode that
2317        enables it to compute how much memory it would need, while actually only
2318        ever using a few hundred bytes of working memory and without too many
2319        tests of the mode. This should make future maintenance and development
2320        easier. A side effect of this work is that the limit of 200 on the nesting
2321        depth of parentheses has been removed (though this was never a serious
2322        limitation, I suspect). However, there is a downside: pcre_compile() now
2323        runs more slowly than before (30% or more, depending on the pattern). I
2324        hope this isn't a big issue. There is no effect on runtime performance.
2325    
2326    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2327        newline (only possible for the last line of a file) and it was a
2328        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2329    
2330    19. Added additional timing features to pcretest. (1) The -tm option now times
2331        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2332        separate command line item, by a number that specifies the number of
2333        repeats to use when timing. The default is 50000; this gives better
2334        precision, but takes uncomfortably long for very large patterns.
2335    
2336    20. Extended pcre_study() to be more clever in cases where a branch of a
2337        subpattern has no definite first character. For example, (a*|b*)[cd] would
2338        previously give no result from pcre_study(). Now it recognizes that the
2339        first character must be a, b, c, or d.
2340    
2341    21. There was an incorrect error "recursive call could loop indefinitely" if
2342        a subpattern (or the entire pattern) that was being tested for matching an
2343        empty string contained only one non-empty item after a nested subpattern.
2344        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2345        incorrectly, because the \d was being skipped in the check.
2346    
2347    22. The pcretest program now has a new pattern option /B and a command line
2348        option -b, which is equivalent to adding /B to every pattern. This causes
2349        it to show the compiled bytecode, without the additional information that
2350        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2351        is the same as /B/I).
2352    
2353    23. A new optimization is now able automatically to treat some sequences such
2354        as a*b as a*+b. More specifically, if something simple (such as a character
2355        or a simple class like \d) has an unlimited quantifier, and is followed by
2356        something that cannot possibly match the quantified thing, the quantifier
2357        is automatically "possessified".
2358    
2359    24. A recursive reference to a subpattern whose number was greater than 39
2360        went wrong under certain circumstances in UTF-8 mode. This bug could also
2361        have affected the operation of pcre_study().
2362    
2363    25. Realized that a little bit of performance could be had by replacing
2364        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2365    
2366    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2367    
2368    27. Possessive quantifiers such as a++ were previously implemented by turning
2369        them into atomic groups such as ($>a+). Now they have their own opcodes,
2370        which improves performance. This includes the automatically created ones
2371        from 23 above.
2372    
2373    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2374        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2375        the first matched character to be a colon. This applied both to named and
2376        numbered groups.
2377    
2378    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2379    
2380    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2381        building PCRE on Windows easier, so I have included it in the distribution.
2382    
2383    31. There is now a check in pcretest against a ridiculously large number being
2384        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2385        loop, the loop is abandoned.
2386    
2387    32. Forward references to subpatterns in conditions such as (?(2)...) where
2388        subpattern 2 is defined later cause pcre_compile() to search forwards in
2389        the pattern for the relevant set of parentheses. This search went wrong
2390        when there were unescaped parentheses in a character class, parentheses
2391        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2392    
2393    33. "Subroutine" calls and backreferences were previously restricted to
2394        referencing subpatterns earlier in the regex. This restriction has now
2395        been removed.
2396    
2397    34. Added a number of extra features that are going to be in Perl 5.10. On the
2398        whole, these are just syntactic alternatives for features that PCRE had
2399        previously implemented using the Python syntax or my own invention. The
2400        other formats are all retained for compatibility.
2401    
2402        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2403            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2404            also .NET compatible.
2405    
2406        (b) A recursion or subroutine call to a named group can now be defined as
2407            (?&name) as well as (?P>name).
2408    
2409        (c) A backreference to a named group can now be defined as \k<name> or
2410            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2411            5.10, are also .NET compatible.
2412    
2413        (d) A conditional reference to a named group can now use the syntax
2414            (?(<name>) or (?('name') as well as (?(name).
2415    
2416        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2417            groups (named and numbered) that are never evaluated inline, but can be
2418            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2419            is always false. There may be only one alternative in such a group.
2420    
2421        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2422            as the simple (?(R). The condition is true only if the most recent
2423            recursion is that of the given number or name. It does not search out
2424            through the entire recursion stack.
2425    
2426        (g) The escape \gN or \g{N} has been added, where N is a positive or
2427            negative number, specifying an absolute or relative reference.
2428    
2429    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2430        some "unreachable code" warnings.
2431    
2432    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2433        things, this adds five new scripts.
2434    
2435    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2436        There were also incompatibilities regarding the handling of \Q..\E inside
2437        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2438        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2439    
2440    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2441        matches an empty string, and forcibly breaks the loop. There were bugs in
2442        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2443        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2444        separate and independent bugs (that affected different cases) have been
2445        fixed.
2446    
2447    39. Refactored the code to abolish the use of different opcodes for small
2448        capturing bracket numbers. This is a tidy that I avoided doing when I
2449        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2450        The new approach is not only tidier, it makes it possible to reduce the
2451        memory needed to fix the previous bug (38).
2452    
2453    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2454        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2455        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2456        mode.
2457    
2458    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2459        report.
2460    
2461    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2462        copy construction and assignment in the C++ wrapper.
2463    
2464    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2465        couple of bugs that could have given wrong results in the "--newline=crlf"
2466        case.
2467    
2468    44. Added a number of casts and did some reorganization of signed/unsigned int
2469        variables following suggestions from Dair Grant. Also renamed the variable
2470        "this" as "item" because it is a C++ keyword.
2471    
2472    45. Arranged for dftables to add
2473    
2474          #include "pcre_internal.h"
2475    
2476        to pcre_chartables.c because without it, gcc 4.x may remove the array
2477        definition from the final binary if PCRE is built into a static library and
2478        dead code stripping is activated.
2479    
2480    46. For an unanchored pattern, if a match attempt fails at the start of a
2481        newline sequence, and the newline setting is CRLF or ANY, and the next two
2482        characters are CRLF, advance by two characters instead of one.
2483    
2484    
2485    Version 6.7 04-Jul-06
2486    ---------------------
2487    
2488     1. In order to handle tests when input lines are enormously long, pcretest has
2489        been re-factored so that it automatically extends its buffers when
2490        necessary. The code is crude, but this _is_ just a test program. The
2491        default size has been increased from 32K to 50K.
2492    
2493     2. The code in pcre_study() was using the value of the re argument before
2494        testing it for NULL. (Of course, in any sensible call of the function, it
2495        won't be NULL.)
2496    
2497     3. The memmove() emulation function in pcre_internal.h, which is used on
2498        systems that lack both memmove() and bcopy() - that is, hardly ever -
2499        was missing a "static" storage class specifier.
2500    
2501     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2502        containing an extended class (one that cannot be represented by a bitmap
2503        because it contains high-valued characters or Unicode property items, e.g.
2504        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2505        pattern, but PCRE should not loop if you do not (it no longer does).
2506        [Detail: two cases were found: (a) a repeated subpattern containing an
2507        extended class; (b) a recursive reference to a subpattern that followed a
2508        previous extended class. It wasn't skipping over the extended class
2509        correctly when UTF-8 mode was not set.]
2510    
2511     5. A negated single-character class was not being recognized as fixed-length
2512        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2513        compile error "lookbehind assertion is not fixed length".
2514    
2515     6. The RunPerlTest auxiliary script was showing an unexpected difference
2516        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2517        write a Perl script that can interpret lines of an input file either as
2518        byte characters or as UTF-8, which is what "perltest" was being required to
2519        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2520        can't do is switch easily at run time between having the "use utf8;" pragma
2521        or not. In the end, I fudged it by using the RunPerlTest script to insert
2522        "use utf8;" explicitly for the UTF-8 tests.
2523    
2524     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2525        the end of the subject string, contrary to the documentation and to what
2526        Perl does. This was true of both matching functions. Now it matches only at
2527        the start of the subject and immediately after *internal* newlines.
2528    
2529     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2530        a pointer to an int instead of a pointer to an unsigned long int. This
2531        caused problems on 64-bit systems.
2532    
2533     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2534        instance of the 'standard' template library not being so standard".
2535    
2536    10. There was no check on the number of named subpatterns nor the maximum
2537        length of a subpattern name. The product of these values is used to compute
2538        the size of the memory block for a compiled pattern. By supplying a very
2539        long subpattern name and a large number of named subpatterns, the size
2540        computation could be caused to overflow. This is now prevented by limiting
2541        the length of names to 32 characters, and the number of named subpatterns
2542        to 10,000.
2543    
2544    11. Subpatterns that are repeated with specific counts have to be replicated in
2545        the compiled pattern. The size of memory for this was computed from the
2546        length of the subpattern and the repeat count. The latter is limited to
2547        65535, but there was no limit on the former, meaning that integer overflow
2548        could in principle occur. The compiled length of a repeated subpattern is
2549        now limited to 30,000 bytes in order to prevent this.
2550    
2551    12. Added the optional facility to have named substrings with the same name.
2552    
2553    13. Added the ability to use a named substring as a condition, using the
2554        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2555        are numbers (not recommended). Forward references are permitted.
2556    
2557    14. Added forward references in named backreferences (if you see what I mean).
2558    
2559    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2560        pattern could run off the end of the subject. For example, the pattern
2561        "(?s)(.{1,5})"8 did this with the subject "ab".
2562    
2563    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2564        PCRE_CASELESS was set when matching characters that were quantified with ?
2565        or *.
2566    
2567    17. A character class other than a single negated character that had a minimum
2568        but no maximum quantifier - for example [ab]{6,} - was not handled
2569        correctly by pce_dfa_exec(). It would match only one character.
2570    
2571    18. A valid (though odd) pattern that looked like a POSIX character
2572        class but used an invalid character after [ (for example [[,abc,]]) caused
2573        pcre_compile() to give the error "Failed: internal error: code overflow" or
2574        in some cases to crash with a glibc free() error. This could even happen if
2575        the pattern terminated after [[ but there just happened to be a sequence of
2576        letters, a binary zero, and a closing ] in the memory that followed.
2577    
2578    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2579        over the years. Originally (before any Unicode support), just the bottom 8
2580        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2581        output from "man perlunicode" includes this:
2582    
2583          The regular expression compiler produces polymorphic opcodes.  That
2584          is, the pattern adapts to the data and automatically switches to
2585          the Unicode character scheme when presented with Unicode data--or
2586          instead uses a traditional byte scheme when presented with byte
2587          data.
2588    
2589        Sadly, a wide octal escape does not cause a switch, and in a string with
2590        no other multibyte characters, these octal escapes are treated as before.
2591        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2592        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2593        Unicode string.
2594    
2595        I have not perpetrated such confusion in PCRE. Up till now, it took just
2596        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2597        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2598        translate to the appropriate multibyte character.
2599    
2600    29. Applied some refactoring to reduce the number of warnings from Microsoft
2601        and Borland compilers. This has included removing the fudge introduced
2602        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2603        a warning about an unused variable.
2604    
2605    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2606        characters since release 4.0, because Perl (from release 5.004) does not.
2607        [Or at least, is documented not to: some releases seem to be in conflict
2608        with the documentation.] However, when a pattern was studied with
2609        pcre_study() and all its branches started with \s, PCRE still included VT
2610        as a possible starting character. Of course, this did no harm; it just
2611        caused an unnecessary match attempt.
2612    
2613    22. Removed a now-redundant internal flag bit that recorded the fact that case
2614        dependency changed within the pattern. This was once needed for "required
2615        byte" processing, but is no longer used. This recovers a now-scarce options
2616        bit. Also moved the least significant internal flag bit to the most-
2617        significant bit of the word, which was not previously used (hangover from
2618        the days when it was an int rather than a uint) to free up another bit for
2619        the future.
2620    
2621    23. Added support for CRLF line endings as well as CR and LF. As well as the
2622        default being selectable at build time, it can now be changed at runtime
2623        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2624        specify that it is scanning data with non-default line endings.
2625    
2626    24. Changed the definition of CXXLINK to make it agree with the definition of
2627        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2628    
2629    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2630        recursions. This makes a big different to stack usage for some patterns.
2631    
2632    26. If a subpattern containing a named recursion or subroutine reference such
2633        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2634        the space required for the compiled pattern went wrong and gave too small a
2635        value. Depending on the environment, this could lead to "Failed: internal
2636        error: code overflow at offset 49" or "glibc detected double free or
2637        corruption" errors.
2638    
2639    27. Applied patches from Google (a) to support the new newline modes and (b) to
2640        advance over multibyte UTF-8 characters in GlobalReplace.
2641    
2642    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2643        difference for some implementation of PCRE in some Windows version.
2644    
2645    29. Added some extra testing facilities to pcretest:
2646    
2647        \q<number>   in a data line sets the "match limit" value
2648        \Q<number>   in a data line sets the "match recursion limt" value
2649        -S <number>  sets the stack size, where <number> is in megabytes
2650    
2651        The -S option isn't available for Windows.
2652    
2653    
2654    Version 6.6 06-Feb-06
2655    ---------------------
2656    
2657     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2658        in pcreposix.h. I have copied the definition from pcre.h.
2659    
2660     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2661        because pcre.h is no longer a built file.
2662    
2663     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2664        not normally included in the compiled code.
2665    
2666    
2667    Version 6.5 01-Feb-06
2668    ---------------------
2669    
2670     1. When using the partial match feature with pcre_dfa_exec(), it was not
2671        anchoring the second and subsequent partial matches at the new starting
2672        point. This could lead to incorrect results. For example, with the pattern
2673        /1234/, partially matching against "123" and then "a4" gave a match.
2674    
2675     2. Changes to pcregrep:
2676    
2677        (a) All non-match returns from pcre_exec() were being treated as failures
2678            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2679            error message is output. Some extra information is given for the
2680            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2681            probably the only errors that are likely to be caused by users (by
2682            specifying a regex that has nested indefinite repeats, for instance).
2683            If there are more than 20 of these errors, pcregrep is abandoned.
2684    
2685        (b) A binary zero was treated as data while matching, but terminated the
2686            output line if it was written out. This has been fixed: binary zeroes
2687            are now no different to any other data bytes.
2688    
2689        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2690            used to set a locale for matching. The --locale=xxxx long option has
2691            been added (no short equivalent) to specify a locale explicitly on the
2692            pcregrep command, overriding the environment variables.
2693    
2694        (d) When -B was used with -n, some line numbers in the output were one less
2695            than they should have been.
2696    
2697        (e) Added the -o (--only-matching) option.
2698    
2699        (f) If -A or -C was used with -c (count only), some lines of context were
2700            accidentally printed for the final match.
2701    
2702        (g) Added the -H (--with-filename) option.
2703    
2704        (h) The combination of options -rh failed to suppress file names for files
2705            that were found from directory arguments.
2706    
2707        (i) Added the -D (--devices) and -d (--directories) options.
2708    
2709        (j) Added the -F (--fixed-strings) option.
2710    
2711        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2712    
2713        (l) Added the --colo(u)r option.
2714    
2715        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2716            is not present by default.
2717    
2718     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2719        items such as (?R) or (?1), when the recursion could match a number of
2720        alternatives. If it matched one of the alternatives, but subsequently,
2721        outside the recursion, there was a failure, the code tried to back up into
2722        the recursion. However, because of the way PCRE is implemented, this is not
2723        possible, and the result was an incorrect result from the match.
2724    
2725        In order to prevent this happening, the specification of recursion has
2726        been changed so that all such subpatterns are automatically treated as
2727        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2728    
2729     4. I had overlooked the fact that, in some locales, there are characters for
2730        which isalpha() is true but neither isupper() nor islower() are true. In
2731        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2732        and ordfeminine) are like this. This affected the treatment of \w and \W
2733        when they appeared in character classes, but not when they appeared outside
2734        a character class. The bit map for "word" characters is now created
2735        separately from the results of isalnum() instead of just taking it from the
2736        upper, lower, and digit maps. (Plus the underscore character, of course.)
2737    
2738     5. The above bug also affected the handling of POSIX character classes such as
2739        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2740        permanent tables. Instead, the bit maps for such a class were previously
2741        created as the appropriate unions of the upper, lower, and digit bitmaps.
2742        Now they are created by subtraction from the [[:word:]] class, which has
2743        its own bitmap.
2744    
2745     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2746        It is created by subtracting the vertical space characters (\x09, \x0a,
2747        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2748        subtraction was done in the overall bitmap for a character class, meaning
2749        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2750        be recognized. This bug has been fixed.
2751    
2752     7. Patches from the folks at Google:
2753    
2754          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2755          real life, but is still worth protecting against".
2756    
2757          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2758          regular expressions".
2759    
2760          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2761          have it.
2762    
2763          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2764          "configure" and the latter not, in order to fix a problem somebody had
2765          with compiling the Arg class on HP-UX.
2766    
2767          (e) Improve the error-handling of the C++ wrapper a little bit.
2768    
2769          (f) New tests for checking recursion limiting.
2770    
2771     8. The pcre_memmove() function, which is used only if the environment does not
2772        have a standard memmove() function (and is therefore rarely compiled),
2773        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2774        returning a result (though PCRE never actually uses the result).
2775    
2776     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2777        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2778        returned instead of calling malloc() with an overflowing number that would
2779        most likely cause subsequent chaos.
2780    
2781    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2782    
2783    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2784        with this option is matched, the nmatch and pmatch options of regexec() are
2785        ignored.
2786    
2787    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2788        provided in case anyone wants to the the POSIX interface with UTF-8
2789        strings.
2790    
2791    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2792        C++ linking (needed for some HP-UX environments).
2793    
2794    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2795        (unused parameter) and in the pcre_printint() function (omitted "default"
2796        switch label when the default is to do nothing).
2797    
2798    15. Added some code to make it possible, when PCRE is compiled as a C++
2799        library, to replace subject pointers for pcre_exec() with a smart pointer
2800        class, thus making it possible to process discontinuous strings.
2801    
2802    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2803        much the same function. They were added by different people who were trying
2804        to make PCRE easy to compile on non-Unix systems. It has been suggested
2805        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2806        for compiling on Windows systems. I have therefore replaced it with
2807        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2808        defaults to "extern" for C or "extern C" for C++, which works fine on
2809        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2810        SCOPE with something explicit in config.h. In addition:
2811    
2812        (a) pcreposix.h still had just "extern" instead of either of these macros;
2813            I have replaced it with PCRE_DATA_SCOPE.
2814    
2815        (b) Functions such as _pcre_xclass(), which are internal to the library,
2816            but external in the C sense, all had PCRE_EXPORT in their definitions.
2817            This is apparently wrong for the Windows case, so I have removed it.
2818            (It makes no difference on Unix-like systems.)
2819    
2820    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2821        of recursive calls to match(). This is different to MATCH_LIMIT because
2822        that limits the total number of calls to match(), not all of which increase
2823        the depth of recursion. Limiting the recursion depth limits the amount of
2824        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2825        when PCRE is compiled, and changed at run time. A patch from Google adds
2826        this functionality to the C++ interface.
2827    
2828    18. Changes to the handling of Unicode character properties:
2829    
2830        (a) Updated the table to Unicode 4.1.0.
2831    
2832        (b) Recognize characters that are not in the table as "Cn" (undefined).
2833    
2834        (c) I revised the way the table is implemented to a much improved format
2835            which includes recognition of ranges. It now supports the ranges that
2836            are defined in UnicodeData.txt, and it also amalgamates other
2837            characters into ranges. This has reduced the number of entries in the
2838            table from around 16,000 to around 3,000, thus reducing its size
2839            considerably. I realized I did not need to use a tree structure after
2840            all - a binary chop search is just as efficient. Having reduced the
2841            number of entries, I extended their size from 6 bytes to 8 bytes to
2842            allow for more data.
2843    
2844        (d) Added support for Unicode script names via properties such as \p{Han}.
2845    
2846    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2847        matching that character.
2848    
2849    20. When matching a repeated Unicode property with a minimum greater than zero,
2850        (for example \pL{2,}), PCRE could look past the end of the subject if it
2851        reached it while seeking the minimum number of characters. This could
2852        happen only if some of the characters were more than one byte long, because
2853        there is a check for at least the minimum number of bytes.
2854    
2855    21. Refactored the implementation of \p and \P so as to be more general, to
2856        allow for more different types of property in future. This has changed the
2857        compiled form incompatibly. Anybody with saved compiled patterns that use
2858        \p or \P will have to recompile them.
2859    
2860    22. Added "Any" and "L&" to the supported property types.
2861    
2862    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2863        but give a compile time error if the value is greater than 0xff.
2864    
2865    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2866        accidentally not being installed or uninstalled.
2867    
2868    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2869        made were to insert the current release number. This seemed silly, because
2870        it made things harder for people building PCRE on systems that don't run
2871        "configure". I have turned pcre.h into a distributed file, no longer built
2872        by "configure", with the version identification directly included. There is
2873        no longer a pcre.h.in file.
2874    
2875        However, this change necessitated a change to the pcre-config script as
2876        well. It is built from pcre-config.in, and one of the substitutions was the
2877        release number. I have updated configure.ac so that ./configure now finds
2878        the release number by grepping pcre.h.
2879    
2880    26. Added the ability to run the tests under valgrind.
2881    
2882    
2883    Version 6.4 05-Sep-05
2884    ---------------------
2885    
2886     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2887        "--" to be printed when multiple files were scanned, even when none of the
2888        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2889        consider it to be a bug, and have restored the previous behaviour.
2890    
2891     2. A couple of code tidies to get rid of compiler warnings.
2892    
2893     3. The pcretest program used to cheat by referring to symbols in the library
2894        whose names begin with _pcre_. These are internal symbols that are not
2895        really supposed to be visible externally, and in some environments it is
2896        possible to suppress them. The cheating is now confined to including
2897        certain files from the library's source, which is a bit cleaner.
2898    
2899     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2900        file's purpose clearer.
2901    
2902     5. Reorganized pcre_ucp_findchar().
2903    
2904    
2905    Version 6.3 15-Aug-05
2906    ---------------------
2907    
2908     1. The file libpcre.pc.in did not have general read permission in the tarball.
2909    
2910     2. There were some problems when building without C++ support:
2911    
2912        (a) If C++ support was not built, "make install" and "make test" still
2913            tried to test it.
2914    
2915        (b) There were problems when the value of CXX was explicitly set. Some
2916            changes have been made to try to fix these, and ...
2917    
2918        (c) --disable-cpp can now be used to explicitly disable C++ support.
2919    
2920        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2921            backslash in a target when C++ was disabled. This confuses some
2922            versions of "make", apparently. Using an intermediate variable solves
2923            this. (Same for CPP_LOBJ.)
2924    
2925     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2926        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2927        necessary on certain architectures.
2928    
2929     4. Added a setting of -export-symbols-regex to the link command to remove
2930        those symbols that are exported in the C sense, but actually are local
2931        within the library, and not documented. Their names all begin with
2932        "_pcre_". This is not a perfect job, because (a) we have to except some
2933        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2934        available (and never for static libraries). I have made a note to try to
2935        find a way round (a) in the future.
2936    
2937    
2938    Version 6.2 01-Aug-05
2939    ---------------------
2940    
2941     1. There was no test for integer overflow of quantifier values. A construction
2942        such as {1111111111111111} would give undefined results. What is worse, if
2943        a minimum quantifier for a parenthesized subpattern overflowed and became
2944        negative, the calculation of the memory size went wrong. This could have
2945        led to memory overwriting.
2946    
2947     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2948    
2949     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2950        operating environments where this matters.
2951    
2952     4. Applied Giuseppe Maxia's patch to add additional features for controlling
2953        PCRE options from within the C++ wrapper.
2954    
2955     5. Named capturing subpatterns were not being correctly counted when a pattern
2956        was compiled. This caused two problems: (a) If there were more than 100
2957        such subpatterns, the calculation of the memory needed for the whole
2958        compiled pattern went wrong, leading to an overflow error. (b) Numerical
2959        back references of the form \12, where the number was greater than 9, were
2960        not recognized as back references, even though there were sufficient
2961        previous subpatterns.
2962    
2963     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2964        versions of gcc, e.g. 2.95.4.
2965    
2966    
2967    Version 6.1 21-Jun-05
2968    ---------------------
2969    
2970     1. There was one reference to the variable "posix" in pcretest.c that was not
2971        surrounded by "#if !defined NOPOSIX".
2972    
2973     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2974        the cross-check on the old pcre_info() function, for the benefit of the
2975        cut-down version of PCRE that is currently imported into Exim.
2976    
2977     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2978        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2979        patterns that start (?i) but I don't think that matters. The use of (?i) is
2980        just an example; this all applies to the other options as well.
2981    
2982     4. Since libtool seems to echo the compile commands it is issuing, the output
2983        from "make" can be reduced a bit by putting "@" in front of each libtool
2984        compile command.
2985    
2986     5. Patch from the folks at Google for configure.in to be a bit more thorough
2987        in checking for a suitable C++ installation before trying to compile the
2988        C++ stuff. This should fix a reported problem when a compiler was present,
2989        but no suitable headers.
2990    
2991     6. The man pages all had just "PCRE" as their title. I have changed them to
2992        be the relevant file name. I have also arranged that these names are
2993        retained in the file doc/pcre.txt, which is a concatenation in text format
2994        of all the man pages except the little individual ones for each function.
2995    
2996     7. The NON-UNIX-USE file had not been updated for the different set of source
2997        files that come with release 6. I also added a few comments about the C++
2998        wrapper.
2999    
3000    
3001    Version 6.0 07-Jun-05
3002    ---------------------
3003    
3004     1. Some minor internal re-organization to help with my DFA experiments.
3005    
3006     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3007        didn't matter for the library itself when fully configured, but did matter
3008        when compiling without UCP support, or within Exim, where the ucp files are
3009        not imported.
3010    
3011     3. Refactoring of the library code to split up the various functions into
3012        different source modules. The addition of the new DFA matching code (see
3013        below) to a single monolithic source would have made it really too
3014        unwieldy, quite apart from causing all the code to be include in a
3015        statically linked application, when only some functions are used. This is
3016        relevant even without the DFA addition now that patterns can be compiled in
3017        one application and matched in another.
3018    
3019        The downside of splitting up is that there have to be some external
3020        functions and data tables that are used internally in different modules of
3021        the library but which are not part of the API. These have all had their
3022        names changed to start with "_pcre_" so that they are unlikely to clash
3023        with other external names.
3024    
3025     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3026        a different (DFA) algorithm. Although it is slower than the original
3027        function, it does have some advantages for certain types of matching
3028        problem.
3029    
3030     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3031        including restarting after a partial match.
3032    
3033     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3034        defined when compiling for Windows was sent to me. I have put it into the
3035        code, though I have no means of testing or verifying it.
3036    
3037     7. Added the pcre_refcount() auxiliary function.
3038    
3039     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3040        match before or at the first newline in the subject string. In pcretest,
3041        the /f option on a pattern can be used to set this.
3042    
3043     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3044        would behave wrongly. This has been present in PCRE since release 4.0.
3045    
3046    10. A number of changes to the pcregrep command:
3047    
3048        (a) Refactored how -x works; insert ^(...)$ instead of setting
3049            PCRE_ANCHORED and checking the length, in preparation for adding
3050            something similar for -w.
3051    
3052        (b) Added the -w (match as a word) option.
3053    
3054        (c) Refactored the way lines are read and buffered so as to have more
3055            than one at a time available.
3056    
3057        (d) Implemented a pcregrep test script.
3058    
3059        (e) Added the -M (multiline match) option. This allows patterns to match
3060            over several lines of the subject. The buffering ensures that at least
3061            8K, or the rest of the document (whichever is the shorter) is available
3062            for matching (and similarly the previous 8K for lookbehind assertions).
3063    
3064        (f) Changed the --help output so that it now says
3065    
3066              -w, --word-regex(p)
3067    
3068            instead of two lines, one with "regex" and the other with "regexp"
3069            because that confused at least one person since the short forms are the
3070            same. (This required a bit of code, as the output is generated
3071            automatically from a table. It wasn't just a text change.)
3072    
3073        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3074            option but starts with a hyphen. Could be a pattern or a path name
3075            starting with a hyphen, for instance.
3076    
3077        (h) "-" can be given as a file name to represent stdin.
3078    
3079        (i) When file names are being printed, "(standard input)" is used for
3080            the standard input, for compatibility with GNU grep. Previously
3081            "<stdin>" was used.
3082    
3083        (j) The option --label=xxx can be used to supply a name to be used for
3084            stdin when file names are being printed. There is no short form.
3085    
3086        (k) Re-factored the options decoding logic because we are going to add
3087            two more options that take data. Such options can now be given in four
3088            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3089    
3090        (l) Added the -A, -B, and -C options for requesting that lines of context
3091            around matches be printed.
3092    
3093        (m) Added the -L option to print the names of files that do not contain
3094            any matching lines, that is, the complement of -l.
3095    
3096        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3097            continue to scan other files.
3098    
3099        (o) The -s option was incorrectly implemented. For compatibility with other
3100            greps, it now suppresses the error message for a non-existent or non-
3101            accessible file (but not the return code). There is a new option called
3102            -q that suppresses the output of matching lines, which was what -s was
3103            previously doing.
3104    
3105        (p) Added --include and --exclude options to specify files for inclusion
3106            and exclusion when recursing.
3107    
3108    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3109        Hopefully, it now does.
3110    
3111    12. Missing cast in pcre_study().
3112    
3113    13. Added an "uninstall" target to the makefile.
3114    
3115    14. Replaced "extern" in the function prototypes in Makefile.in with
3116        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3117        world, but is set differently for Windows.
3118    
3119    15. Added a second compiling function called pcre_compile2(). The only
3120        difference is that it has an extra argument, which is a pointer to an
3121        integer error code. When there is a compile-time failure, this is set
3122        non-zero, in addition to the error test pointer being set to point to an
3123        error message. The new argument may be NULL if no error number is required
3124        (but then you may as well call pcre_compile(), which is now just a
3125        wrapper). This facility is provided because some applications need a
3126        numeric error indication, but it has also enabled me to tidy up the way
3127        compile-time errors are handled in the POSIX wrapper.
3128    
3129    16. Added VPATH=.libs to the makefile; this should help when building with one
3130        prefix path and installing with another. (Or so I'm told by someone who
3131        knows more about this stuff than I do.)
3132    
3133    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3134        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3135        match everything, including newlines. This is not POSIX-compatible, but
3136        somebody wanted the feature. From pcretest it can be activated by using
3137        both the P and the s flags.
3138    
3139    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3140    
3141    19. libpcre.pc was being incorrectly installed as executable.
3142    
3143    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3144        it now also looks for '\r' so that it will work unmodified on Windows.
3145    
3146    21. Added Google's contributed C++ wrapper to the distribution.
3147    
3148    22. Added some untidy missing memory free() calls in pcretest, to keep
3149        Electric Fence happy when testing.
3150    
3151    
3152    
3153    Version 5.0 13-Sep-04
3154    ---------------------
3155    
3156     1. Internal change: literal characters are no longer packed up into items
3157        containing multiple characters in a single byte-string. Each character
3158        is now matched using a separate opcode. However, there may be more than one
3159        byte in the character in UTF-8 mode.
3160    
3161     2. The pcre_callout_block structure has two new fields: pattern_position and
3162        next_item_length. These contain the offset in the pattern to the next match
3163        item, and its length, respectively.
3164    
3165     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3166        insertion of callouts before each pattern item. Added the /C option to
3167        pcretest to make use of this.
3168    
3169     4. On the advice of a Windows user, the lines
3170    
3171          #if defined(_WIN32) || defined(WIN32)
3172          _setmode( _fileno( stdout ), 0x8000 );
3173          #endif  /* defined(_WIN32) || defined(WIN32) */
3174    
3175        have been added to the source of pcretest. This apparently does useful
3176        magic in relation to line terminators.
3177    
3178     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3179        for the benefit of those environments where the "b" makes a difference.
3180    
3181     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3182        to know about it. I have put a hack into configure.in that adds in code
3183        to set GCC=yes if CC=icc. This seems to end up at a point in the
3184        generated configure script that is early enough to affect the setting of
3185        compiler options, which is what is needed, but I have no means of testing
3186        whether it really works. (The user who reported this had patched the
3187        generated configure script, which of course I cannot do.)
3188    
3189        LATER: After change 22 below (new libtool files), the configure script
3190        seems to know about icc (and also ecc). Therefore, I have commented out
3191        this hack in configure.in.
3192    
3193     7. Added support for pkg-config (2 patches were sent in).
3194    
3195     8. Negated POSIX character classes that used a combination of internal tables
3196        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3197        [[:^ascii]]. Typically, they would match almost any characters. The other
3198        POSIX classes were not broken in this way.
3199    
3200     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3201        to find the match, as PCRE was deluded into thinking that the match had to
3202        start at the start point or following a newline. The same bug applied to
3203        patterns with negative forward assertions or any backward assertions
3204        preceding ".*" at the start, unless the pattern required a fixed first
3205        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3206    
3207    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3208        starting at the last subject character, bytes beyond the end of the subject
3209        string were read.
3210    
3211    11. Renamed the variable "class" as "classbits" to make life easier for C++
3212        users. (Previously there was a macro definition, but it apparently wasn't
3213        enough.)
3214    
3215    12. Added the new field "tables" to the extra data so that tables can be passed
3216        in at exec time, or the internal tables can be re-selected. This allows
3217        a compiled regex to be saved and re-used at a later time by a different
3218        program that might have everything at different addresses.
3219    
3220    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3221        -R library as well as a -L library.
3222    
3223    14. The debugging options of pcretest (-d on the command line or D on a
3224        pattern) showed incorrect output for anything following an extended class
3225        that contained multibyte characters and which was followed by a quantifier.
3226    
3227    15. Added optional support for general category Unicode character properties
3228        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3229        support. It adds about 90K to the size of the library. The meanings of the
3230        inbuilt class escapes such as \d and \s have NOT been changed.
3231    
3232    16. Updated pcredemo.c to include calls to free() to release the memory for the
3233        compiled pattern.
3234    
3235    17. The generated file chartables.c was being created in the source directory
3236        instead of in the building directory. This caused the build to fail if the
3237        source directory was different from the building directory, and was
3238        read-only.
3239    
3240    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3241        file. No doubt somebody will tell me if they don't make sense... Also added
3242        Dan Mooney's comments about building on OpenVMS.
3243    
3244    19. Added support for partial matching via the PCRE_PARTIAL option for
3245        pcre_exec() and the \P data escape in pcretest.
3246    
3247    20. Extended pcretest with 3 new pattern features:
3248    
3249        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3250              write the compiled pattern to the file whose name is "rest-of-line".
3251              This is a straight binary dump of the data, with the saved pointer to
3252              the character tables forced to be NULL. The study data, if any, is
3253              written too. After writing, pcretest reads a new pattern.
3254    
3255        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3256              compiled pattern from the given file. There must not be any
3257              occurrences of "<" in the file name (pretty unlikely); if there are,
3258              pcretest will instead treat the initial "<" as a pattern delimiter.
3259              After reading in the pattern, pcretest goes on to read data lines as
3260              usual.
3261    
3262        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3263              and 16-bit fields in a compiled pattern, to simulate a pattern that
3264              was compiled on a host of opposite endianness.
3265    
3266    21. The pcre-exec() function can now cope with patterns that were compiled on
3267        hosts of opposite endianness, with this restriction:
3268    
3269          As for any compiled expression that is saved and used later, the tables
3270          pointer field cannot be preserved; the extra_data field in the arguments
3271          to pcre_exec() should be used to pass in a tables address if a value
3272          other than the default internal tables were used at compile time.
3273    
3274    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3275        now diagnosed as an error. Previously, most of the time, a negative number
3276        would have been treated as zero, but if in addition "ovector" was passed as
3277        NULL, a crash could occur.
3278    
3279    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3280        new versions from the libtool 1.5 distribution (the last one is a copy of
3281        a file called libtool.m4). This seems to have fixed the need to patch
3282        "configure" to support Darwin 1.3 (which I used to do). However, I still
3283        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3284        workstation).
3285    
3286    24. Changed the PCRE licence to be the more standard "BSD" licence.
3287    
3288    
3289    Version 4.5 01-Dec-03
3290    ---------------------
3291    
3292     1. There has been some re-arrangement of the code for the match() function so
3293        that it can be compiled in a version that does not call itself recursively.
3294        Instead, it keeps those local variables that need separate instances for
3295        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3296        needs to "recurse". Keeping track of where control must go is done by means
3297        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3298        hide most of the details from the main code, and operates only if
3299        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3300        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3301        operating.
3302    
3303        To make it easier for callers to provide specially tailored get/free
3304        functions for this usage, two new functions, pcre_stack_malloc, and
3305        pcre_stack_free, are used. They are always called in strict stacking order,
3306        and the size of block requested is always the same.
3307    
3308        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3309        PCRE has been compiled to use the stack or the heap for recursion. The
3310        -C option of pcretest uses this to show which version is compiled.
3311    
3312        A new data escape \S, is added to pcretest; it causes the amounts of store
3313        obtained and freed by both kinds of malloc/free at match time to be added
3314        to the output.
3315    
3316     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3317        what's available on my current Linux desktop machine.
3318    
3319     3. When matching a UTF-8 string, the test for a valid string at the start has
3320        been extended. If start_offset is not zero, PCRE now checks that it points
3321        to a byte that is the start of a UTF-8 character. If not, it returns
3322        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3323        this is necessary because there may be backward assertions in the pattern.
3324        When matching the same subject several times, it may save resources to use
3325        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3326    
3327     4. The code for checking the validity of UTF-8 strings has been tightened so
3328        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3329        containing "overlong sequences".
3330    
3331     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3332        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3333        should have been "&", but it just so happened that all the cases this let
3334        through by mistake were picked up later in the function.
3335    
3336     6. I had used a variable called "isblank" - this is a C99 function, causing
3337        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3338    
3339     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3340        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3341        so the version is shown; (c) stop "make test" echoing "./RunTest".
3342    
3343     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3344    
3345     9. The prototype for memmove() for systems that don't have it was using
3346        size_t, but the inclusion of the header that defines size_t was later. I've
3347        moved the #includes for the C headers earlier to avoid this.
3348    
3349    10. Added some adjustments to the code to make it easier to compiler on certain
3350        special systems:
3351    
3352          (a) Some "const" qualifiers were missing.
3353          (b) Added the macro EXPORT before all exported functions; by default this
3354              is defined to be empty.
3355          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3356              that it reads its output file name as an argument instead of writing
3357              to the standard output and assuming this can be redirected.
3358    
3359    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3360        class containing characters with values greater than 255, PCRE compilation
3361        went into a loop.
3362    
3363    12. A recursive reference to a subpattern that was within another subpattern
3364        that had a minimum quantifier of zero caused PCRE to crash. For example,
3365        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3366        recursion. If the recursively-called subpattern itself had a zero repeat,
3367        that was OK.
3368    
3369    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3370        buffer into which it was copied (for escape processing) was still set at
3371        1024, so long lines caused crashes.
3372    
3373    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3374        "internal error: code overflow...". This applied to any character class
3375        that was followed by a possessive quantifier.
3376    
3377    15. Modified the Makefile to add libpcre.la as a prerequisite for
3378        libpcreposix.la because I was told this is needed for a parallel build to
3379        work.
3380    
3381    16. If a pattern that contained .* following optional items at the start was
3382        studied, the wrong optimizing data was generated, leading to matching
3383        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3384        matching string must start with a or b or c. The correct conclusion for
3385        this pattern is that a match can start with any character.
3386    
3387    
3388    Version 4.4 13-Aug-03
3389    ---------------------
3390    
3391     1. In UTF-8 mode, a character class containing characters with values between
3392        127 and 255 was not handled correctly if the compiled pattern was studied.
3393        In fixing this, I have also improved the studying algorithm for such
3394        classes (slightly).
3395    
3396     2. Three internal functions had redundant arguments passed to them. Removal
3397        might give a very teeny performance improvement.
3398    
3399     3. Documentation bug: the value of the capture_top field in a callout is *one
3400        more than* the number of the hightest numbered captured substring.
3401    
3402     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3403        in incorrectly linking with a previously installed version. They now link
3404        explicitly with libpcre.la.
3405    
3406     5. configure.in no longer needs to recognize Cygwin specially.
3407    
3408     6. A problem in pcre.in for Windows platforms is fixed.
3409    
3410     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3411        pcretest, it used to include the size of the study block as part of its
3412        output. Unfortunately, the structure contains a field that has a different
3413        size on different hardware architectures. This meant that the tests that
3414        showed this size failed. As the block is currently always of a fixed size,
3415        this information isn't actually particularly useful in pcretest output, so
3416        I have just removed it.
3417    
3418     8. Three pre-processor statements accidentally did not start in column 1.
3419        Sadly, there are *still* compilers around that complain, even though
3420        standard C has not required this for well over a decade. Sigh.
3421    
3422     9. In pcretest, the code for checking callouts passed small integers in the
3423        callout_data field, which is a void * field. However, some picky compilers
3424        complained about the casts involved for this on 64-bit systems. Now
3425        pcretest passes the address of the small integer instead, which should get
3426        rid of the warnings.
3427    
3428    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3429        both compile and run time, and gives an error if an invalid UTF-8 sequence
3430        is found. There is a option for disabling this check in cases where the
3431        string is known to be correct and/or the maximum performance is wanted.
3432    
3433    11. In response to a bug report, I changed one line in Makefile.in from
3434    
3435            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3436        to
3437            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3438    
3439        to look similar to other lines, but I have no way of telling whether this
3440        is the right thing to do, as I do not use Windows. No doubt I'll get told
3441        if it's wrong...
3442    
3443    
3444    Version 4.3 21-May-03
3445    ---------------------
3446    
3447    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3448       Makefile.
3449    
3450    2. Some refactoring to improve the quality of the code:
3451    
3452       (i)   The utf8_table... variables are now declared "const".
3453    
3454       (ii)  The code for \cx, which used the "case flipping" table to upper case
3455             lower case letters, now just substracts 32. This is ASCII-specific,
3456             but the whole concept of \cx is ASCII-specific, so it seems
3457             reasonable.
3458    
3459       (iii) PCRE was using its character types table to recognize decimal and
3460             hexadecimal digits in the pattern. This is silly, because it handles
3461             only 0-9, a-f, and A-F, but the character types table is locale-
3462             specific, which means strange things might happen. A private
3463             table is now used for this - though it costs 256 bytes, a table is
3464             much faster than multiple explicit tests. Of course, the standard
3465             character types table is still used for matching digits in subject
3466             strings against \d.
3467    
3468       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3469             ending in _t are). So I've renamed it as ESC_tee.
3470    
3471    3. The first argument for regexec() in the POSIX wrapper should have been
3472       defined as "const".
3473    
3474    4. Changed pcretest to use malloc() for its buffers so that they can be
3475       Electric Fenced for debugging.
3476    
3477    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3478       to read one or more bytes before the start of the subject string. Often this
3479       had no effect on PCRE's behaviour, but in some circumstances it could
3480       provoke a segmentation fault.
3481    
3482    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3483       to try to read one or more bytes before the start of the subject string.
3484    
3485    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3486       UTF-8 support could misbehave in various ways if the subject string
3487       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3488       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3489       back over UTF-8 characters.)
3490    
3491    
3492    Version 4.2 14-Apr-03
3493    ---------------------
3494    
3495    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3496    
3497    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3498         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3499         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3500         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3501         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3502           and BUILD_EXEEXT
3503         Note: automatic setting of the BUILD variables is not yet working
3504         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3505           compile-time but not at link-time
3506         [LINK]: use for linking executables only
3507         make different versions for Windows and non-Windows
3508         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3509           libraries
3510         [LINK_FOR_BUILD]: new variable
3511         [OBJEXT]: use throughout
3512         [EXEEXT]: use throughout
3513         <winshared>: new target
3514         <wininstall>: new target
3515         <dftables.o>: use native compiler
3516         <dftables>: use native linker
3517         <install>: handle Windows platform correctly
3518         <clean>: ditto
3519         <check>: ditto
3520         copy DLL to top builddir before testing
3521    
3522       As part of these changes, -no-undefined was removed again. This was reported
3523       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3524       in any case.
3525    
3526    3. Some tidies to get rid of compiler warnings:
3527    
3528       . In the match_data structure, match_limit was an unsigned long int, whereas
3529         match_call_count was an int. I've made them both unsigned long ints.
3530    
3531       . In pcretest the fact that a const uschar * doesn't automatically cast to
3532         a void * provoked a warning.
3533    
3534       . Turning on some more compiler warnings threw up some "shadow" variables
3535         and a few more missing casts.
3536    
3537    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3538       option, a class that contained a single character with a value between 128
3539       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3540    
3541    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3542       option, a class that contained several characters, but with at least one
3543       whose value was between 128 and 255 caused PCRE to crash.
3544    
3545    
3546    Version 4.1 12-Mar-03
3547    ---------------------
3548    
3549    1. Compiling with gcc -pedantic found a couple of places where casts were
3550    needed, and a string in dftables.c that was longer than standard compilers are
3551    required to support.
3552    
3553    2. Compiling with Sun's compiler found a few more places where the code could
3554    be tidied up in order to avoid warnings.
3555    
3556    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3557    first of these names is deprecated in the latest Autoconf in favour of the name
3558    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3559    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3560    analogy I have changed it to CFLAGS_FOR_BUILD.
3561    
3562    4. Added -no-undefined to the linking command in the Makefile, because this is
3563    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3564    linking step for the pcreposix library.
3565    
3566    5. PCRE was failing to diagnose the case of two named groups with the same
3567    name.
3568    
3569    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3570    literal character that is needed in the subject for a match, and scans along to
3571    ensure that it is present before embarking on the full matching process. This
3572    saves time in cases of nested unlimited repeats that are never going to match.
3573    Problem: the scan can take a lot of time if the subject is very long (e.g.
3574    megabytes), thus penalizing straightforward matches. It is now done only if the
3575    amount of subject to be scanned is less than 1000 bytes.
3576    
3577    7. A lesser problem with the same optimization is that it was recording the
3578    first character of an anchored pattern as "needed", thus provoking a search
3579    right along the subject, even when the first match of the pattern was going to
3580    fail. The "needed" character is now not set for anchored patterns, unless it
3581    follows something in the pattern that is of non-fixed length. Thus, it still
3582    fulfils its original purpose of finding quick non-matches in cases of nested
3583    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3584    
3585    
3586    Version 4.0 17-Feb-03
3587    ---------------------
3588    
3589    1. If a comment in an extended regex that started immediately after a meta-item
3590    extended to the end of string, PCRE compiled incorrect data. This could lead to
3591    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3592    
3593    2. Moved to autoconf 2.53 and libtool 1.4.2.
3594    
3595    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3596    the special perltest8 script is no longer needed - all the tests can be run
3597    from a single perltest script.
3598    
3599    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3600    by \s. It has now been removed in PCRE. This means it isn't recognized as
3601    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3602    class [:space:] *does* include VT, thereby creating a mess.
3603    
3604    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3605    space and tab.
3606    
3607    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3608    its new features into the main test script, reducing the number of scripts.
3609    
3610    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3611    were backward compatible, and made the (?i) apply to the whole pattern, as if
3612    /i were given. Now it behaves more logically, and applies the option setting
3613    only to what follows. PCRE has been changed to follow suit. However, if it
3614    finds options settings right at the start of the pattern, it extracts them into
3615    the global options, as before. Thus, they show up in the info data.
3616    
3617    8. Added support for the \Q...\E escape sequence. Characters in between are
3618    treated as literals. This is slightly different from Perl in that $ and @ are
3619    also handled as literals inside the quotes. In Perl, they will cause variable
3620    interpolation. Note the following examples:
3621    
3622        Pattern            PCRE matches      Perl matches
3623    
3624        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3625        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3626        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3627    
3628    For compatibility with Perl, \Q...\E sequences are recognized inside character
3629    classes as well as outside them.
3630    
3631    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3632    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3633    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3634    signed/unsigned warnings.
3635    
3636    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3637    option for pcretest, so I've replaced it by a simple function that does just
3638    that job.
3639    
3640    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3641    "pcregrep -".
3642    
3643    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3644    Java package. This provides some syntactic sugar for simple cases of what my
3645    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3646    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3647    item, you can use this simplified notation. Note that only makes sense with
3648    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3649    greediness, whatever the setting of the PCRE_UNGREEDY option.
3650    
3651    13. A change of greediness default within a pattern was not taking effect at
3652    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3653    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3654    was abstracted outside.
3655    
3656    14. PCRE now supports the \G assertion. It is true when the current matching
3657    position is at the start point of the match. This differs from \A when the
3658    starting offset is non-zero. Used with the /g option of pcretest (or similar
3659    code), it works in the same way as it does for Perl's /g option. If all
3660    alternatives of a regex begin with \G, the expression is anchored to the start
3661    match position, and the "anchored" flag is set in the compiled expression.
3662    
3663    15. Some bugs concerning the handling of certain option changes within patterns
3664    have been fixed. These applied to options other than (?ims). For example,
3665    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3666    been the other way round. Some of this was related to change 7 above.
3667    
3668    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3669    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3670    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3671    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3672    
3673    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3674    mode. Unlike ".", it always matches newline, whatever the setting of
3675    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3676    assertions. Perl allows it, but it doesn't (in general) work because it can't
3677    calculate the length of the lookbehind. At least, that's the case for Perl
3678    5.8.0 - I've been told they are going to document that it doesn't work in
3679    future.
3680    
3681    18. Added an error diagnosis for escapes that PCRE does not support: these are
3682    \L, \l, \N, \P, \p, \U, \u, and \X.
3683    
3684    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3685    reading past the end of the pattern in cases such as /[abcd/.
3686    
3687    20. PCRE was getting more memory than necessary for patterns with classes that
3688    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3689    
3690    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3691    compiling PCRE for use with Virtual Pascal.
3692    
3693    22. Small fix to the Makefile to make it work properly if the build is done
3694    outside the source tree.
3695    
3696    23. Added a new extension: a condition to go with recursion. If a conditional
3697    subpattern starts with (?(R) the "true" branch is used if recursion has
3698    happened, whereas the "false" branch is used only at the top level.
3699    
3700    24. When there was a very long string of literal characters (over 255 bytes
3701    without UTF support, over 250 bytes with UTF support), the computation of how
3702    much memory was required could be incorrect, leading to segfaults or other
3703    strange effects.