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

Diff of /code/trunk/ChangeLog

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

revision 35 by nigel, Sat Feb 24 21:39:05 2007 UTC revision 997 by zherczeg, Fri Jul 27 11:31:03 2012 UTC
# Line 1  Line 1 
1  ChangeLog for PCRE  ChangeLog for PCRE
2  ------------------  ------------------
3    
4    Version 8.32
5    ------------
6    
7    1.  Improved JIT compiler optimizations for first character search and single
8        character iterators.
9    
10    2.  Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
11        Patch by Daniel Richard G.
12    
13    3.  Single character iterator optimizations in the JIT compiler.
14    
15    4.  Improved JIT compiler optimizations for character ranges.
16    
17    5.  Rename the "leave" variable names to "quit" to improve WinCE compatibility.
18        Reported by Giuseppe D'Angelo.
19    
20    6.  The PCRE_STARTLINE bit, indicating that a match can occur only at the start
21        of a line, was being set incorrectly in cases where .* appeared inside
22        atomic brackets at the start of a pattern, or where there was a subsequent
23        *PRUNE or *SKIP.
24    
25    7.  Improved instruction cache flush for POWER/PowerPC.
26        Patch by Daniel Richard G.
27    
28    
29    Version 8.31 06-July-2012
30    -------------------------
31    
32    1.  Fixing a wrong JIT test case and some compiler warnings.
33    
34    2.  Removed a bashism from the RunTest script.
35    
36    3.  Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
37        to unsigned type, result still unsigned" that was given by an MS compiler
38        on encountering the code "-sizeof(xxx)".
39    
40    4.  Partial matching support is added to the JIT compiler.
41    
42    5.  Fixed several bugs concerned with partial matching of items that consist
43        of more than one character:
44    
45        (a) /^(..)\1/ did not partially match "aba" because checking references was
46            done on an "all or nothing" basis. This also applied to repeated
47            references.
48    
49        (b) \R did not give a hard partial match if \r was found at the end of the
50            subject.
51    
52        (c) \X did not give a hard partial match after matching one or more
53            characters at the end of the subject.
54    
55        (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
56            a partial match for the string "\r".
57    
58        (e) When newline was set to CRLF, the metacharacter "." did not recognize
59            a partial match for a CR character at the end of the subject string.
60    
61    6.  If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
62        running pcretest, the text "(JIT)" added to the output whenever JIT is
63        actually used to run the match.
64    
65    7.  Individual JIT compile options can be set in pcretest by following -s+[+]
66        or /S+[+] with a digit between 1 and 7.
67    
68    8.  OP_NOT now supports any UTF character not just single-byte ones.
69    
70    9.  (*MARK) control verb is now supported by the JIT compiler.
71    
72    10. The command "./RunTest list" lists the available tests without actually
73        running any of them. (Because I keep forgetting what they all are.)
74    
75    11. Add PCRE_INFO_MAXLOOKBEHIND.
76    
77    12. Applied a (slightly modified) user-supplied patch that improves performance
78        when the heap is used for recursion (compiled with --disable-stack-for-
79        recursion). Instead of malloc and free for each heap frame each time a
80        logical recursion happens, frames are retained on a chain and re-used where
81        possible. This sometimes gives as much as 30% improvement.
82    
83    13. As documented, (*COMMIT) is now confined to within a recursive subpattern
84        call.
85    
86    14. As documented, (*COMMIT) is now confined to within a positive assertion.
87    
88    15. It is now possible to link pcretest with libedit as an alternative to
89        libreadline.
90    
91    16. (*COMMIT) control verb is now supported by the JIT compiler.
92    
93    17. The Unicode data tables have been updated to Unicode 6.1.0.
94    
95    18. Added --file-list option to pcregrep.
96    
97    19. Added binary file support to pcregrep, including the -a, --binary-files,
98        -I, and --text options.
99    
100    20. The madvise function is renamed for posix_madvise for QNX compatibility
101        reasons. Fixed by Giuseppe D'Angelo.
102    
103    21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
104    
105    22. Changed the option for creating symbolic links for 16-bit man pages from
106        -s to -sf so that re-installing does not cause issues.
107    
108    23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
109    
110    24. Fixed a very old bug in pcretest that caused errors with restarted DFA
111        matches in certain environments (the workspace was not being correctly
112        retained). Also added to pcre_dfa_exec() a simple plausibility check on
113        some of the workspace data at the beginning of a restart.
114    
115    25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
116        was not doing so when it should - probably a typo introduced by SVN 528
117        (change 8.10/14).
118    
119    26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
120        \w+ when the character tables indicated that \x{c4} was a word character.
121        There were several related cases, all because the tests for doing a table
122        lookup were testing for characters less than 127 instead of 255.
123    
124    27. If a pattern contains capturing parentheses that are not used in a match,
125        their slots in the ovector are set to -1. For those that are higher than
126        any matched groups, this happens at the end of processing. In the case when
127        there were back references that the ovector was too small to contain
128        (causing temporary malloc'd memory to be used during matching), and the
129        highest capturing number was not used, memory off the end of the ovector
130        was incorrectly being set to -1. (It was using the size of the temporary
131        memory instead of the true size.)
132    
133    28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
134        ovector size, it uses memory at the end of the block that it has got.
135    
136    29. Check for an overlong MARK name and give an error at compile time. The
137        limit is 255 for the 8-bit library and 65535 for the 16-bit library.
138    
139    30. JIT compiler update.
140    
141    31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
142        Rill for the patch.
143    
144    32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
145    
146    33. Variable renamings in the PCRE-JIT compiler. No functionality change.
147    
148    34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
149        SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
150        was enabled.
151    
152    35. Improve JIT code generation for greedy plus quantifier.
153    
154    36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
155        1 to "aa" instead of to an empty string. The bug affected repeated groups
156        that could potentially match an empty string.
157    
158    37. Optimizing single character iterators in JIT.
159    
160    38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
161        the same checks as \x{...} characters in non-JavaScript mode. Specifically,
162        codepoints that are too big for the mode are faulted, and in a UTF mode,
163        disallowed codepoints are also faulted.
164    
165    39. If PCRE was compiled with UTF support, in three places in the DFA
166        matcher there was code that should only have been obeyed in UTF mode, but
167        was being obeyed unconditionally. In 8-bit mode this could cause incorrect
168        processing when bytes with values greater than 127 were present. In 16-bit
169        mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
170        both cases the values are those that cannot be the first data item in a UTF
171        character. The three items that might have provoked this were recursions,
172        possessively repeated groups, and atomic groups.
173    
174    40. Ensure that libpcre is explicitly listed in the link commands for pcretest
175        and pcregrep, because some OS require shared objects to be explicitly
176        passed to ld, causing the link step to fail if they are not.
177    
178    41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
179        mode, patterns that started with \h* or \R* might be incorrectly matched.
180    
181    
182    Version 8.30 04-February-2012
183    -----------------------------
184    
185    1.  Renamed "isnumber" as "is_a_number" because in some Mac environments this
186        name is defined in ctype.h.
187    
188    2.  Fixed a bug in fixed-length calculation for lookbehinds that would show up
189        only in quite long subpatterns.
190    
191    3.  Removed the function pcre_info(), which has been obsolete and deprecated
192        since it was replaced by pcre_fullinfo() in February 2000.
193    
194    4.  For a non-anchored pattern, if (*SKIP) was given with a name that did not
195        match a (*MARK), and the match failed at the start of the subject, a
196        reference to memory before the start of the subject could occur. This bug
197        was introduced by fix 17 of release 8.21.
198    
199    5.  A reference to an unset group with zero minimum repetition was giving
200        totally wrong answers (in non-JavaScript-compatibility mode). For example,
201        /(another)?(\1?)test/ matched against "hello world test". This bug was
202        introduced in release 8.13.
203    
204    6.  Add support for 16-bit character strings (a large amount of work involving
205        many changes and refactorings).
206    
207    7.  RunGrepTest failed on msys because \r\n was replaced by whitespace when the
208        command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
209        from a file.
210    
211    8.  Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
212        rounding is not applied in this particular case).
213    
214    9.  The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
215        if they appear, or are escaped, in patterns.
216    
217    10. Get rid of a number of -Wunused-but-set-variable warnings.
218    
219    11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
220        "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
221        Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
222        also returns the mark "x". This bug applied to capturing parentheses,
223        non-capturing parentheses, and atomic parentheses. It also applied to some
224        assertions.
225    
226    12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
227        information out of configure.ac instead of relying on pcre.h.generic, which
228        is not stored in the repository.
229    
230    13. Applied Dmitry V. Levin's patch for a more portable method for linking with
231        -lreadline.
232    
233    14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
234    
235    15. Applied Graycode's patch to put the top-level frame on the stack rather
236        than the heap when not using the stack for recursion. This gives a
237        performance improvement in many cases when recursion is not deep.
238    
239    16. Experimental code added to "pcretest -C" to output the stack frame size.
240    
241    
242    Version 8.21 12-Dec-2011
243    ------------------------
244    
245    1.  Updating the JIT compiler.
246    
247    2.  JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
248        are added as well.
249    
250    3.  Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
251        PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
252        calling _pcre_jit_exec. Some extra comments are added.
253    
254    4.  (*MARK) settings inside atomic groups that do not contain any capturing
255        parentheses, for example, (?>a(*:m)), were not being passed out. This bug
256        was introduced by change 18 for 8.20.
257    
258    5.  Supporting of \x, \U and \u in JavaScript compatibility mode based on the
259        ECMA-262 standard.
260    
261    6.  Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
262        erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
263        This bug was probably introduced by change 9 of 8.13.
264    
265    7.  While fixing 6 above, I noticed that a number of other items were being
266        incorrectly rejected as "not fixed length". This arose partly because newer
267        opcodes had not been added to the fixed-length checking code. I have (a)
268        corrected the bug and added tests for these items, and (b) arranged for an
269        error to occur if an unknown opcode is encountered while checking for fixed
270        length instead of just assuming "not fixed length". The items that were
271        rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
272        (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
273        repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
274    
275    8.  A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
276        being incorrectly compiled and would have given unpredicatble results.
277    
278    9.  A possessively repeated subpattern with minimum repeat count greater than
279        one behaved incorrectly. For example, (A){2,}+ behaved as if it was
280        (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
281        the first (A) could occur when it should not.
282    
283    10. Add a cast and remove a redundant test from the code.
284    
285    11. JIT should use pcre_malloc/pcre_free for allocation.
286    
287    12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
288        best practice nowadays, and helps with cross-compiling. (If the exec_prefix
289        is anything other than /usr, -L is still shown).
290    
291    13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
292    
293    14. Perl does not support \N without a following name in a [] class; PCRE now
294        also gives an error.
295    
296    15. If a forward reference was repeated with an upper limit of around 2000,
297        it caused the error "internal error: overran compiling workspace". The
298        maximum number of forward references (including repeats) was limited by the
299        internal workspace, and dependent on the LINK_SIZE. The code has been
300        rewritten so that the workspace expands (via pcre_malloc) if necessary, and
301        the default depends on LINK_SIZE. There is a new upper limit (for safety)
302        of around 200,000 forward references. While doing this, I also speeded up
303        the filling in of repeated forward references.
304    
305    16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
306        incorrectly expecting the subject to contain another "a" after the start.
307    
308    17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
309        in the match, the SKIP should be ignored. This was not happening; instead
310        the SKIP was being treated as NOMATCH. For patterns such as
311        /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
312        tested.
313    
314    18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
315        now much more compatible with Perl, in particular in cases where the result
316        is a non-match for a non-anchored pattern. For example, if
317        /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
318        "m", where previously it did not return a name. A side effect of this
319        change is that for partial matches, the last encountered mark name is
320        returned, as for non matches. A number of tests that were previously not
321        Perl-compatible have been moved into the Perl-compatible test files. The
322        refactoring has had the pleasing side effect of removing one argument from
323        the match() function, thus reducing its stack requirements.
324    
325    19. If the /S+ option was used in pcretest to study a pattern using JIT,
326        subsequent uses of /S (without +) incorrectly behaved like /S+.
327    
328    21. Retrieve executable code size support for the JIT compiler and fixing
329        some warnings.
330    
331    22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
332        not work when the shorter character appeared right at the end of the
333        subject string.
334    
335    23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
336        systems.
337    
338    24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
339        output it when the /M option is used in pcretest.
340    
341    25. The CheckMan script was not being included in the distribution. Also, added
342        an explicit "perl" to run Perl scripts from the PrepareRelease script
343        because this is reportedly needed in Windows.
344    
345    26. If study data was being save in a file and studying had not found a set of
346        "starts with" bytes for the pattern, the data written to the file (though
347        never used) was taken from uninitialized memory and so caused valgrind to
348        complain.
349    
350    27. Updated RunTest.bat as provided by Sheri Pierce.
351    
352    28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
353    
354    29. Computation of memory usage for the table of capturing group names was
355        giving an unnecessarily large value.
356    
357    
358    Version 8.20 21-Oct-2011
359    ------------------------
360    
361    1.  Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
362        a POSIX class. After further experiments with Perl, which convinced me that
363        Perl has bugs and confusions, a closing square bracket is no longer allowed
364        in a POSIX name. This bug also affected patterns with classes that started
365        with full stops.
366    
367    2.  If a pattern such as /(a)b|ac/ is matched against "ac", there is no
368        captured substring, but while checking the failing first alternative,
369        substring 1 is temporarily captured. If the output vector supplied to
370        pcre_exec() was not big enough for this capture, the yield of the function
371        was still zero ("insufficient space for captured substrings"). This cannot
372        be totally fixed without adding another stack variable, which seems a lot
373        of expense for a edge case. However, I have improved the situation in cases
374        such as /(a)(b)x|abc/ matched against "abc", where the return code
375        indicates that fewer than the maximum number of slots in the ovector have
376        been set.
377    
378    3.  Related to (2) above: when there are more back references in a pattern than
379        slots in the output vector, pcre_exec() uses temporary memory during
380        matching, and copies in the captures as far as possible afterwards. It was
381        using the entire output vector, but this conflicts with the specification
382        that only 2/3 is used for passing back captured substrings. Now it uses
383        only the first 2/3, for compatibility. This is, of course, another edge
384        case.
385    
386    4.  Zoltan Herczeg's just-in-time compiler support has been integrated into the
387        main code base, and can be used by building with --enable-jit. When this is
388        done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
389        runtime --no-jit option is given.
390    
391    5.  When the number of matches in a pcre_dfa_exec() run exactly filled the
392        ovector, the return from the function was zero, implying that there were
393        other matches that did not fit. The correct "exactly full" value is now
394        returned.
395    
396    6.  If a subpattern that was called recursively or as a subroutine contained
397        (*PRUNE) or any other control that caused it to give a non-standard return,
398        invalid errors such as "Error -26 (nested recursion at the same subject
399        position)" or even infinite loops could occur.
400    
401    7.  If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
402        computing the minimum length on reaching *ACCEPT, and so ended up with the
403        wrong value of 1 rather than 0. Further investigation indicates that
404        computing a minimum subject length in the presence of *ACCEPT is difficult
405        (think back references, subroutine calls), and so I have changed the code
406        so that no minimum is registered for a pattern that contains *ACCEPT.
407    
408    8.  If (*THEN) was present in the first (true) branch of a conditional group,
409        it was not handled as intended. [But see 16 below.]
410    
411    9.  Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
412        Sheri Pierce.
413    
414    10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
415        the first byte in a match must be "a".
416    
417    11. Change 17 for 8.13 increased the recursion depth for patterns like
418        /a(?:.)*?a/ drastically. I've improved things by remembering whether a
419        pattern contains any instances of (*THEN). If it does not, the old
420        optimizations are restored. It would be nice to do this on a per-group
421        basis, but at the moment that is not feasible.
422    
423    12. In some environments, the output of pcretest -C is CRLF terminated. This
424        broke RunTest's code that checks for the link size. A single white space
425        character after the value is now allowed for.
426    
427    13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
428        For "fr", it uses the Windows-specific input and output files.
429    
430    14. If (*THEN) appeared in a group that was called recursively or as a
431        subroutine, it did not work as intended. [But see next item.]
432    
433    15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
434        pattern fragments (but not containing any | characters). If A and B are
435        matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
436        was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
437        D. In other words, Perl considers parentheses that do not contain any |
438        characters to be part of a surrounding alternative, whereas PCRE was
439        treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
440        differently. PCRE now behaves in the same way as Perl, except in the case
441        of subroutine/recursion calls such as (?1) which have in any case always
442        been different (but PCRE had them first :-).
443    
444    16. Related to 15 above: Perl does not treat the | in a conditional group as
445        creating alternatives. Such a group is treated in the same way as an
446        ordinary group without any | characters when processing (*THEN). PCRE has
447        been changed to match Perl's behaviour.
448    
449    17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
450        RunGrepTest script failed.
451    
452    18. Change 22 for version 13 caused atomic groups to use more stack. This is
453        inevitable for groups that contain captures, but it can lead to a lot of
454        stack use in large patterns. The old behaviour has been restored for atomic
455        groups that do not contain any capturing parentheses.
456    
457    19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
458        suppress the check for a minimum subject length at run time. (If it was
459        given to pcre_exec() or pcre_dfa_exec() it did work.)
460    
461    20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
462        fail to work when decoding hex characters in data strings in EBCDIC
463        environments.
464    
465    21. It appears that in at least one Mac OS environment, the isxdigit() function
466        is implemented as a macro that evaluates to its argument more than once,
467        contravening the C 90 Standard (I haven't checked a later standard). There
468        was an instance in pcretest which caused it to go wrong when processing
469        \x{...} escapes in subject strings. The has been rewritten to avoid using
470        things like p++ in the argument of isxdigit().
471    
472    
473    Version 8.13 16-Aug-2011
474    ------------------------
475    
476    1.  The Unicode data tables have been updated to Unicode 6.0.0.
477    
478    2.  Two minor typos in pcre_internal.h have been fixed.
479    
480    3.  Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
481        pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
482        in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
483    
484    4.  There were a number of related bugs in the code for matching backrefences
485        caselessly in UTF-8 mode when codes for the characters concerned were
486        different numbers of bytes. For example, U+023A and U+2C65 are an upper
487        and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
488        (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
489        code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
490        2-byte code at the end of the subject (it thought there wasn't enough data
491        left).
492    
493    5.  Comprehensive information about what went wrong is now returned by
494        pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
495        as the output vector has at least 2 elements. The offset of the start of
496        the failing character and a reason code are placed in the vector.
497    
498    6.  When the UTF-8 string check fails for pcre_compile(), the offset that is
499        now returned is for the first byte of the failing character, instead of the
500        last byte inspected. This is an incompatible change, but I hope it is small
501        enough not to be a problem. It makes the returned offset consistent with
502        pcre_exec() and pcre_dfa_exec().
503    
504    7.  pcretest now gives a text phrase as well as the error number when
505        pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
506        failure, the offset and reason code are output.
507    
508    8.  When \R was used with a maximizing quantifier it failed to skip backwards
509        over a \r\n pair if the subsequent match failed. Instead, it just skipped
510        back over a single character (\n). This seems wrong (because it treated the
511        two characters as a single entity when going forwards), conflicts with the
512        documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
513        behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
514        has been changed.
515    
516    9.  Some internal refactoring has changed the processing so that the handling
517        of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
518        time (the PCRE_DOTALL option was changed this way some time ago: version
519        7.7 change 16). This has made it possible to abolish the OP_OPT op code,
520        which was always a bit of a fudge. It also means that there is one less
521        argument for the match() function, which reduces its stack requirements
522        slightly. This change also fixes an incompatibility with Perl: the pattern
523        (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
524    
525    10. More internal refactoring has drastically reduced the number of recursive
526        calls to match() for possessively repeated groups such as (abc)++ when
527        using pcre_exec().
528    
529    11. While implementing 10, a number of bugs in the handling of groups were
530        discovered and fixed:
531    
532        (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
533        (a|)*(?1) gave a compile-time internal error.
534        ((a|)+)+  did not notice that the outer group could match an empty string.
535        (^a|^)+   was not marked as anchored.
536        (.*a|.*)+ was not marked as matching at start or after a newline.
537    
538    12. Yet more internal refactoring has removed another argument from the match()
539        function. Special calls to this function are now indicated by setting a
540        value in a variable in the "match data" data block.
541    
542    13. Be more explicit in pcre_study() instead of relying on "default" for
543        opcodes that mean there is no starting character; this means that when new
544        ones are added and accidentally left out of pcre_study(), testing should
545        pick them up.
546    
547    14. The -s option of pcretest has been documented for ages as being an old
548        synonym of -m (show memory usage). I have changed it to mean "force study
549        for every regex", that is, assume /S for every regex. This is similar to -i
550        and -d etc. It's slightly incompatible, but I'm hoping nobody is still
551        using it. It makes it easier to run collections of tests with and without
552        study enabled, and thereby test pcre_study() more easily. All the standard
553        tests are now run with and without -s (but some patterns can be marked as
554        "never study" - see 20 below).
555    
556    15. When (*ACCEPT) was used in a subpattern that was called recursively, the
557        restoration of the capturing data to the outer values was not happening
558        correctly.
559    
560    16. If a recursively called subpattern ended with (*ACCEPT) and matched an
561        empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
562        pattern had matched an empty string, and so incorrectly returned a no
563        match.
564    
565    17. There was optimizing code for the last branch of non-capturing parentheses,
566        and also for the obeyed branch of a conditional subexpression, which used
567        tail recursion to cut down on stack usage. Unfortunately, now that there is
568        the possibility of (*THEN) occurring in these branches, tail recursion is
569        no longer possible because the return has to be checked for (*THEN). These
570        two optimizations have therefore been removed. [But see 8.20/11 above.]
571    
572    18. If a pattern containing \R was studied, it was assumed that \R always
573        matched two bytes, thus causing the minimum subject length to be
574        incorrectly computed because \R can also match just one byte.
575    
576    19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
577        was incorrectly computed.
578    
579    20. If /S is present twice on a test pattern in pcretest input, it now
580        *disables* studying, thereby overriding the use of -s on the command line
581        (see 14 above). This is necessary for one or two tests to keep the output
582        identical in both cases.
583    
584    21. When (*ACCEPT) was used in an assertion that matched an empty string and
585        PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
586    
587    22. When an atomic group that contained a capturing parenthesis was
588        successfully matched, but the branch in which it appeared failed, the
589        capturing was not being forgotten if a higher numbered group was later
590        captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
591        group 1 to "a", when in fact it should be unset. This applied to multi-
592        branched capturing and non-capturing groups, repeated or not, and also to
593        positive assertions (capturing in negative assertions does not happen
594        in PCRE) and also to nested atomic groups.
595    
596    23. Add the ++ qualifier feature to pcretest, to show the remainder of the
597        subject after a captured substring, to make it easier to tell which of a
598        number of identical substrings has been captured.
599    
600    24. The way atomic groups are processed by pcre_exec() has been changed so that
601        if they are repeated, backtracking one repetition now resets captured
602        values correctly. For example, if ((?>(a+)b)+aabab) is matched against
603        "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
604        "aaa". Previously, it would have been "a". As part of this code
605        refactoring, the way recursive calls are handled has also been changed.
606    
607    25. If an assertion condition captured any substrings, they were not passed
608        back unless some other capturing happened later. For example, if
609        (?(?=(a))a) was matched against "a", no capturing was returned.
610    
611    26. When studying a pattern that contained subroutine calls or assertions,
612        the code for finding the minimum length of a possible match was handling
613        direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
614        group 1 called group 2 while simultaneously a separate group 2 called group
615        1). A stack overflow occurred in this case. I have fixed this by limiting
616        the recursion depth to 10.
617    
618    27. Updated RunTest.bat in the distribution to the version supplied by Tom
619        Fortmann. This supports explicit test numbers on the command line, and has
620        argument validation and error reporting.
621    
622    28. An instance of \X with an unlimited repeat could fail if at any point the
623        first character it looked at was a mark character.
624    
625    29. Some minor code refactoring concerning Unicode properties and scripts
626        should reduce the stack requirement of match() slightly.
627    
628    30. Added the '=' option to pcretest to check the setting of unused capturing
629        slots at the end of the pattern, which are documented as being -1, but are
630        not included in the return count.
631    
632    31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
633        compiled something random. Now it gives a compile-time error (as does
634        Perl).
635    
636    32. A *MARK encountered during the processing of a positive assertion is now
637        recorded and passed back (compatible with Perl).
638    
639    33. If --only-matching or --colour was set on a pcregrep call whose pattern
640        had alternative anchored branches, the search for a second match in a line
641        was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
642        matched the line "0102" twice. The same bug affected patterns that started
643        with a backwards assertion. For example /\b01|\b02/ also matched "0102"
644        twice.
645    
646    34. Previously, PCRE did not allow quantification of assertions. However, Perl
647        does, and because of capturing effects, quantifying parenthesized
648        assertions may at times be useful. Quantifiers are now allowed for
649        parenthesized assertions.
650    
651    35. A minor code tidy in pcre_compile() when checking options for \R usage.
652    
653    36. \g was being checked for fancy things in a character class, when it should
654        just be a literal "g".
655    
656    37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
657        appearance of a nested POSIX class supersedes an apparent external class.
658        For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
659        unescaped square brackets may also appear as part of class names. For
660        example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
661        more like Perl. (But see 8.20/1 above.)
662    
663    38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
664        was because it thought it was \N{name}, which is not supported).
665    
666    39. Add minix to OS list not supporting the -S option in pcretest.
667    
668    40. PCRE tries to detect cases of infinite recursion at compile time, but it
669        cannot analyze patterns in sufficient detail to catch mutual recursions
670        such as ((?1))((?2)). There is now a runtime test that gives an error if a
671        subgroup is called recursively as a subpattern for a second time at the
672        same position in the subject string. In previous releases this might have
673        been caught by the recursion limit, or it might have run out of stack.
674    
675    41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
676        happen only once. PCRE was, however incorrectly giving a compile time error
677        "recursive call could loop indefinitely" because it cannot analyze the
678        pattern in sufficient detail. The compile time test no longer happens when
679        PCRE is compiling a conditional subpattern, but actual runaway loops are
680        now caught at runtime (see 40 above).
681    
682    42. It seems that Perl allows any characters other than a closing parenthesis
683        to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
684        has been changed to be the same.
685    
686    43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
687        as not to get warnings when autogen.sh is called. Also changed
688        AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
689    
690    44. To help people who use pcregrep to scan files containing exceedingly long
691        lines, the following changes have been made:
692    
693        (a) The default value of the buffer size parameter has been increased from
694            8K to 20K. (The actual buffer used is three times this size.)
695    
696        (b) The default can be changed by ./configure --with-pcregrep-bufsize when
697            PCRE is built.
698    
699        (c) A --buffer-size=n option has been added to pcregrep, to allow the size
700            to be set at run time.
701    
702        (d) Numerical values in pcregrep options can be followed by K or M, for
703            example --buffer-size=50K.
704    
705        (e) If a line being scanned overflows pcregrep's buffer, an error is now
706            given and the return code is set to 2.
707    
708    45. Add a pointer to the latest mark to the callout data block.
709    
710    46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
711        partial match of an empty string instead of no match. This was specific to
712        the use of ".".
713    
714    47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
715        complete match instead of a partial match. This bug was dependent on both
716        the PCRE_UTF8 and PCRE_DOTALL options being set.
717    
718    48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
719        starting byte set, because \b was not being ignored.
720    
721    
722    Version 8.12 15-Jan-2011
723    ------------------------
724    
725    1.  Fixed some typos in the markup of the man pages, and wrote a script that
726        checks for such things as part of the documentation building process.
727    
728    2.  On a big-endian 64-bit system, pcregrep did not correctly process the
729        --match-limit and --recursion-limit options (added for 8.11). In
730        particular, this made one of the standard tests fail. (The integer value
731        went into the wrong half of a long int.)
732    
733    3.  If the --colour option was given to pcregrep with -v (invert match), it
734        did strange things, either producing crazy output, or crashing. It should,
735        of course, ignore a request for colour when reporting lines that do not
736        match.
737    
738    4.  Another pcregrep bug caused similar problems if --colour was specified with
739        -M (multiline) and the pattern match finished with a line ending.
740    
741    5.  In pcregrep, when a pattern that ended with a literal newline sequence was
742        matched in multiline mode, the following line was shown as part of the
743        match. This seems wrong, so I have changed it.
744    
745    6.  Another pcregrep bug in multiline mode, when --colour was specified, caused
746        the check for further matches in the same line (so they could be coloured)
747        to overrun the end of the current line. If another match was found, it was
748        incorrectly shown (and then shown again when found in the next line).
749    
750    7.  If pcregrep was compiled under Windows, there was a reference to the
751        function pcregrep_exit() before it was defined. I am assuming this was
752        the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
753        reported by a user. I've moved the definition above the reference.
754    
755    
756    Version 8.11 10-Dec-2010
757    ------------------------
758    
759    1.  (*THEN) was not working properly if there were untried alternatives prior
760        to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
761        backtracked to try for "b" instead of moving to the next alternative branch
762        at the same level (in this case, to look for "c"). The Perl documentation
763        is clear that when (*THEN) is backtracked onto, it goes to the "next
764        alternative in the innermost enclosing group".
765    
766    2.  (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
767        such as   (A(*COMMIT)B(*THEN)C|D)  any failure after matching A should
768        result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
769        (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
770        (*THEN).
771    
772    3.  If \s appeared in a character class, it removed the VT character from
773        the class, even if it had been included by some previous item, for example
774        in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
775        of \s, but is part of the POSIX "space" class.)
776    
777    4.  A partial match never returns an empty string (because you can always
778        match an empty string at the end of the subject); however the checking for
779        an empty string was starting at the "start of match" point. This has been
780        changed to the "earliest inspected character" point, because the returned
781        data for a partial match starts at this character. This means that, for
782        example, /(?<=abc)def/ gives a partial match for the subject "abc"
783        (previously it gave "no match").
784    
785    5.  Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
786        of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
787        previously a full match would be given. However, setting PCRE_PARTIAL_HARD
788        has an implication that the given string is incomplete (because a partial
789        match is preferred over a full match). For this reason, these items now
790        give a partial match in this situation. [Aside: previously, the one case
791        /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
792        match rather than a full match, which was wrong by the old rules, but is
793        now correct.]
794    
795    6.  There was a bug in the handling of #-introduced comments, recognized when
796        PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
797        If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
798        UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
799        scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
800        but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
801        places in pcre_compile().
802    
803    7.  Related to (6) above, when pcre_compile() was skipping #-introduced
804        comments when looking ahead for named forward references to subpatterns,
805        the only newline sequence it recognized was NL. It now handles newlines
806        according to the set newline convention.
807    
808    8.  SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
809        former, but used strtoul(), whereas pcretest avoided strtoul() but did not
810        cater for a lack of strerror(). These oversights have been fixed.
811    
812    9.  Added --match-limit and --recursion-limit to pcregrep.
813    
814    10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
815    
816    11. When the -o option was used, pcregrep was setting a return code of 1, even
817        when matches were found, and --line-buffered was not being honoured.
818    
819    12. Added an optional parentheses number to the -o and --only-matching options
820        of pcregrep.
821    
822    13. Imitating Perl's /g action for multiple matches is tricky when the pattern
823        can match an empty string. The code to do it in pcretest and pcredemo
824        needed fixing:
825    
826        (a) When the newline convention was "crlf", pcretest got it wrong, skipping
827            only one byte after an empty string match just before CRLF (this case
828            just got forgotten; "any" and "anycrlf" were OK).
829    
830        (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
831            mode when an empty string match preceded an ASCII character followed by
832            a non-ASCII character. (The code for advancing by one character rather
833            than one byte was nonsense.)
834    
835        (c) The pcredemo.c sample program did not have any code at all to handle
836            the cases when CRLF is a valid newline sequence.
837    
838    14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
839        as a starting offset was within the subject string. There is now a new
840        error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
841        negative or greater than the length of the string. In order to test this,
842        pcretest is extended to allow the setting of negative starting offsets.
843    
844    15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
845        starting offset points to the beginning of a UTF-8 character was
846        unnecessarily clumsy. I tidied it up.
847    
848    16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
849        bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
850    
851    17. Nobody had reported that the --include_dir option, which was added in
852        release 7.7 should have been called --include-dir (hyphen, not underscore)
853        for compatibility with GNU grep. I have changed it to --include-dir, but
854        left --include_dir as an undocumented synonym, and the same for
855        --exclude-dir, though that is not available in GNU grep, at least as of
856        release 2.5.4.
857    
858    18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
859        characters from a string of bytes) have been redefined so as not to use
860        loops, in order to improve performance in some environments. At the same
861        time, I abstracted some of the common code into auxiliary macros to save
862        repetition (this should not affect the compiled code).
863    
864    19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
865        compile-time error is now given if \c is not followed by an ASCII
866        character, that is, a byte less than 128. (In EBCDIC mode, the code is
867        different, and any byte value is allowed.)
868    
869    20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
870        START_OPTIMIZE option, which is now allowed at compile time - but just
871        passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
872        to pcregrep and other applications that have no direct access to PCRE
873        options. The new /Y option in pcretest sets this option when calling
874        pcre_compile().
875    
876    21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
877        back references. Groups containing recursive back references were forced to
878        be atomic by that change, but in the case of named groups, the amount of
879        memory required was incorrectly computed, leading to "Failed: internal
880        error: code overflow". This has been fixed.
881    
882    22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
883        pcretest.c, to avoid build problems in some Borland environments.
884    
885    
886    Version 8.10 25-Jun-2010
887    ------------------------
888    
889    1.  Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
890        THEN.
891    
892    2.  (*ACCEPT) was not working when inside an atomic group.
893    
894    3.  Inside a character class, \B is treated as a literal by default, but
895        faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
896        causes the error). The code is unchanged, but I tidied the documentation.
897    
898    4.  Inside a character class, PCRE always treated \R and \X as literals,
899        whereas Perl faults them if its -w option is set. I have changed PCRE so
900        that it faults them when PCRE_EXTRA is set.
901    
902    5.  Added support for \N, which always matches any character other than
903        newline. (It is the same as "." when PCRE_DOTALL is not set.)
904    
905    6.  When compiling pcregrep with newer versions of gcc which may have
906        FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
907        declared with attribute warn_unused_result" were given. Just casting the
908        result to (void) does not stop the warnings; a more elaborate fudge is
909        needed. I've used a macro to implement this.
910    
911    7.  Minor change to pcretest.c to avoid a compiler warning.
912    
913    8.  Added four artifical Unicode properties to help with an option to make
914        \s etc use properties (see next item). The new properties are: Xan
915        (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
916    
917    9.  Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
918        use Unicode properties. (*UCP) at the start of a pattern can be used to set
919        this option. Modified pcretest to add /W to test this facility. Added
920        REG_UCP to make it available via the POSIX interface.
921    
922    10. Added --line-buffered to pcregrep.
923    
924    11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
925        studied, and the match started with a letter with a code point greater than
926        127 whose first byte was different to the first byte of the other case of
927        the letter, the other case of this starting letter was not recognized
928        (#976).
929    
930    12. If a pattern that was studied started with a repeated Unicode property
931        test, for example, \p{Nd}+, there was the theoretical possibility of
932        setting up an incorrect bitmap of starting bytes, but fortunately it could
933        not have actually happened in practice until change 8 above was made (it
934        added property types that matched character-matching opcodes).
935    
936    13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
937        possible starting bytes for non-anchored patterns.
938    
939    14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
940        \R, and also a number of cases that involve Unicode properties, both
941        explicit and implicit when PCRE_UCP is set.
942    
943    15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
944        input, it could crash or give wrong results if characters with values
945        greater than 0xc0 were present in the subject string. (Detail: it assumed
946        UTF-8 input when processing these items.)
947    
948    16. Added a lot of (int) casts to avoid compiler warnings in systems where
949        size_t is 64-bit (#991).
950    
951    17. Added a check for running out of memory when PCRE is compiled with
952        --disable-stack-for-recursion (#990).
953    
954    18. If the last data line in a file for pcretest does not have a newline on
955        the end, a newline was missing in the output.
956    
957    19. The default pcre_chartables.c file recognizes only ASCII characters (values
958        less than 128) in its various bitmaps. However, there is a facility for
959        generating tables according to the current locale when PCRE is compiled. It
960        turns out that in some environments, 0x85 and 0xa0, which are Unicode space
961        characters, are recognized by isspace() and therefore were getting set in
962        these tables, and indeed these tables seem to approximate to ISO 8859. This
963        caused a problem in UTF-8 mode when pcre_study() was used to create a list
964        of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
965        which of course cannot start UTF-8 characters. I have changed the code so
966        that only real ASCII characters (less than 128) and the correct starting
967        bytes for UTF-8 encodings are set for characters greater than 127 when in
968        UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
969        altogether.)
970    
971    20. Added the /T option to pcretest so as to be able to run tests with non-
972        standard character tables, thus making it possible to include the tests
973        used for 19 above in the standard set of tests.
974    
975    21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
976        reference to a subpattern the other side of a comment that contains an
977        opening parenthesis caused either an internal compiling error, or a
978        reference to the wrong subpattern.
979    
980    
981    Version 8.02 19-Mar-2010
982    ------------------------
983    
984    1.  The Unicode data tables have been updated to Unicode 5.2.0.
985    
986    2.  Added the option --libs-cpp to pcre-config, but only when C++ support is
987        configured.
988    
989    3.  Updated the licensing terms in the pcregexp.pas file, as agreed with the
990        original author of that file, following a query about its status.
991    
992    4.  On systems that do not have stdint.h (e.g. Solaris), check for and include
993        inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
994    
995    5.  A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
996        quantifier applied to a forward-referencing subroutine call, could compile
997        incorrect code or give the error "internal error: previously-checked
998        referenced subpattern not found".
999    
1000    6.  Both MS Visual Studio and Symbian OS have problems with initializing
1001        variables to point to external functions. For these systems, therefore,
1002        pcre_malloc etc. are now initialized to local functions that call the
1003        relevant global functions.
1004    
1005    7.  There were two entries missing in the vectors called coptable and poptable
1006        in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1007        I've fixed the data, and added a kludgy way of testing at compile time that
1008        the lengths are correct (equal to the number of opcodes).
1009    
1010    8.  Following on from 7, I added a similar kludge to check the length of the
1011        eint vector in pcreposix.c.
1012    
1013    9.  Error texts for pcre_compile() are held as one long string to avoid too
1014        much relocation at load time. To find a text, the string is searched,
1015        counting zeros. There was no check for running off the end of the string,
1016        which could happen if a new error number was added without updating the
1017        string.
1018    
1019    10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1020    
1021    11. \K was not working if it appeared in an atomic group or in a group that
1022        was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1023        \K is "not well defined" if used in an assertion. PCRE now accepts it if
1024        the assertion is positive, but not if it is negative.
1025    
1026    12. Change 11 fortuitously reduced the size of the stack frame used in the
1027        "match()" function of pcre_exec.c by one pointer. Forthcoming
1028        implementation of support for (*MARK) will need an extra pointer on the
1029        stack; I have reserved it now, so that the stack frame size does not
1030        decrease.
1031    
1032    13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1033        item in branch that calls a recursion is a subroutine call - as in the
1034        second branch in the above example - was incorrectly given the compile-
1035        time error "recursive call could loop indefinitely" because pcre_compile()
1036        was not correctly checking the subroutine for matching a non-empty string.
1037    
1038    14. The checks for overrunning compiling workspace could trigger after an
1039        overrun had occurred. This is a "should never occur" error, but it can be
1040        triggered by pathological patterns such as hundreds of nested parentheses.
1041        The checks now trigger 100 bytes before the end of the workspace.
1042    
1043    15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1044    
1045    
1046    Version 8.01 19-Jan-2010
1047    ------------------------
1048    
1049    1.  If a pattern contained a conditional subpattern with only one branch (in
1050        particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1051        computed the wrong minimum data length (which is of course zero for such
1052        subpatterns). This could cause incorrect "no match" results.
1053    
1054    2.  For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1055        the pattern is reset in the first branch, pcre_compile() failed with
1056        "internal error: code overflow at offset...". This happened only when
1057        the reset was to the original external option setting. (An optimization
1058        abstracts leading options settings into an external setting, which was the
1059        cause of this.)
1060    
1061    3.  A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1062        of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1063        assertion pattern did not match (meaning that the assertion was true), it
1064        was incorrectly treated as false if the SKIP had been reached during the
1065        matching. This also applied to assertions used as conditions.
1066    
1067    4.  If an item that is not supported by pcre_dfa_exec() was encountered in an
1068        assertion subpattern, including such a pattern used as a condition,
1069        unpredictable results occurred, instead of the error return
1070        PCRE_ERROR_DFA_UITEM.
1071    
1072    5.  The C++ GlobalReplace function was not working like Perl for the special
1073        situation when an empty string is matched. It now does the fancy magic
1074        stuff that is necessary.
1075    
1076    6.  In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1077        removed. (These were left over from very, very early versions of PCRE.)
1078    
1079    7.  Some cosmetic changes to the code to make life easier when compiling it
1080        as part of something else:
1081    
1082        (a) Change DEBUG to PCRE_DEBUG.
1083    
1084        (b) In pcre_compile(), rename the member of the "branch_chain" structure
1085            called "current" as "current_branch", to prevent a collision with the
1086            Linux macro when compiled as a kernel module.
1087    
1088        (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1089            prevent a collision with the Linux macro when compiled as a kernel
1090            module.
1091    
1092    8.  In pcre_compile() there are some checks for integer overflows that used to
1093        cast potentially large values to (double). This has been changed to that
1094        when building, a check for int64_t is made, and if it is found, it is used
1095        instead, thus avoiding the use of floating point arithmetic. (There is no
1096        other use of FP in PCRE.) If int64_t is not found, the fallback is to
1097        double.
1098    
1099    9.  Added two casts to avoid signed/unsigned warnings from VS Studio Express
1100        2005 (difference between two addresses compared to an unsigned value).
1101    
1102    10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1103        custom one, because of the following reported problem in Windows:
1104    
1105          - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1106              under Win32.
1107          - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1108              therefore missing the function definition.
1109          - The compiler thus generates a "C" signature for the test function.
1110          - The linker fails to find the "C" function.
1111          - PCRE fails to configure if asked to do so against libbz2.
1112    
1113    11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1114        messages were output:
1115    
1116          Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1117          rerunning libtoolize, to keep the correct libtool macros in-tree.
1118          Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1119    
1120        I have done both of these things.
1121    
1122    12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1123        most of the time, it *can* run out if it is given a pattern that contains a
1124        runaway infinite recursion. I updated the discussion in the pcrestack man
1125        page.
1126    
1127    13. Now that we have gone to the x.xx style of version numbers, the minor
1128        version may start with zero. Using 08 or 09 is a bad idea because users
1129        might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1130        interpreted as invalid octal numbers. I've updated the previous comment in
1131        configure.ac, and also added a check that gives an error if 08 or 09 are
1132        used.
1133    
1134    14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1135        causing partial matching to fail when the end of the subject matched \W
1136        in a UTF-8 pattern where \W was quantified with a minimum of 3.
1137    
1138    15. There were some discrepancies between the declarations in pcre_internal.h
1139        of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1140        their definitions. The declarations used "const uschar *" and the
1141        definitions used USPTR. Even though USPTR is normally defined as "const
1142        unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1143        reported that: "This difference in casting confuses some C++ compilers, for
1144        example, SunCC recognizes above declarations as different functions and
1145        generates broken code for hbpcre." I have changed the declarations to use
1146        USPTR.
1147    
1148    16. GNU libtool is named differently on some systems. The autogen.sh script now
1149        tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1150        (FreeBSD).
1151    
1152    17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
1153        (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
1154        comment: "Figure out how to create a longlong from a string: strtoll and
1155        equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
1156        instance, but it only takes 2 args instead of 3!"
1157    
1158    18. A subtle bug concerned with back references has been fixed by a change of
1159        specification, with a corresponding code fix. A pattern such as
1160        ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
1161        refers, was giving matches when it shouldn't. For example, xa=xaaa would
1162        match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
1163        same bug. Such groups have to be quantified to be useful, or contained
1164        inside another quantified group. (If there's no repetition, the reference
1165        can never match.) The problem arises because, having left the group and
1166        moved on to the rest of the pattern, a later failure that backtracks into
1167        the group uses the captured value from the final iteration of the group
1168        rather than the correct earlier one. I have fixed this in PCRE by forcing
1169        any group that contains a reference to itself to be an atomic group; that
1170        is, there cannot be any backtracking into it once it has completed. This is
1171        similar to recursive and subroutine calls.
1172    
1173    
1174    Version 8.00 19-Oct-09
1175    ----------------------
1176    
1177    1.  The table for translating pcre_compile() error codes into POSIX error codes
1178        was out-of-date, and there was no check on the pcre_compile() error code
1179        being within the table. This could lead to an OK return being given in
1180        error.
1181    
1182    2.  Changed the call to open a subject file in pcregrep from fopen(pathname,
1183        "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1184        in a Windows environment.
1185    
1186    3.  The pcregrep --count option prints the count for each file even when it is
1187        zero, as does GNU grep. However, pcregrep was also printing all files when
1188        --files-with-matches was added. Now, when both options are given, it prints
1189        counts only for those files that have at least one match. (GNU grep just
1190        prints the file name in this circumstance, but including the count seems
1191        more useful - otherwise, why use --count?) Also ensured that the
1192        combination -clh just lists non-zero counts, with no names.
1193    
1194    4.  The long form of the pcregrep -F option was incorrectly implemented as
1195        --fixed_strings instead of --fixed-strings. This is an incompatible change,
1196        but it seems right to fix it, and I didn't think it was worth preserving
1197        the old behaviour.
1198    
1199    5.  The command line items --regex=pattern and --regexp=pattern were not
1200        recognized by pcregrep, which required --regex pattern or --regexp pattern
1201        (with a space rather than an '='). The man page documented the '=' forms,
1202        which are compatible with GNU grep; these now work.
1203    
1204    6.  No libpcreposix.pc file was created for pkg-config; there was just
1205        libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1206    
1207    7.  Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1208        when UCP support is not needed, by modifying the Python script that
1209        generates it from Unicode data files. This should not matter if the module
1210        is correctly used as a library, but I received one complaint about 50K of
1211        unwanted data. My guess is that the person linked everything into his
1212        program rather than using a library. Anyway, it does no harm.
1213    
1214    8.  A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1215        was a minimum greater than 1 for a wide character in a possessive
1216        repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1217        which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1218        character. Chaos in the form of incorrect output or a compiling loop could
1219        result.
1220    
1221    9.  The restrictions on what a pattern can contain when partial matching is
1222        requested for pcre_exec() have been removed. All patterns can now be
1223        partially matched by this function. In addition, if there are at least two
1224        slots in the offset vector, the offset of the earliest inspected character
1225        for the match and the offset of the end of the subject are set in them when
1226        PCRE_ERROR_PARTIAL is returned.
1227    
1228    10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1229        synonymous with PCRE_PARTIAL, for backwards compatibility, and
1230        PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1231        and may be more useful for multi-segment matching.
1232    
1233    11. Partial matching with pcre_exec() is now more intuitive. A partial match
1234        used to be given if ever the end of the subject was reached; now it is
1235        given only if matching could not proceed because another character was
1236        needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1237        string "Z", which now yields "no match" instead of "partial match". In the
1238        case of pcre_dfa_exec(), "no match" is given if every matching path for the
1239        final character ended with (*FAIL).
1240    
1241    12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1242        if the pattern had a "must contain" character that was already found in the
1243        earlier partial match, unless partial matching was again requested. For
1244        example, with the pattern /dog.(body)?/, the "must contain" character is
1245        "g". If the first part-match was for the string "dog", restarting with
1246        "sbody" failed. This bug has been fixed.
1247    
1248    13. The string returned by pcre_dfa_exec() after a partial match has been
1249        changed so that it starts at the first inspected character rather than the
1250        first character of the match. This makes a difference only if the pattern
1251        starts with a lookbehind assertion or \b or \B (\K is not supported by
1252        pcre_dfa_exec()). It's an incompatible change, but it makes the two
1253        matching functions compatible, and I think it's the right thing to do.
1254    
1255    14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1256        so that the demonstration program is easily available in environments where
1257        PCRE has not been installed from source.
1258    
1259    15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1260        libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1261        library.
1262    
1263    16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1264        It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1265        is not the first non-POSIX option to be added. Clearly some people find
1266        these options useful.
1267    
1268    17. If a caller to the POSIX matching function regexec() passes a non-zero
1269        value for nmatch with a NULL value for pmatch, the value of
1270        nmatch is forced to zero.
1271    
1272    18. RunGrepTest did not have a test for the availability of the -u option of
1273        the diff command, as RunTest does. It now checks in the same way as
1274        RunTest, and also checks for the -b option.
1275    
1276    19. If an odd number of negated classes containing just a single character
1277        interposed, within parentheses, between a forward reference to a named
1278        subpattern and the definition of the subpattern, compilation crashed with
1279        an internal error, complaining that it could not find the referenced
1280        subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1281        [The bug was that it was starting one character too far in when skipping
1282        over the character class, thus treating the ] as data rather than
1283        terminating the class. This meant it could skip too much.]
1284    
1285    20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1286        /g option in pcretest when the pattern contains \K, which makes it possible
1287        to have an empty string match not at the start, even when the pattern is
1288        anchored. Updated pcretest and pcredemo to use this option.
1289    
1290    21. If the maximum number of capturing subpatterns in a recursion was greater
1291        than the maximum at the outer level, the higher number was returned, but
1292        with unset values at the outer level. The correct (outer level) value is
1293        now given.
1294    
1295    22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1296        PCRE did not set those parentheses (unlike Perl). I have now found a way to
1297        make it do so. The string so far is captured, making this feature
1298        compatible with Perl.
1299    
1300    23. The tests have been re-organized, adding tests 11 and 12, to make it
1301        possible to check the Perl 5.10 features against Perl 5.10.
1302    
1303    24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1304        pattern matches a fixed length string. PCRE did not allow this; now it
1305        does. Neither allows recursion.
1306    
1307    25. I finally figured out how to implement a request to provide the minimum
1308        length of subject string that was needed in order to match a given pattern.
1309        (It was back references and recursion that I had previously got hung up
1310        on.) This code has now been added to pcre_study(); it finds a lower bound
1311        to the length of subject needed. It is not necessarily the greatest lower
1312        bound, but using it to avoid searching strings that are too short does give
1313        some useful speed-ups. The value is available to calling programs via
1314        pcre_fullinfo().
1315    
1316    26. While implementing 25, I discovered to my embarrassment that pcretest had
1317        not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1318        study optimizations had never been tested with that matching function.
1319        Oops. What is worse, even when it was passed study data, there was a bug in
1320        pcre_dfa_exec() that meant it never actually used it. Double oops. There
1321        were also very few tests of studied patterns with pcre_dfa_exec().
1322    
1323    27. If (?| is used to create subpatterns with duplicate numbers, they are now
1324        allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1325        on the other side of the coin, they are no longer allowed to have different
1326        names, because these cannot be distinguished in PCRE, and this has caused
1327        confusion. (This is a difference from Perl.)
1328    
1329    28. When duplicate subpattern names are present (necessarily with different
1330        numbers, as required by 27 above), and a test is made by name in a
1331        conditional pattern, either for a subpattern having been matched, or for
1332        recursion in such a pattern, all the associated numbered subpatterns are
1333        tested, and the overall condition is true if the condition is true for any
1334        one of them. This is the way Perl works, and is also more like the way
1335        testing by number works.
1336    
1337    
1338    Version 7.9 11-Apr-09
1339    ---------------------
1340    
1341    1.  When building with support for bzlib/zlib (pcregrep) and/or readline
1342        (pcretest), all targets were linked against these libraries. This included
1343        libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1344        libraries. This caused unwanted dependencies to be created. This problem
1345        has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1346        pcretest is linked with readline.
1347    
1348    2.  The "typedef int BOOL" in pcre_internal.h that was included inside the
1349        "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1350        moved outside it again, because FALSE and TRUE are already defined in AIX,
1351        but BOOL is not.
1352    
1353    3.  The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1354        PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1355    
1356    4.  The pcregrep documentation said spaces were inserted as well as colons (or
1357        hyphens) following file names and line numbers when outputting matching
1358        lines. This is not true; no spaces are inserted. I have also clarified the
1359        wording for the --colour (or --color) option.
1360    
1361    5.  In pcregrep, when --colour was used with -o, the list of matching strings
1362        was not coloured; this is different to GNU grep, so I have changed it to be
1363        the same.
1364    
1365    6.  When --colo(u)r was used in pcregrep, only the first matching substring in
1366        each matching line was coloured. Now it goes on to look for further matches
1367        of any of the test patterns, which is the same behaviour as GNU grep.
1368    
1369    7.  A pattern that could match an empty string could cause pcregrep to loop; it
1370        doesn't make sense to accept an empty string match in pcregrep, so I have
1371        locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1372        seems to be how GNU grep behaves.
1373    
1374    8.  The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1375        start or after a newline", because the conditional assertion was not being
1376        correctly handled. The rule now is that both the assertion and what follows
1377        in the first alternative must satisfy the test.
1378    
1379    9.  If auto-callout was enabled in a pattern with a conditional group whose
1380        condition was an assertion, PCRE could crash during matching, both with
1381        pcre_exec() and pcre_dfa_exec().
1382    
1383    10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1384        used for matching.
1385    
1386    11. Unicode property support in character classes was not working for
1387        characters (bytes) greater than 127 when not in UTF-8 mode.
1388    
1389    12. Added the -M command line option to pcretest.
1390    
1391    14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1392    
1393    15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1394    
1395    16. Added comments and documentation about mis-use of no_arg in the C++
1396        wrapper.
1397    
1398    17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1399        from Martin Jerabek that uses macro names for all relevant character and
1400        string constants.
1401    
1402    18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1403        SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1404        SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1405        these, but not everybody uses configure.
1406    
1407    19. A conditional group that had only one branch was not being correctly
1408        recognized as an item that could match an empty string. This meant that an
1409        enclosing group might also not be so recognized, causing infinite looping
1410        (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1411        with the subject "ab", where knowledge that the repeated group can match
1412        nothing is needed in order to break the loop.
1413    
1414    20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1415        exec(), but without supplying a callout function, matching went wrong.
1416    
1417    21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1418        leak if the size of the offset vector was greater than 30. When the vector
1419        is smaller, the saved offsets during recursion go onto a local stack
1420        vector, but for larger vectors malloc() is used. It was failing to free
1421        when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1422        error, in fact).
1423    
1424    22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1425        heapframe that is used only when UTF-8 support is enabled. This caused no
1426        problem, but was untidy.
1427    
1428    23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1429        CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1430        included within another project.
1431    
1432    24. Steven Van Ingelgem's patches to add more options to the CMake support,
1433        slightly modified by me:
1434    
1435          (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1436              not building pcregrep.
1437    
1438          (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1439              if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1440    
1441    25. Forward references, both numeric and by name, in patterns that made use of
1442        duplicate group numbers, could behave incorrectly or give incorrect errors,
1443        because when scanning forward to find the reference group, PCRE was not
1444        taking into account the duplicate group numbers. A pattern such as
1445        ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1446    
1447    26. Changed a few more instances of "const unsigned char *" to USPTR, making
1448        the feature of a custom pointer more persuasive (as requested by a user).
1449    
1450    27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1451        pcretest.c, inside #ifndefs, because it seems they are sometimes already
1452        pre-defined.
1453    
1454    28. Added support for (*UTF8) at the start of a pattern.
1455    
1456    29. Arrange for flags added by the "release type" setting in CMake to be shown
1457        in the configuration summary.
1458    
1459    
1460    Version 7.8 05-Sep-08
1461    ---------------------
1462    
1463    1.  Replaced UCP searching code with optimized version as implemented for Ad
1464        Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1465        stage table and inline lookup instead of a function, giving speed ups of 2
1466        to 5 times on some simple patterns that I tested. Permission was given to
1467        distribute the MultiStage2.py script that generates the tables (it's not in
1468        the tarball, but is in the Subversion repository).
1469    
1470    2.  Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1471        scripts.
1472    
1473    3.  Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1474        a group with a zero qualifier. The result of the study could be incorrect,
1475        or the function might crash, depending on the pattern.
1476    
1477    4.  Caseless matching was not working for non-ASCII characters in back
1478        references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1479        It now works when Unicode Property Support is available.
1480    
1481    5.  In pcretest, an escape such as \x{de} in the data was always generating
1482        a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1483        non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1484        truncation.
1485    
1486    6.  Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1487    
1488    7.  Added two (int) casts to pcregrep when printing the difference of two
1489        pointers, in case they are 64-bit values.
1490    
1491    8.  Added comments about Mac OS X stack usage to the pcrestack man page and to
1492        test 2 if it fails.
1493    
1494    9.  Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1495        and a #define of that name to empty if it is not externally set. This is to
1496        allow users of MSVC to set it if necessary.
1497    
1498    10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1499        the convenience functions in the pcre_get.c source file.
1500    
1501    11. An option change at the start of a pattern that had top-level alternatives
1502        could cause overwriting and/or a crash. This command provoked a crash in
1503        some environments:
1504    
1505          printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1506    
1507        This potential security problem was recorded as CVE-2008-2371.
1508    
1509    12. For a pattern where the match had to start at the beginning or immediately
1510        after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1511        pcre_dfa_exec() could read past the end of the passed subject if there was
1512        no match. To help with detecting such bugs (e.g. with valgrind), I modified
1513        pcretest so that it places the subject at the end of its malloc-ed buffer.
1514    
1515    13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1516        exec() might read past the end of the data buffer in UTF-8 mode.
1517    
1518    14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1519        the data contained the byte 0x85 as part of a UTF-8 character within its
1520        first line. This applied both to normal and DFA matching.
1521    
1522    15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1523        /^[^d]*?$/8 failed to match "abc".
1524    
1525    16. Added a missing copyright notice to pcrecpp_internal.h.
1526    
1527    17. Make it more clear in the documentation that values returned from
1528        pcre_exec() in ovector are byte offsets, not character counts.
1529    
1530    18. Tidied a few places to stop certain compilers from issuing warnings.
1531    
1532    19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1533        supplied by Stefan Weber. I made a further small update for 7.8 because
1534        there is a change of source arrangements: the pcre_searchfuncs.c module is
1535        replaced by pcre_ucd.c.
1536    
1537    
1538    Version 7.7 07-May-08
1539    ---------------------
1540    
1541    1.  Applied Craig's patch to sort out a long long problem: "If we can't convert
1542        a string to a long long, pretend we don't even have a long long." This is
1543        done by checking for the strtoq, strtoll, and _strtoi64 functions.
1544    
1545    2.  Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1546        pre-7.6 versions, which defined a global no_arg variable instead of putting
1547        it in the RE class. (See also #8 below.)
1548    
1549    3.  Remove a line of dead code, identified by coverity and reported by Nuno
1550        Lopes.
1551    
1552    4.  Fixed two related pcregrep bugs involving -r with --include or --exclude:
1553    
1554        (1) The include/exclude patterns were being applied to the whole pathnames
1555            of files, instead of just to the final components.
1556    
1557        (2) If there was more than one level of directory, the subdirectories were
1558            skipped unless they satisfied the include/exclude conditions. This is
1559            inconsistent with GNU grep (and could even be seen as contrary to the
1560            pcregrep specification - which I improved to make it absolutely clear).
1561            The action now is always to scan all levels of directory, and just
1562            apply the include/exclude patterns to regular files.
1563    
1564    5.  Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1565        --exclude_dir in the tests to avoid scanning .svn directories.
1566    
1567    6.  Applied Craig's patch to the QuoteMeta function so that it escapes the
1568        NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1569        doesn't support NULs in patterns.
1570    
1571    7.  Added some missing "const"s to declarations of static tables in
1572        pcre_compile.c and pcre_dfa_exec.c.
1573    
1574    8.  Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1575        caused by fix #2  above. (Subsequently also a second patch to fix the
1576        first patch. And a third patch - this was a messy problem.)
1577    
1578    9.  Applied Craig's patch to remove the use of push_back().
1579    
1580    10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1581        matching function regexec().
1582    
1583    11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1584        which, however, unlike Perl's \g{...}, are subroutine calls, not back
1585        references. PCRE supports relative numbers with this syntax (I don't think
1586        Oniguruma does).
1587    
1588    12. Previously, a group with a zero repeat such as (...){0} was completely
1589        omitted from the compiled regex. However, this means that if the group
1590        was called as a subroutine from elsewhere in the pattern, things went wrong
1591        (an internal error was given). Such groups are now left in the compiled
1592        pattern, with a new opcode that causes them to be skipped at execution
1593        time.
1594    
1595    13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1596        to the way PCRE behaves:
1597    
1598        (a) A lone ] character is dis-allowed (Perl treats it as data).
1599    
1600        (b) A back reference to an unmatched subpattern matches an empty string
1601            (Perl fails the current match path).
1602    
1603        (c) A data ] in a character class must be notated as \] because if the
1604            first data character in a class is ], it defines an empty class. (In
1605            Perl it is not possible to have an empty class.) The empty class []
1606            never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1607            The negative empty class [^] matches any one character, independently
1608            of the DOTALL setting.
1609    
1610    14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1611        non-existent subpattern following a character class starting with ']' and
1612        containing () gave an internal compiling error instead of "reference to
1613        non-existent subpattern". Fortunately, when the pattern did exist, the
1614        compiled code was correct. (When scanning forwards to check for the
1615        existencd of the subpattern, it was treating the data ']' as terminating
1616        the class, so got the count wrong. When actually compiling, the reference
1617        was subsequently set up correctly.)
1618    
1619    15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1620        it was being rejected as not supported by pcre_dfa_exec(), even though
1621        other assertions are supported. I have made pcre_dfa_exec() support
1622        (*FAIL).
1623    
1624    16. The implementation of 13c above involved the invention of a new opcode,
1625        OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1626        cannot be changed at match time, I realized I could make a small
1627        improvement to matching performance by compiling OP_ALLANY instead of
1628        OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1629        on the OP_ANY path.
1630    
1631    17. Compiling pcretest on Windows with readline support failed without the
1632        following two fixes: (1) Make the unistd.h include conditional on
1633        HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1634    
1635    18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1636        ncurses library to be included for pcretest when ReadLine support is
1637        requested, but also to allow for it to be overridden. This patch came from
1638        Daniel Bergström.
1639    
1640    19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1641        as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1642        any errors with the current Unicode tables. Thanks to Peter Kankowski for
1643        spotting this.
1644    
1645    
1646    Version 7.6 28-Jan-08
1647    ---------------------
1648    
1649    1.  A character class containing a very large number of characters with
1650        codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1651        overflow.
1652    
1653    2.  Patch to cut out the "long long" test in pcrecpp_unittest when
1654        HAVE_LONG_LONG is not defined.
1655    
1656    3.  Applied Christian Ehrlicher's patch to update the CMake build files to
1657        bring them up to date and include new features. This patch includes:
1658    
1659        - Fixed PH's badly added libz and libbz2 support.
1660        - Fixed a problem with static linking.
1661        - Added pcredemo. [But later removed - see 7 below.]
1662        - Fixed dftables problem and added an option.
1663        - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1664            HAVE_LONG_LONG.
1665        - Added readline support for pcretest.
1666        - Added an listing of the option settings after cmake has run.
1667    
1668    4.  A user submitted a patch to Makefile that makes it easy to create
1669        "pcre.dll" under mingw when using Configure/Make. I added stuff to
1670        Makefile.am that cause it to include this special target, without
1671        affecting anything else. Note that the same mingw target plus all
1672        the other distribution libraries and programs are now supported
1673        when configuring with CMake (see 6 below) instead of with
1674        Configure/Make.
1675    
1676    5.  Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1677        This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1678        exported in the Windows port". It has not yet been confirmed that the patch
1679        solves the problem, but it does no harm.
1680    
1681    6.  Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1682        NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1683        with CMake, and also correct the comment about stack recursion.
1684    
1685    7.  Remove the automatic building of pcredemo from the ./configure system and
1686        from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1687        of a program that users should build themselves after PCRE is installed, so
1688        building it automatically is not really right. What is more, it gave
1689        trouble in some build environments.
1690    
1691    8.  Further tidies to CMakeLists.txt from Sheri and Christian.
1692    
1693    
1694    Version 7.5 10-Jan-08
1695    ---------------------
1696    
1697    1.  Applied a patch from Craig: "This patch makes it possible to 'ignore'
1698        values in parens when parsing an RE using the C++ wrapper."
1699    
1700    2.  Negative specials like \S did not work in character classes in UTF-8 mode.
1701        Characters greater than 255 were excluded from the class instead of being
1702        included.
1703    
1704    3.  The same bug as (2) above applied to negated POSIX classes such as
1705        [:^space:].
1706    
1707    4.  PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1708        defined or documented. It seems to have been a typo for PCRE_STATIC, so
1709        I have changed it.
1710    
1711    5.  The construct (?&) was not diagnosed as a syntax error (it referenced the
1712        first named subpattern) and a construct such as (?&a) would reference the
1713        first named subpattern whose name started with "a" (in other words, the
1714        length check was missing). Both these problems are fixed. "Subpattern name
1715        expected" is now given for (?&) (a zero-length name), and this patch also
1716        makes it give the same error for \k'' (previously it complained that that
1717        was a reference to a non-existent subpattern).
1718    
1719    6.  The erroneous patterns (?+-a) and (?-+a) give different error messages;
1720        this is right because (?- can be followed by option settings as well as by
1721        digits. I have, however, made the messages clearer.
1722    
1723    7.  Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1724        than the number used in the conditional) now cause a compile-time error.
1725        This is actually not compatible with Perl, which accepts such patterns, but
1726        treats the conditional as always being FALSE (as PCRE used to), but it
1727        seems to me that giving a diagnostic is better.
1728    
1729    8.  Change "alphameric" to the more common word "alphanumeric" in comments
1730        and messages.
1731    
1732    9.  Fix two occurrences of "backslash" in comments that should have been
1733        "backspace".
1734    
1735    10. Remove two redundant lines of code that can never be obeyed (their function
1736        was moved elsewhere).
1737    
1738    11. The program that makes PCRE's Unicode character property table had a bug
1739        which caused it to generate incorrect table entries for sequences of
1740        characters that have the same character type, but are in different scripts.
1741        It amalgamated them into a single range, with the script of the first of
1742        them. In other words, some characters were in the wrong script. There were
1743        thirteen such cases, affecting characters in the following ranges:
1744    
1745          U+002b0 - U+002c1
1746          U+0060c - U+0060d
1747          U+0061e - U+00612
1748          U+0064b - U+0065e
1749          U+0074d - U+0076d
1750          U+01800 - U+01805
1751          U+01d00 - U+01d77
1752          U+01d9b - U+01dbf
1753          U+0200b - U+0200f
1754          U+030fc - U+030fe
1755          U+03260 - U+0327f
1756          U+0fb46 - U+0fbb1
1757          U+10450 - U+1049d
1758    
1759    12. The -o option (show only the matching part of a line) for pcregrep was not
1760        compatible with GNU grep in that, if there was more than one match in a
1761        line, it showed only the first of them. It now behaves in the same way as
1762        GNU grep.
1763    
1764    13. If the -o and -v options were combined for pcregrep, it printed a blank
1765        line for every non-matching line. GNU grep prints nothing, and pcregrep now
1766        does the same. The return code can be used to tell if there were any
1767        non-matching lines.
1768    
1769    14. Added --file-offsets and --line-offsets to pcregrep.
1770    
1771    15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1772        infinitely looping recursion. The bug was that positive lookaheads were not
1773        being skipped when checking for a possible empty match (negative lookaheads
1774        and both kinds of lookbehind were skipped).
1775    
1776    16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1777        inclusion of <windows.h> to before rather than after the definition of
1778        INVALID_FILE_ATTRIBUTES (patch from David Byron).
1779    
1780    17. Specifying a possessive quantifier with a specific limit for a Unicode
1781        character property caused pcre_compile() to compile bad code, which led at
1782        runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1783        are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1784        caused the error; without that there was no problem.
1785    
1786    18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1787    
1788    19. Added --enable-pcretest-libreadline.
1789    
1790    20. In pcrecpp.cc, the variable 'count' was incremented twice in
1791        RE::GlobalReplace(). As a result, the number of replacements returned was
1792        double what it should be. I removed one of the increments, but Craig sent a
1793        later patch that removed the other one (the right fix) and added unit tests
1794        that check the return values (which was not done before).
1795    
1796    21. Several CMake things:
1797    
1798        (1) Arranged that, when cmake is used on Unix, the libraries end up with
1799            the names libpcre and libpcreposix, not just pcre and pcreposix.
1800    
1801        (2) The above change means that pcretest and pcregrep are now correctly
1802            linked with the newly-built libraries, not previously installed ones.
1803    
1804        (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1805    
1806    22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1807        crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1808        UTF-8 newline character). The key issue is that the pattern starts .*;
1809        this means that the match must be either at the beginning, or after a
1810        newline. The bug was in the code for advancing after a failed match and
1811        checking that the new position followed a newline. It was not taking
1812        account of UTF-8 characters correctly.
1813    
1814    23. PCRE was behaving differently from Perl in the way it recognized POSIX
1815        character classes. PCRE was not treating the sequence [:...:] as a
1816        character class unless the ... were all letters. Perl, however, seems to
1817        allow any characters between [: and :], though of course it rejects as
1818        unknown any "names" that contain non-letters, because all the known class
1819        names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1820        for example, whereas PCRE did not - it did not recognize a POSIX character
1821        class. This seemed a bit dangerous, so the code has been changed to be
1822        closer to Perl. The behaviour is not identical to Perl, because PCRE will
1823        diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1824        treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1825        Perl does, and where it didn't before.
1826    
1827    24. Rewrite so as to remove the single use of %n from pcregrep because in some
1828        Windows environments %n is disabled by default.
1829    
1830    
1831    Version 7.4 21-Sep-07
1832    ---------------------
1833    
1834    1.  Change 7.3/28 was implemented for classes by looking at the bitmap. This
1835        means that a class such as [\s] counted as "explicit reference to CR or
1836        LF". That isn't really right - the whole point of the change was to try to
1837        help when there was an actual mention of one of the two characters. So now
1838        the change happens only if \r or \n (or a literal CR or LF) character is
1839        encountered.
1840    
1841    2.  The 32-bit options word was also used for 6 internal flags, but the numbers
1842        of both had grown to the point where there were only 3 bits left.
1843        Fortunately, there was spare space in the data structure, and so I have
1844        moved the internal flags into a new 16-bit field to free up more option
1845        bits.
1846    
1847    3.  The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1848        but did not set the internal JCHANGED flag - either of these is enough to
1849        control the way the "get" function works - but the PCRE_INFO_JCHANGED
1850        facility is supposed to tell if (?J) was ever used, so now (?J) at the
1851        start sets both bits.
1852    
1853    4.  Added options (at build time, compile time, exec time) to change \R from
1854        matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1855    
1856    5.  doc/pcresyntax.html was missing from the distribution.
1857    
1858    6.  Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1859        compatibility, even though it is no longer used.
1860    
1861    7.  Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1862        strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1863        windows.h file is present (where different names are used). [This was
1864        reversed later after testing - see 16 below.]
1865    
1866    8.  Changed all #include <config.h> to #include "config.h". There were also
1867        some further <pcre.h> cases that I changed to "pcre.h".
1868    
1869    9.  When pcregrep was used with the --colour option, it missed the line ending
1870        sequence off the lines that it output.
1871    
1872    10. It was pointed out to me that arrays of string pointers cause lots of
1873        relocations when a shared library is dynamically loaded. A technique of
1874        using a single long string with a table of offsets can drastically reduce
1875        these. I have refactored PCRE in four places to do this. The result is
1876        dramatic:
1877    
1878          Originally:                          290
1879          After changing UCP table:            187
1880          After changing error message table:   43
1881          After changing table of "verbs"       36
1882          After changing table of Posix names   22
1883    
1884        Thanks to the folks working on Gregex for glib for this insight.
1885    
1886    11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1887        unicode-properties was also set.
1888    
1889    12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1890    
1891    13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1892        checked only for CRLF.
1893    
1894    14. Added casts to pcretest.c to avoid compiler warnings.
1895    
1896    15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1897    
1898    16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1899        and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1900        entirely. This removes changes made in 7 above.
1901    
1902    17. The CMake files have been updated, and there is now more information about
1903        building with CMake in the NON-UNIX-USE document.
1904    
1905    
1906    Version 7.3 28-Aug-07
1907    ---------------------
1908    
1909     1. In the rejigging of the build system that eventually resulted in 7.1, the
1910        line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1911        brackets there is not right, since it causes compilers to look for an
1912        installed pcre.h, not the version that is in the source that is being
1913        compiled (which of course may be different). I have changed it back to:
1914    
1915          #include "pcre.h"
1916    
1917        I have a vague recollection that the change was concerned with compiling in
1918        different directories, but in the new build system, that is taken care of
1919        by the VPATH setting the Makefile.
1920    
1921     2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1922        when the subject happened to end in the byte 0x85 (e.g. if the last
1923        character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1924        characters but of course it shouldn't be taken as a newline when it is part
1925        of another character. The bug was that, for an unlimited repeat of . in
1926        not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1927        characters when looking for a newline.
1928    
1929     3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1930    
1931     4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1932        in debug output.
1933    
1934     5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1935        long printing in the pcrecpp unittest when running under MinGW.
1936    
1937     6. ESC_K was left out of the EBCDIC table.
1938    
1939     7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1940        parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1941        limit also applies to "virtual nesting" when a pattern is recursive, and in
1942        this case 1000 isn't so big. I have been able to remove this limit at the
1943        expense of backing off one optimization in certain circumstances. Normally,
1944        when pcre_exec() would call its internal match() function recursively and
1945        immediately return the result unconditionally, it uses a "tail recursion"
1946        feature to save stack. However, when a subpattern that can match an empty
1947        string has an unlimited repetition quantifier, it no longer makes this
1948        optimization. That gives it a stack frame in which to save the data for
1949        checking that an empty string has been matched. Previously this was taken
1950        from the 1000-entry workspace that had been reserved. So now there is no
1951        explicit limit, but more stack is used.
1952    
1953     8. Applied Daniel's patches to solve problems with the import/export magic
1954        syntax that is required for Windows, and which was going wrong for the
1955        pcreposix and pcrecpp parts of the library. These were overlooked when this
1956        problem was solved for the main library.
1957    
1958     9. There were some crude static tests to avoid integer overflow when computing
1959        the size of patterns that contain repeated groups with explicit upper
1960        limits. As the maximum quantifier is 65535, the maximum group length was
1961        set at 30,000 so that the product of these two numbers did not overflow a
1962        32-bit integer. However, it turns out that people want to use groups that
1963        are longer than 30,000 bytes (though not repeat them that many times).
1964        Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1965        made it possible to implement the integer overflow checks in a much more
1966        dynamic way, which I have now done. The artificial limitation on group
1967        length has been removed - we now have only the limit on the total length of
1968        the compiled pattern, which depends on the LINK_SIZE setting.
1969    
1970    10. Fixed a bug in the documentation for get/copy named substring when
1971        duplicate names are permitted. If none of the named substrings are set, the
1972        functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1973        empty string.
1974    
1975    11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1976        instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1977        because the ] is interpreted as the first data character and the
1978        terminating ] is not found. PCRE has been made compatible with Perl in this
1979        regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1980        cause memory overwriting.
1981    
1982    10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1983        string has been matched (to stop an infinite loop). It was not recognizing
1984        a conditional subpattern that could match an empty string if that
1985        subpattern was within another subpattern. For example, it looped when
1986        trying to match  (((?(1)X|))*)  but it was OK with  ((?(1)X|)*)  where the
1987        condition was not nested. This bug has been fixed.
1988    
1989    12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1990        past the start of the subject in the presence of bytes with the top bit
1991        set, for example "\x8aBCD".
1992    
1993    13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1994        (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1995    
1996    14. Optimized (?!) to (*FAIL).
1997    
1998    15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1999        This restricts code points to be within the range 0 to 0x10FFFF, excluding
2000        the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2001        full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2002        does: it's just the validity check that is more restrictive.
2003    
2004    16. Inserted checks for integer overflows during escape sequence (backslash)
2005        processing, and also fixed erroneous offset values for syntax errors during
2006        backslash processing.
2007    
2008    17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2009        for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2010    
2011    18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2012        caused an overrun.
2013    
2014    19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2015        something other than just ASCII characters) inside a group that had an
2016        unlimited repeat caused a loop at compile time (while checking to see
2017        whether the group could match an empty string).
2018    
2019    20. Debugging a pattern containing \p or \P could cause a crash. For example,
2020        [\P{Any}] did so. (Error in the code for printing property names.)
2021    
2022    21. An orphan \E inside a character class could cause a crash.
2023    
2024    22. A repeated capturing bracket such as (A)? could cause a wild memory
2025        reference during compilation.
2026    
2027    23. There are several functions in pcre_compile() that scan along a compiled
2028        expression for various reasons (e.g. to see if it's fixed length for look
2029        behind). There were bugs in these functions when a repeated \p or \P was
2030        present in the pattern. These operators have additional parameters compared
2031        with \d, etc, and these were not being taken into account when moving along
2032        the compiled data. Specifically:
2033    
2034        (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2035            length.
2036    
2037        (b) An item such as \pL+ within a repeated group could cause crashes or
2038            loops.
2039    
2040        (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2041            "reference to non-existent subpattern" error.
2042    
2043        (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2044    
2045    24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2046        characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2047    
2048    25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2049    
2050    26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2051        character were causing crashes (broken optimization).
2052    
2053    27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2054        \p or \P) caused a compile-time loop.
2055    
2056    28. More problems have arisen in unanchored patterns when CRLF is a valid line
2057        break. For example, the unstudied pattern [\r\n]A does not match the string
2058        "\r\nA" because change 7.0/46 below moves the current point on by two
2059        characters after failing to match at the start. However, the pattern \nA
2060        *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2061        the same is true. There doesn't seem any very clean way out of this, but
2062        what I have chosen to do makes the common cases work: PCRE now takes note
2063        of whether there can be an explicit match for \r or \n anywhere in the
2064        pattern, and if so, 7.0/46 no longer applies. As part of this change,
2065        there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2066        pattern has explicit CR or LF references.
2067    
2068    29. Added (*CR) etc for changing newline setting at start of pattern.
2069    
2070    
2071    Version 7.2 19-Jun-07
2072    ---------------------
2073    
2074     1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2075        which is apparently normally available under Windows.
2076    
2077     2. Re-jig the pcregrep tests with different newline settings in an attempt
2078        to make them independent of the local environment's newline setting.
2079    
2080     3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2081    
2082     4. Some of the "internals" tests were previously cut out when the link size
2083        was not 2, because the output contained actual offsets. The recent new
2084        "Z" feature of pcretest means that these can be cut out, making the tests
2085        usable with all link sizes.
2086    
2087     5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2088        stack recursion. This gives a massive performance boost under BSD, but just
2089        a small improvement under Linux. However, it saves one field in the frame
2090        in all cases.
2091    
2092     6. Added more features from the forthcoming Perl 5.10:
2093    
2094        (a) (?-n) (where n is a string of digits) is a relative subroutine or
2095            recursion call. It refers to the nth most recently opened parentheses.
2096    
2097        (b) (?+n) is also a relative subroutine call; it refers to the nth next
2098            to be opened parentheses.
2099    
2100        (c) Conditions that refer to capturing parentheses can be specified
2101            relatively, for example, (?(-2)... or (?(+3)...
2102    
2103        (d) \K resets the start of the current match so that everything before
2104            is not part of it.
2105    
2106        (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2107    
2108        (f) \g{name} is another synonym - part of Perl 5.10's unification of
2109            reference syntax.
2110    
2111        (g) (?| introduces a group in which the numbering of parentheses in each
2112            alternative starts with the same number.
2113    
2114        (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2115    
2116     7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2117        PCRE_INFO_JCHANGED.
2118    
2119     8. A pattern such as  (.*(.)?)*  caused pcre_exec() to fail by either not
2120        terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2121        for detecting groups that can match an empty string.
2122    
2123     9. A pattern with a very large number of alternatives (more than several
2124        hundred) was running out of internal workspace during the pre-compile
2125        phase, where pcre_compile() figures out how much memory will be needed. A
2126        bit of new cunning has reduced the workspace needed for groups with
2127        alternatives. The 1000-alternative test pattern now uses 12 bytes of
2128        workspace instead of running out of the 4096 that are available.
2129    
2130    10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2131    
2132    11. Applied patch from Google to remove an optimization that didn't quite work.
2133        The report of the bug said:
2134    
2135          pcrecpp::RE("a*").FullMatch("aaa") matches, while
2136          pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2137          pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2138    
2139    12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2140        it matched the wrong number of bytes.
2141    
2142    
2143    Version 7.1 24-Apr-07
2144    ---------------------
2145    
2146     1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2147        that is more "standard", making use of automake and other Autotools. There
2148        is some re-arrangement of the files and adjustment of comments consequent
2149        on this.
2150    
2151     2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
2152        for recursive directory scanning broke on some systems because the files
2153        are not scanned in any specific order and on different systems the order
2154        was different. A call to "sort" has been inserted into RunGrepTest for the
2155        approprate test as a short-term fix. In the longer term there may be an
2156        alternative.
2157    
2158     3. I had an email from Eric Raymond about problems translating some of PCRE's
2159        man pages to HTML (despite the fact that I distribute HTML pages, some
2160        people do their own conversions for various reasons). The problems
2161        concerned the use of low-level troff macros .br and .in. I have therefore
2162        removed all such uses from the man pages (some were redundant, some could
2163        be replaced by .nf/.fi pairs). The 132html script that I use to generate
2164        HTML has been updated to handle .nf/.fi and to complain if it encounters
2165        .br or .in.
2166    
2167     4. Updated comments in configure.ac that get placed in config.h.in and also
2168        arranged for config.h to be included in the distribution, with the name
2169        config.h.generic, for the benefit of those who have to compile without
2170        Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2171    
2172     5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2173        Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2174        makevp.bat for the current PCRE, using the additional files
2175        makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2176    
2177     6. A Windows user reported a minor discrepancy with test 2, which turned out
2178        to be caused by a trailing space on an input line that had got lost in his
2179        copy. The trailing space was an accident, so I've just removed it.
2180    
2181     7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2182        that is needed.
2183    
2184     8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2185        as "const" (a) because they are and (b) because it helps the PHP
2186        maintainers who have recently made a script to detect big data structures
2187        in the php code that should be moved to the .rodata section. I remembered
2188        to update Builducptable as well, so it won't revert if ucptable.h is ever
2189        re-created.
2190    
2191     9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2192        pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2193        order to be able to cut out the UTF-8 tables in the latter when UTF-8
2194        support is not required. This saves 1.5-2K of code, which is important in
2195        some applications.
2196    
2197        Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2198        so as not to refer to the tables, even though these functions will never be
2199        called when UTF-8 support is disabled. Otherwise there are problems with a
2200        shared library.
2201    
2202    10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2203    
2204        (a) It was defining its arguments as char * instead of void *.
2205    
2206        (b) It was assuming that all moves were upwards in memory; this was true
2207            a long time ago when I wrote it, but is no longer the case.
2208    
2209        The emulated memove() is provided for those environments that have neither
2210        memmove() nor bcopy(). I didn't think anyone used it these days, but that
2211        is clearly not the case, as these two bugs were recently reported.
2212    
2213    11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2214        and Detrail to create the HTML documentation, the .txt form of the man
2215        pages, and it removes trailing spaces from listed files. It also creates
2216        pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2217        case, it wraps all the #defines with #ifndefs. This script should be run
2218        before "make dist".
2219    
2220    12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2221        with Unicode property support.
2222    
2223        (a) For a maximizing quantifier, if the two different cases of the
2224            character were of different lengths in their UTF-8 codings (there are
2225            some cases like this - I found 11), and the matching function had to
2226            back up over a mixture of the two cases, it incorrectly assumed they
2227            were both the same length.
2228    
2229        (b) When PCRE was configured to use the heap rather than the stack for
2230            recursion during matching, it was not correctly preserving the data for
2231            the other case of a UTF-8 character when checking ahead for a match
2232            while processing a minimizing repeat. If the check also involved
2233            matching a wide character, but failed, corruption could cause an
2234            erroneous result when trying to check for a repeat of the original
2235            character.
2236    
2237    13. Some tidying changes to the testing mechanism:
2238    
2239        (a) The RunTest script now detects the internal link size and whether there
2240            is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2241            values substituted by "configure". (The RunGrepTest script already did
2242            this for UTF-8.) The configure.ac script no longer substitutes the
2243            relevant variables.
2244    
2245        (b) The debugging options /B and /D in pcretest show the compiled bytecode
2246            with length and offset values. This means that the output is different
2247            for different internal link sizes. Test 2 is skipped for link sizes
2248            other than 2 because of this, bypassing the problem. Unfortunately,
2249            there was also a test in test 3 (the locale tests) that used /B and
2250            failed for link sizes other than 2. Rather than cut the whole test out,
2251            I have added a new /Z option to pcretest that replaces the length and
2252            offset values with spaces. This is now used to make test 3 independent
2253            of link size. (Test 2 will be tidied up later.)
2254    
2255    14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2256        segmentation fault instead of returning the appropriate error message.
2257    
2258    15. In multiline mode when the newline sequence was set to "any", the pattern
2259        ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2260        This doesn't seem right; it now treats the CRLF combination as the line
2261        ending, and so does not match in that case. It's only a pattern such as ^$
2262        that would hit this one: something like ^ABC$ would have failed after \r
2263        and then tried again after \r\n.
2264    
2265    16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2266        in an attempt to make files that differ only in their line terminators
2267        compare equal. This works on Linux.
2268    
2269    17. Under certain error circumstances pcregrep might try to free random memory
2270        as it exited. This is now fixed, thanks to valgrind.
2271    
2272    19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2273        "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2274        was because its rules for how to advance for /g after matching an empty
2275        string at the end of a line did not allow for this case. They now check for
2276        it specially.
2277    
2278    20. pcretest is supposed to handle patterns and data of any length, by
2279        extending its buffers when necessary. It was getting this wrong when the
2280        buffer for a data line had to be extended.
2281    
2282    21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2283        CRLF as a newline sequence.
2284    
2285    22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2286        out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2287        I have nevertheless tidied it up.
2288    
2289    23. Added some casts to kill warnings from HP-UX ia64 compiler.
2290    
2291    24. Added a man page for pcre-config.
2292    
2293    
2294    Version 7.0 19-Dec-06
2295    ---------------------
2296    
2297     1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2298        moving to gcc 4.1.1.
2299    
2300     2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2301        sys/time.h, which is documented as needed for this function. It doesn't
2302        seem to matter on Linux, but it showed up on some releases of OS X.
2303    
2304     3. It seems that there are systems where bytes whose values are greater than
2305        127 match isprint() in the "C" locale. The "C" locale should be the
2306        default when a C program starts up. In most systems, only ASCII printing
2307        characters match isprint(). This difference caused the output from pcretest
2308        to vary, making some of the tests fail. I have changed pcretest so that:
2309    
2310        (a) When it is outputting text in the compiled version of a pattern, bytes
2311            other than 32-126 are always shown as hex escapes.
2312    
2313        (b) When it is outputting text that is a matched part of a subject string,
2314            it does the same, unless a different locale has been set for the match
2315            (using the /L modifier). In this case, it uses isprint() to decide.
2316    
2317     4. Fixed a major bug that caused incorrect computation of the amount of memory
2318        required for a compiled pattern when options that changed within the
2319        pattern affected the logic of the preliminary scan that determines the
2320        length. The relevant options are -x, and -i in UTF-8 mode. The result was
2321        that the computed length was too small. The symptoms of this bug were
2322        either the PCRE error "internal error: code overflow" from pcre_compile(),
2323        or a glibc crash with a message such as "pcretest: free(): invalid next
2324        size (fast)". Examples of patterns that provoked this bug (shown in
2325        pcretest format) are:
2326    
2327          /(?-x: )/x
2328          /(?x)(?-x: \s*#\s*)/
2329          /((?i)[\x{c0}])/8
2330          /(?i:[\x{c0}])/8
2331    
2332        HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2333        is now done differently.
2334    
2335     5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2336        wrapper classes; (b) implement a new function in the C++ scanner that is
2337        more efficient than the old way of doing things because it avoids levels of
2338        recursion in the regex matching; (c) add a paragraph to the documentation
2339        for the FullMatch() function.
2340    
2341     6. The escape sequence \n was being treated as whatever was defined as
2342        "newline". Not only was this contrary to the documentation, which states
2343        that \n is character 10 (hex 0A), but it also went horribly wrong when
2344        "newline" was defined as CRLF. This has been fixed.
2345    
2346     7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2347        was being set to -1 for the "end of line" case (supposedly a value that no
2348        character can have). Though this value is never used (the check for end of
2349        line is "zero bytes in current character"), it caused compiler complaints.
2350        I've changed it to 0xffffffff.
2351    
2352     8. In pcre_version.c, the version string was being built by a sequence of
2353        C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2354        string (as it is for production releases) called a macro with an empty
2355        argument. The C standard says the result of this is undefined. The gcc
2356        compiler treats it as an empty string (which was what was wanted) but it is
2357        reported that Visual C gives an error. The source has been hacked around to
2358        avoid this problem.
2359    
2360     9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2361        builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2362        instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2363        of them did).
2364    
2365    10. Originally, pcretest opened its input and output without "b"; then I was
2366        told that "b" was needed in some environments, so it was added for release
2367        5.0 to both the input and output. (It makes no difference on Unix-like
2368        systems.) Later I was told that it is wrong for the input on Windows. I've
2369        now abstracted the modes into two macros, to make it easier to fiddle with
2370        them, and removed "b" from the input mode under Windows.
2371    
2372    11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2373    
2374    12. Added -help and --help to pcretest as an official way of being reminded
2375        of the options.
2376    
2377    13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2378        and pcrecpp.cc because they annoy compilers at high warning levels.
2379    
2380    14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2381    
2382    15. Fixed an occurrence of == in configure.ac that should have been = (shell
2383        scripts are not C programs :-) and which was not noticed because it works
2384        on Linux.
2385    
2386    16. pcretest is supposed to handle any length of pattern and data line (as one
2387        line or as a continued sequence of lines) by extending its input buffer if
2388        necessary. This feature was broken for very long pattern lines, leading to
2389        a string of junk being passed to pcre_compile() if the pattern was longer
2390        than about 50K.
2391    
2392    17. I have done a major re-factoring of the way pcre_compile() computes the
2393        amount of memory needed for a compiled pattern. Previously, there was code
2394        that made a preliminary scan of the pattern in order to do this. That was
2395        OK when PCRE was new, but as the facilities have expanded, it has become
2396        harder and harder to keep it in step with the real compile phase, and there
2397        have been a number of bugs (see for example, 4 above). I have now found a
2398        cunning way of running the real compile function in a "fake" mode that
2399        enables it to compute how much memory it would need, while actually only
2400        ever using a few hundred bytes of working memory and without too many
2401        tests of the mode. This should make future maintenance and development
2402        easier. A side effect of this work is that the limit of 200 on the nesting
2403        depth of parentheses has been removed (though this was never a serious
2404        limitation, I suspect). However, there is a downside: pcre_compile() now
2405        runs more slowly than before (30% or more, depending on the pattern). I
2406        hope this isn't a big issue. There is no effect on runtime performance.
2407    
2408    18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2409        newline (only possible for the last line of a file) and it was a
2410        pattern that set a locale (followed by /Lsomething), pcretest crashed.
2411    
2412    19. Added additional timing features to pcretest. (1) The -tm option now times
2413        matching only, not compiling. (2) Both -t and -tm can be followed, as a
2414        separate command line item, by a number that specifies the number of
2415        repeats to use when timing. The default is 50000; this gives better
2416        precision, but takes uncomfortably long for very large patterns.
2417    
2418    20. Extended pcre_study() to be more clever in cases where a branch of a
2419        subpattern has no definite first character. For example, (a*|b*)[cd] would
2420        previously give no result from pcre_study(). Now it recognizes that the
2421        first character must be a, b, c, or d.
2422    
2423    21. There was an incorrect error "recursive call could loop indefinitely" if
2424        a subpattern (or the entire pattern) that was being tested for matching an
2425        empty string contained only one non-empty item after a nested subpattern.
2426        For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2427        incorrectly, because the \d was being skipped in the check.
2428    
2429    22. The pcretest program now has a new pattern option /B and a command line
2430        option -b, which is equivalent to adding /B to every pattern. This causes
2431        it to show the compiled bytecode, without the additional information that
2432        -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2433        is the same as /B/I).
2434    
2435    23. A new optimization is now able automatically to treat some sequences such
2436        as a*b as a*+b. More specifically, if something simple (such as a character
2437        or a simple class like \d) has an unlimited quantifier, and is followed by
2438        something that cannot possibly match the quantified thing, the quantifier
2439        is automatically "possessified".
2440    
2441    24. A recursive reference to a subpattern whose number was greater than 39
2442        went wrong under certain circumstances in UTF-8 mode. This bug could also
2443        have affected the operation of pcre_study().
2444    
2445    25. Realized that a little bit of performance could be had by replacing
2446        (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2447    
2448    26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2449    
2450    27. Possessive quantifiers such as a++ were previously implemented by turning
2451        them into atomic groups such as ($>a+). Now they have their own opcodes,
2452        which improves performance. This includes the automatically created ones
2453        from 23 above.
2454    
2455    28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2456        lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2457        the first matched character to be a colon. This applied both to named and
2458        numbered groups.
2459    
2460    29. The ucpinternal.h header file was missing its idempotency #ifdef.
2461    
2462    30. I was sent a "project" file called libpcre.a.dev which I understand makes
2463        building PCRE on Windows easier, so I have included it in the distribution.
2464    
2465    31. There is now a check in pcretest against a ridiculously large number being
2466        returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2467        loop, the loop is abandoned.
2468    
2469    32. Forward references to subpatterns in conditions such as (?(2)...) where
2470        subpattern 2 is defined later cause pcre_compile() to search forwards in
2471        the pattern for the relevant set of parentheses. This search went wrong
2472        when there were unescaped parentheses in a character class, parentheses
2473        escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2474    
2475    33. "Subroutine" calls and backreferences were previously restricted to
2476        referencing subpatterns earlier in the regex. This restriction has now
2477        been removed.
2478    
2479    34. Added a number of extra features that are going to be in Perl 5.10. On the
2480        whole, these are just syntactic alternatives for features that PCRE had
2481        previously implemented using the Python syntax or my own invention. The
2482        other formats are all retained for compatibility.
2483    
2484        (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2485            as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2486            also .NET compatible.
2487    
2488        (b) A recursion or subroutine call to a named group can now be defined as
2489            (?&name) as well as (?P>name).
2490    
2491        (c) A backreference to a named group can now be defined as \k<name> or
2492            \k'name' as well as (?P=name). The new forms, as well as being in Perl
2493            5.10, are also .NET compatible.
2494    
2495        (d) A conditional reference to a named group can now use the syntax
2496            (?(<name>) or (?('name') as well as (?(name).
2497    
2498        (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2499            groups (named and numbered) that are never evaluated inline, but can be
2500            called as "subroutines" from elsewhere. In effect, the DEFINE condition
2501            is always false. There may be only one alternative in such a group.
2502    
2503        (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2504            as the simple (?(R). The condition is true only if the most recent
2505            recursion is that of the given number or name. It does not search out
2506            through the entire recursion stack.
2507    
2508        (g) The escape \gN or \g{N} has been added, where N is a positive or
2509            negative number, specifying an absolute or relative reference.
2510    
2511    35. Tidied to get rid of some further signed/unsigned compiler warnings and
2512        some "unreachable code" warnings.
2513    
2514    36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2515        things, this adds five new scripts.
2516    
2517    37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2518        There were also incompatibilities regarding the handling of \Q..\E inside
2519        character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2520        hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2521    
2522    38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2523        matches an empty string, and forcibly breaks the loop. There were bugs in
2524        this code in non-simple cases. For a pattern such as  ^(a()*)*  matched
2525        against  aaaa  the result was just "a" rather than "aaaa", for example. Two
2526        separate and independent bugs (that affected different cases) have been
2527        fixed.
2528    
2529    39. Refactored the code to abolish the use of different opcodes for small
2530        capturing bracket numbers. This is a tidy that I avoided doing when I
2531        removed the limit on the number of capturing brackets for 3.5 back in 2001.
2532        The new approach is not only tidier, it makes it possible to reduce the
2533        memory needed to fix the previous bug (38).
2534    
2535    40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2536        sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2537        processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2538        mode.
2539    
2540    41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2541        report.
2542    
2543    42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2544        copy construction and assignment in the C++ wrapper.
2545    
2546    43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2547        couple of bugs that could have given wrong results in the "--newline=crlf"
2548        case.
2549    
2550    44. Added a number of casts and did some reorganization of signed/unsigned int
2551        variables following suggestions from Dair Grant. Also renamed the variable
2552        "this" as "item" because it is a C++ keyword.
2553    
2554    45. Arranged for dftables to add
2555    
2556          #include "pcre_internal.h"
2557    
2558        to pcre_chartables.c because without it, gcc 4.x may remove the array
2559        definition from the final binary if PCRE is built into a static library and
2560        dead code stripping is activated.
2561    
2562    46. For an unanchored pattern, if a match attempt fails at the start of a
2563        newline sequence, and the newline setting is CRLF or ANY, and the next two
2564        characters are CRLF, advance by two characters instead of one.
2565    
2566    
2567    Version 6.7 04-Jul-06
2568    ---------------------
2569    
2570     1. In order to handle tests when input lines are enormously long, pcretest has
2571        been re-factored so that it automatically extends its buffers when
2572        necessary. The code is crude, but this _is_ just a test program. The
2573        default size has been increased from 32K to 50K.
2574    
2575     2. The code in pcre_study() was using the value of the re argument before
2576        testing it for NULL. (Of course, in any sensible call of the function, it
2577        won't be NULL.)
2578    
2579     3. The memmove() emulation function in pcre_internal.h, which is used on
2580        systems that lack both memmove() and bcopy() - that is, hardly ever -
2581        was missing a "static" storage class specifier.
2582    
2583     4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2584        containing an extended class (one that cannot be represented by a bitmap
2585        because it contains high-valued characters or Unicode property items, e.g.
2586        [\pZ]). Almost always one would set UTF-8 mode when processing such a
2587        pattern, but PCRE should not loop if you do not (it no longer does).
2588        [Detail: two cases were found: (a) a repeated subpattern containing an
2589        extended class; (b) a recursive reference to a subpattern that followed a
2590        previous extended class. It wasn't skipping over the extended class
2591        correctly when UTF-8 mode was not set.]
2592    
2593     5. A negated single-character class was not being recognized as fixed-length
2594        in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2595        compile error "lookbehind assertion is not fixed length".
2596    
2597     6. The RunPerlTest auxiliary script was showing an unexpected difference
2598        between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2599        write a Perl script that can interpret lines of an input file either as
2600        byte characters or as UTF-8, which is what "perltest" was being required to
2601        do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2602        can't do is switch easily at run time between having the "use utf8;" pragma
2603        or not. In the end, I fudged it by using the RunPerlTest script to insert
2604        "use utf8;" explicitly for the UTF-8 tests.
2605    
2606     7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2607        the end of the subject string, contrary to the documentation and to what
2608        Perl does. This was true of both matching functions. Now it matches only at
2609        the start of the subject and immediately after *internal* newlines.
2610    
2611     8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2612        a pointer to an int instead of a pointer to an unsigned long int. This
2613        caused problems on 64-bit systems.
2614    
2615     9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2616        instance of the 'standard' template library not being so standard".
2617    
2618    10. There was no check on the number of named subpatterns nor the maximum
2619        length of a subpattern name. The product of these values is used to compute
2620        the size of the memory block for a compiled pattern. By supplying a very
2621        long subpattern name and a large number of named subpatterns, the size
2622        computation could be caused to overflow. This is now prevented by limiting
2623        the length of names to 32 characters, and the number of named subpatterns
2624        to 10,000.
2625    
2626    11. Subpatterns that are repeated with specific counts have to be replicated in
2627        the compiled pattern. The size of memory for this was computed from the
2628        length of the subpattern and the repeat count. The latter is limited to
2629        65535, but there was no limit on the former, meaning that integer overflow
2630        could in principle occur. The compiled length of a repeated subpattern is
2631        now limited to 30,000 bytes in order to prevent this.
2632    
2633    12. Added the optional facility to have named substrings with the same name.
2634    
2635    13. Added the ability to use a named substring as a condition, using the
2636        Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2637        are numbers (not recommended). Forward references are permitted.
2638    
2639    14. Added forward references in named backreferences (if you see what I mean).
2640    
2641    15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2642        pattern could run off the end of the subject. For example, the pattern
2643        "(?s)(.{1,5})"8 did this with the subject "ab".
2644    
2645    16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2646        PCRE_CASELESS was set when matching characters that were quantified with ?
2647        or *.
2648    
2649    17. A character class other than a single negated character that had a minimum
2650        but no maximum quantifier - for example [ab]{6,} - was not handled
2651        correctly by pce_dfa_exec(). It would match only one character.
2652    
2653    18. A valid (though odd) pattern that looked like a POSIX character
2654        class but used an invalid character after [ (for example [[,abc,]]) caused
2655        pcre_compile() to give the error "Failed: internal error: code overflow" or
2656        in some cases to crash with a glibc free() error. This could even happen if
2657        the pattern terminated after [[ but there just happened to be a sequence of
2658        letters, a binary zero, and a closing ] in the memory that followed.
2659    
2660    19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2661        over the years. Originally (before any Unicode support), just the bottom 8
2662        bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2663        output from "man perlunicode" includes this:
2664    
2665          The regular expression compiler produces polymorphic opcodes.  That
2666          is, the pattern adapts to the data and automatically switches to
2667          the Unicode character scheme when presented with Unicode data--or
2668          instead uses a traditional byte scheme when presented with byte
2669          data.
2670    
2671        Sadly, a wide octal escape does not cause a switch, and in a string with
2672        no other multibyte characters, these octal escapes are treated as before.
2673        Thus, in Perl, the pattern  /\500/ actually matches \100 but the pattern
2674        /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2675        Unicode string.
2676    
2677        I have not perpetrated such confusion in PCRE. Up till now, it took just
2678        the bottom 8 bits, as in old Perl. I have now made octal escapes with
2679        values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2680        translate to the appropriate multibyte character.
2681    
2682    29. Applied some refactoring to reduce the number of warnings from Microsoft
2683        and Borland compilers. This has included removing the fudge introduced
2684        seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2685        a warning about an unused variable.
2686    
2687    21. PCRE has not included VT (character 0x0b) in the set of whitespace
2688        characters since release 4.0, because Perl (from release 5.004) does not.
2689        [Or at least, is documented not to: some releases seem to be in conflict
2690        with the documentation.] However, when a pattern was studied with
2691        pcre_study() and all its branches started with \s, PCRE still included VT
2692        as a possible starting character. Of course, this did no harm; it just
2693        caused an unnecessary match attempt.
2694    
2695    22. Removed a now-redundant internal flag bit that recorded the fact that case
2696        dependency changed within the pattern. This was once needed for "required
2697        byte" processing, but is no longer used. This recovers a now-scarce options
2698        bit. Also moved the least significant internal flag bit to the most-
2699        significant bit of the word, which was not previously used (hangover from
2700        the days when it was an int rather than a uint) to free up another bit for
2701        the future.
2702    
2703    23. Added support for CRLF line endings as well as CR and LF. As well as the
2704        default being selectable at build time, it can now be changed at runtime
2705        via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2706        specify that it is scanning data with non-default line endings.
2707    
2708    24. Changed the definition of CXXLINK to make it agree with the definition of
2709        LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2710    
2711    25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2712        recursions. This makes a big different to stack usage for some patterns.
2713    
2714    26. If a subpattern containing a named recursion or subroutine reference such
2715        as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2716        the space required for the compiled pattern went wrong and gave too small a
2717        value. Depending on the environment, this could lead to "Failed: internal
2718        error: code overflow at offset 49" or "glibc detected double free or
2719        corruption" errors.
2720    
2721    27. Applied patches from Google (a) to support the new newline modes and (b) to
2722        advance over multibyte UTF-8 characters in GlobalReplace.
2723    
2724    28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2725        difference for some implementation of PCRE in some Windows version.
2726    
2727    29. Added some extra testing facilities to pcretest:
2728    
2729        \q<number>   in a data line sets the "match limit" value
2730        \Q<number>   in a data line sets the "match recursion limt" value
2731        -S <number>  sets the stack size, where <number> is in megabytes
2732    
2733        The -S option isn't available for Windows.
2734    
2735    
2736    Version 6.6 06-Feb-06
2737    ---------------------
2738    
2739     1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2740        in pcreposix.h. I have copied the definition from pcre.h.
2741    
2742     2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2743        because pcre.h is no longer a built file.
2744    
2745     3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2746        not normally included in the compiled code.
2747    
2748    
2749    Version 6.5 01-Feb-06
2750    ---------------------
2751    
2752     1. When using the partial match feature with pcre_dfa_exec(), it was not
2753        anchoring the second and subsequent partial matches at the new starting
2754        point. This could lead to incorrect results. For example, with the pattern
2755        /1234/, partially matching against "123" and then "a4" gave a match.
2756    
2757     2. Changes to pcregrep:
2758    
2759        (a) All non-match returns from pcre_exec() were being treated as failures
2760            to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2761            error message is output. Some extra information is given for the
2762            PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2763            probably the only errors that are likely to be caused by users (by
2764            specifying a regex that has nested indefinite repeats, for instance).
2765            If there are more than 20 of these errors, pcregrep is abandoned.
2766    
2767        (b) A binary zero was treated as data while matching, but terminated the
2768            output line if it was written out. This has been fixed: binary zeroes
2769            are now no different to any other data bytes.
2770    
2771        (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2772            used to set a locale for matching. The --locale=xxxx long option has
2773            been added (no short equivalent) to specify a locale explicitly on the
2774            pcregrep command, overriding the environment variables.
2775    
2776        (d) When -B was used with -n, some line numbers in the output were one less
2777            than they should have been.
2778    
2779        (e) Added the -o (--only-matching) option.
2780    
2781        (f) If -A or -C was used with -c (count only), some lines of context were
2782            accidentally printed for the final match.
2783    
2784        (g) Added the -H (--with-filename) option.
2785    
2786        (h) The combination of options -rh failed to suppress file names for files
2787            that were found from directory arguments.
2788    
2789        (i) Added the -D (--devices) and -d (--directories) options.
2790    
2791        (j) Added the -F (--fixed-strings) option.
2792    
2793        (k) Allow "-" to be used as a file name for -f as well as for a data file.
2794    
2795        (l) Added the --colo(u)r option.
2796    
2797        (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2798            is not present by default.
2799    
2800     3. A nasty bug was discovered in the handling of recursive patterns, that is,
2801        items such as (?R) or (?1), when the recursion could match a number of
2802        alternatives. If it matched one of the alternatives, but subsequently,
2803        outside the recursion, there was a failure, the code tried to back up into
2804        the recursion. However, because of the way PCRE is implemented, this is not
2805        possible, and the result was an incorrect result from the match.
2806    
2807        In order to prevent this happening, the specification of recursion has
2808        been changed so that all such subpatterns are automatically treated as
2809        atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2810    
2811     4. I had overlooked the fact that, in some locales, there are characters for
2812        which isalpha() is true but neither isupper() nor islower() are true. In
2813        the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2814        and ordfeminine) are like this. This affected the treatment of \w and \W
2815        when they appeared in character classes, but not when they appeared outside
2816        a character class. The bit map for "word" characters is now created
2817        separately from the results of isalnum() instead of just taking it from the
2818        upper, lower, and digit maps. (Plus the underscore character, of course.)
2819    
2820     5. The above bug also affected the handling of POSIX character classes such as
2821        [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2822        permanent tables. Instead, the bit maps for such a class were previously
2823        created as the appropriate unions of the upper, lower, and digit bitmaps.
2824        Now they are created by subtraction from the [[:word:]] class, which has
2825        its own bitmap.
2826    
2827     6. The [[:blank:]] character class matches horizontal, but not vertical space.
2828        It is created by subtracting the vertical space characters (\x09, \x0a,
2829        \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2830        subtraction was done in the overall bitmap for a character class, meaning
2831        that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2832        be recognized. This bug has been fixed.
2833    
2834     7. Patches from the folks at Google:
2835    
2836          (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2837          real life, but is still worth protecting against".
2838    
2839          (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2840          regular expressions".
2841    
2842          (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2843          have it.
2844    
2845          (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2846          "configure" and the latter not, in order to fix a problem somebody had
2847          with compiling the Arg class on HP-UX.
2848    
2849          (e) Improve the error-handling of the C++ wrapper a little bit.
2850    
2851          (f) New tests for checking recursion limiting.
2852    
2853     8. The pcre_memmove() function, which is used only if the environment does not
2854        have a standard memmove() function (and is therefore rarely compiled),
2855        contained two bugs: (a) use of int instead of size_t, and (b) it was not
2856        returning a result (though PCRE never actually uses the result).
2857    
2858     9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2859        large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2860        returned instead of calling malloc() with an overflowing number that would
2861        most likely cause subsequent chaos.
2862    
2863    10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2864    
2865    11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2866        with this option is matched, the nmatch and pmatch options of regexec() are
2867        ignored.
2868    
2869    12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2870        provided in case anyone wants to the the POSIX interface with UTF-8
2871        strings.
2872    
2873    13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2874        C++ linking (needed for some HP-UX environments).
2875    
2876    14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2877        (unused parameter) and in the pcre_printint() function (omitted "default"
2878        switch label when the default is to do nothing).
2879    
2880    15. Added some code to make it possible, when PCRE is compiled as a C++
2881        library, to replace subject pointers for pcre_exec() with a smart pointer
2882        class, thus making it possible to process discontinuous strings.
2883    
2884    16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2885        much the same function. They were added by different people who were trying
2886        to make PCRE easy to compile on non-Unix systems. It has been suggested
2887        that PCRE_EXPORT be abolished now that there is more automatic apparatus
2888        for compiling on Windows systems. I have therefore replaced it with
2889        PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2890        defaults to "extern" for C or "extern C" for C++, which works fine on
2891        Unix-like systems. It is now possible to override the value of PCRE_DATA_
2892        SCOPE with something explicit in config.h. In addition:
2893    
2894        (a) pcreposix.h still had just "extern" instead of either of these macros;
2895            I have replaced it with PCRE_DATA_SCOPE.
2896    
2897        (b) Functions such as _pcre_xclass(), which are internal to the library,
2898            but external in the C sense, all had PCRE_EXPORT in their definitions.
2899            This is apparently wrong for the Windows case, so I have removed it.
2900            (It makes no difference on Unix-like systems.)
2901    
2902    17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2903        of recursive calls to match(). This is different to MATCH_LIMIT because
2904        that limits the total number of calls to match(), not all of which increase
2905        the depth of recursion. Limiting the recursion depth limits the amount of
2906        stack (or heap if NO_RECURSE is set) that is used. The default can be set
2907        when PCRE is compiled, and changed at run time. A patch from Google adds
2908        this functionality to the C++ interface.
2909    
2910    18. Changes to the handling of Unicode character properties:
2911    
2912        (a) Updated the table to Unicode 4.1.0.
2913    
2914        (b) Recognize characters that are not in the table as "Cn" (undefined).
2915    
2916        (c) I revised the way the table is implemented to a much improved format
2917            which includes recognition of ranges. It now supports the ranges that
2918            are defined in UnicodeData.txt, and it also amalgamates other
2919            characters into ranges. This has reduced the number of entries in the
2920            table from around 16,000 to around 3,000, thus reducing its size
2921            considerably. I realized I did not need to use a tree structure after
2922            all - a binary chop search is just as efficient. Having reduced the
2923            number of entries, I extended their size from 6 bytes to 8 bytes to
2924            allow for more data.
2925    
2926        (d) Added support for Unicode script names via properties such as \p{Han}.
2927    
2928    19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2929        matching that character.
2930    
2931    20. When matching a repeated Unicode property with a minimum greater than zero,
2932        (for example \pL{2,}), PCRE could look past the end of the subject if it
2933        reached it while seeking the minimum number of characters. This could
2934        happen only if some of the characters were more than one byte long, because
2935        there is a check for at least the minimum number of bytes.
2936    
2937    21. Refactored the implementation of \p and \P so as to be more general, to
2938        allow for more different types of property in future. This has changed the
2939        compiled form incompatibly. Anybody with saved compiled patterns that use
2940        \p or \P will have to recompile them.
2941    
2942    22. Added "Any" and "L&" to the supported property types.
2943    
2944    23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2945        but give a compile time error if the value is greater than 0xff.
2946    
2947    24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2948        accidentally not being installed or uninstalled.
2949    
2950    25. The pcre.h file was built from pcre.h.in, but the only changes that were
2951        made were to insert the current release number. This seemed silly, because
2952        it made things harder for people building PCRE on systems that don't run
2953        "configure". I have turned pcre.h into a distributed file, no longer built
2954        by "configure", with the version identification directly included. There is
2955        no longer a pcre.h.in file.
2956    
2957        However, this change necessitated a change to the pcre-config script as
2958        well. It is built from pcre-config.in, and one of the substitutions was the
2959        release number. I have updated configure.ac so that ./configure now finds
2960        the release number by grepping pcre.h.
2961    
2962    26. Added the ability to run the tests under valgrind.
2963    
2964    
2965    Version 6.4 05-Sep-05
2966    ---------------------
2967    
2968     1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2969        "--" to be printed when multiple files were scanned, even when none of the
2970        -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2971        consider it to be a bug, and have restored the previous behaviour.
2972    
2973     2. A couple of code tidies to get rid of compiler warnings.
2974    
2975     3. The pcretest program used to cheat by referring to symbols in the library
2976        whose names begin with _pcre_. These are internal symbols that are not
2977        really supposed to be visible externally, and in some environments it is
2978        possible to suppress them. The cheating is now confined to including
2979        certain files from the library's source, which is a bit cleaner.
2980    
2981     4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2982        file's purpose clearer.
2983    
2984     5. Reorganized pcre_ucp_findchar().
2985    
2986    
2987    Version 6.3 15-Aug-05
2988    ---------------------
2989    
2990     1. The file libpcre.pc.in did not have general read permission in the tarball.
2991    
2992     2. There were some problems when building without C++ support:
2993    
2994        (a) If C++ support was not built, "make install" and "make test" still
2995            tried to test it.
2996    
2997        (b) There were problems when the value of CXX was explicitly set. Some
2998            changes have been made to try to fix these, and ...
2999    
3000        (c) --disable-cpp can now be used to explicitly disable C++ support.
3001    
3002        (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3003            backslash in a target when C++ was disabled. This confuses some
3004            versions of "make", apparently. Using an intermediate variable solves
3005            this. (Same for CPP_LOBJ.)
3006    
3007     3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3008        (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3009        necessary on certain architectures.
3010    
3011     4. Added a setting of -export-symbols-regex to the link command to remove
3012        those symbols that are exported in the C sense, but actually are local
3013        within the library, and not documented. Their names all begin with
3014        "_pcre_". This is not a perfect job, because (a) we have to except some
3015        symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3016        available (and never for static libraries). I have made a note to try to
3017        find a way round (a) in the future.
3018    
3019    
3020    Version 6.2 01-Aug-05
3021    ---------------------
3022    
3023     1. There was no test for integer overflow of quantifier values. A construction
3024        such as {1111111111111111} would give undefined results. What is worse, if
3025        a minimum quantifier for a parenthesized subpattern overflowed and became
3026        negative, the calculation of the memory size went wrong. This could have
3027        led to memory overwriting.
3028    
3029     2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3030    
3031     3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3032        operating environments where this matters.
3033    
3034     4. Applied Giuseppe Maxia's patch to add additional features for controlling
3035        PCRE options from within the C++ wrapper.
3036    
3037     5. Named capturing subpatterns were not being correctly counted when a pattern
3038        was compiled. This caused two problems: (a) If there were more than 100
3039        such subpatterns, the calculation of the memory needed for the whole
3040        compiled pattern went wrong, leading to an overflow error. (b) Numerical
3041        back references of the form \12, where the number was greater than 9, were
3042        not recognized as back references, even though there were sufficient
3043        previous subpatterns.
3044    
3045     6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3046        versions of gcc, e.g. 2.95.4.
3047    
3048    
3049    Version 6.1 21-Jun-05
3050    ---------------------
3051    
3052     1. There was one reference to the variable "posix" in pcretest.c that was not
3053        surrounded by "#if !defined NOPOSIX".
3054    
3055     2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3056        the cross-check on the old pcre_info() function, for the benefit of the
3057        cut-down version of PCRE that is currently imported into Exim.
3058    
3059     3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3060        allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3061        patterns that start (?i) but I don't think that matters. The use of (?i) is
3062        just an example; this all applies to the other options as well.
3063    
3064     4. Since libtool seems to echo the compile commands it is issuing, the output
3065        from "make" can be reduced a bit by putting "@" in front of each libtool
3066        compile command.
3067    
3068     5. Patch from the folks at Google for configure.in to be a bit more thorough
3069        in checking for a suitable C++ installation before trying to compile the
3070        C++ stuff. This should fix a reported problem when a compiler was present,
3071        but no suitable headers.
3072    
3073     6. The man pages all had just "PCRE" as their title. I have changed them to
3074        be the relevant file name. I have also arranged that these names are
3075        retained in the file doc/pcre.txt, which is a concatenation in text format
3076        of all the man pages except the little individual ones for each function.
3077    
3078     7. The NON-UNIX-USE file had not been updated for the different set of source
3079        files that come with release 6. I also added a few comments about the C++
3080        wrapper.
3081    
3082    
3083    Version 6.0 07-Jun-05
3084    ---------------------
3085    
3086     1. Some minor internal re-organization to help with my DFA experiments.
3087    
3088     2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3089        didn't matter for the library itself when fully configured, but did matter
3090        when compiling without UCP support, or within Exim, where the ucp files are
3091        not imported.
3092    
3093     3. Refactoring of the library code to split up the various functions into
3094        different source modules. The addition of the new DFA matching code (see
3095        below) to a single monolithic source would have made it really too
3096        unwieldy, quite apart from causing all the code to be include in a
3097        statically linked application, when only some functions are used. This is
3098        relevant even without the DFA addition now that patterns can be compiled in
3099        one application and matched in another.
3100    
3101        The downside of splitting up is that there have to be some external
3102        functions and data tables that are used internally in different modules of
3103        the library but which are not part of the API. These have all had their
3104        names changed to start with "_pcre_" so that they are unlikely to clash
3105        with other external names.
3106    
3107     4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3108        a different (DFA) algorithm. Although it is slower than the original
3109        function, it does have some advantages for certain types of matching
3110        problem.
3111    
3112     5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3113        including restarting after a partial match.
3114    
3115     6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3116        defined when compiling for Windows was sent to me. I have put it into the
3117        code, though I have no means of testing or verifying it.
3118    
3119     7. Added the pcre_refcount() auxiliary function.
3120    
3121     8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3122        match before or at the first newline in the subject string. In pcretest,
3123        the /f option on a pattern can be used to set this.
3124    
3125     9. A repeated \w when used in UTF-8 mode with characters greater than 256
3126        would behave wrongly. This has been present in PCRE since release 4.0.
3127    
3128    10. A number of changes to the pcregrep command:
3129    
3130        (a) Refactored how -x works; insert ^(...)$ instead of setting
3131            PCRE_ANCHORED and checking the length, in preparation for adding
3132            something similar for -w.
3133    
3134        (b) Added the -w (match as a word) option.
3135    
3136        (c) Refactored the way lines are read and buffered so as to have more
3137            than one at a time available.
3138    
3139        (d) Implemented a pcregrep test script.
3140    
3141        (e) Added the -M (multiline match) option. This allows patterns to match
3142            over several lines of the subject. The buffering ensures that at least
3143            8K, or the rest of the document (whichever is the shorter) is available
3144            for matching (and similarly the previous 8K for lookbehind assertions).
3145    
3146        (f) Changed the --help output so that it now says
3147    
3148              -w, --word-regex(p)
3149    
3150            instead of two lines, one with "regex" and the other with "regexp"
3151            because that confused at least one person since the short forms are the
3152            same. (This required a bit of code, as the output is generated
3153            automatically from a table. It wasn't just a text change.)
3154    
3155        (g) -- can be used to terminate pcregrep options if the next thing isn't an
3156            option but starts with a hyphen. Could be a pattern or a path name
3157            starting with a hyphen, for instance.
3158    
3159        (h) "-" can be given as a file name to represent stdin.
3160    
3161        (i) When file names are being printed, "(standard input)" is used for
3162            the standard input, for compatibility with GNU grep. Previously
3163            "<stdin>" was used.
3164    
3165        (j) The option --label=xxx can be used to supply a name to be used for
3166            stdin when file names are being printed. There is no short form.
3167    
3168        (k) Re-factored the options decoding logic because we are going to add
3169            two more options that take data. Such options can now be given in four
3170            different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3171    
3172        (l) Added the -A, -B, and -C options for requesting that lines of context
3173            around matches be printed.
3174    
3175        (m) Added the -L option to print the names of files that do not contain
3176            any matching lines, that is, the complement of -l.
3177    
3178        (n) The return code is 2 if any file cannot be opened, but pcregrep does
3179            continue to scan other files.
3180    
3181        (o) The -s option was incorrectly implemented. For compatibility with other
3182            greps, it now suppresses the error message for a non-existent or non-
3183            accessible file (but not the return code). There is a new option called
3184            -q that suppresses the output of matching lines, which was what -s was
3185            previously doing.
3186    
3187        (p) Added --include and --exclude options to specify files for inclusion
3188            and exclusion when recursing.
3189    
3190    11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3191        Hopefully, it now does.
3192    
3193    12. Missing cast in pcre_study().
3194    
3195    13. Added an "uninstall" target to the makefile.
3196    
3197    14. Replaced "extern" in the function prototypes in Makefile.in with
3198        "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3199        world, but is set differently for Windows.
3200    
3201    15. Added a second compiling function called pcre_compile2(). The only
3202        difference is that it has an extra argument, which is a pointer to an
3203        integer error code. When there is a compile-time failure, this is set
3204        non-zero, in addition to the error test pointer being set to point to an
3205        error message. The new argument may be NULL if no error number is required
3206        (but then you may as well call pcre_compile(), which is now just a
3207        wrapper). This facility is provided because some applications need a
3208        numeric error indication, but it has also enabled me to tidy up the way
3209        compile-time errors are handled in the POSIX wrapper.
3210    
3211    16. Added VPATH=.libs to the makefile; this should help when building with one
3212        prefix path and installing with another. (Or so I'm told by someone who
3213        knows more about this stuff than I do.)
3214    
3215    17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3216        passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3217        match everything, including newlines. This is not POSIX-compatible, but
3218        somebody wanted the feature. From pcretest it can be activated by using
3219        both the P and the s flags.
3220    
3221    18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3222    
3223    19. libpcre.pc was being incorrectly installed as executable.
3224    
3225    20. A couple of places in pcretest check for end-of-line by looking for '\n';
3226        it now also looks for '\r' so that it will work unmodified on Windows.
3227    
3228    21. Added Google's contributed C++ wrapper to the distribution.
3229    
3230    22. Added some untidy missing memory free() calls in pcretest, to keep
3231        Electric Fence happy when testing.
3232    
3233    
3234    
3235    Version 5.0 13-Sep-04
3236    ---------------------
3237    
3238     1. Internal change: literal characters are no longer packed up into items
3239        containing multiple characters in a single byte-string. Each character
3240        is now matched using a separate opcode. However, there may be more than one
3241        byte in the character in UTF-8 mode.
3242    
3243     2. The pcre_callout_block structure has two new fields: pattern_position and
3244        next_item_length. These contain the offset in the pattern to the next match
3245        item, and its length, respectively.
3246    
3247     3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3248        insertion of callouts before each pattern item. Added the /C option to
3249        pcretest to make use of this.
3250    
3251     4. On the advice of a Windows user, the lines
3252    
3253          #if defined(_WIN32) || defined(WIN32)
3254          _setmode( _fileno( stdout ), 0x8000 );
3255          #endif  /* defined(_WIN32) || defined(WIN32) */
3256    
3257        have been added to the source of pcretest. This apparently does useful
3258        magic in relation to line terminators.
3259    
3260     5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3261        for the benefit of those environments where the "b" makes a difference.
3262    
3263     6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3264        to know about it. I have put a hack into configure.in that adds in code
3265        to set GCC=yes if CC=icc. This seems to end up at a point in the
3266        generated configure script that is early enough to affect the setting of
3267        compiler options, which is what is needed, but I have no means of testing
3268        whether it really works. (The user who reported this had patched the
3269        generated configure script, which of course I cannot do.)
3270    
3271        LATER: After change 22 below (new libtool files), the configure script
3272        seems to know about icc (and also ecc). Therefore, I have commented out
3273        this hack in configure.in.
3274    
3275     7. Added support for pkg-config (2 patches were sent in).
3276    
3277     8. Negated POSIX character classes that used a combination of internal tables
3278        were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3279        [[:^ascii]]. Typically, they would match almost any characters. The other
3280        POSIX classes were not broken in this way.
3281    
3282     9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3283        to find the match, as PCRE was deluded into thinking that the match had to
3284        start at the start point or following a newline. The same bug applied to
3285        patterns with negative forward assertions or any backward assertions
3286        preceding ".*" at the start, unless the pattern required a fixed first
3287        character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3288    
3289    10. In UTF-8 mode, when moving forwards in the subject after a failed match
3290        starting at the last subject character, bytes beyond the end of the subject
3291        string were read.
3292    
3293    11. Renamed the variable "class" as "classbits" to make life easier for C++
3294        users. (Previously there was a macro definition, but it apparently wasn't
3295        enough.)
3296    
3297    12. Added the new field "tables" to the extra data so that tables can be passed
3298        in at exec time, or the internal tables can be re-selected. This allows
3299        a compiled regex to be saved and re-used at a later time by a different
3300        program that might have everything at different addresses.
3301    
3302    13. Modified the pcre-config script so that, when run on Solaris, it shows a
3303        -R library as well as a -L library.
3304    
3305    14. The debugging options of pcretest (-d on the command line or D on a
3306        pattern) showed incorrect output for anything following an extended class
3307        that contained multibyte characters and which was followed by a quantifier.
3308    
3309    15. Added optional support for general category Unicode character properties
3310        via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3311        support. It adds about 90K to the size of the library. The meanings of the
3312        inbuilt class escapes such as \d and \s have NOT been changed.
3313    
3314    16. Updated pcredemo.c to include calls to free() to release the memory for the
3315        compiled pattern.
3316    
3317    17. The generated file chartables.c was being created in the source directory
3318        instead of in the building directory. This caused the build to fail if the
3319        source directory was different from the building directory, and was
3320        read-only.
3321    
3322    18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3323        file. No doubt somebody will tell me if they don't make sense... Also added
3324        Dan Mooney's comments about building on OpenVMS.
3325    
3326    19. Added support for partial matching via the PCRE_PARTIAL option for
3327        pcre_exec() and the \P data escape in pcretest.
3328    
3329    20. Extended pcretest with 3 new pattern features:
3330    
3331        (i)   A pattern option of the form ">rest-of-line" causes pcretest to
3332              write the compiled pattern to the file whose name is "rest-of-line".
3333              This is a straight binary dump of the data, with the saved pointer to
3334              the character tables forced to be NULL. The study data, if any, is
3335              written too. After writing, pcretest reads a new pattern.
3336    
3337        (ii)  If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3338              compiled pattern from the given file. There must not be any
3339              occurrences of "<" in the file name (pretty unlikely); if there are,
3340              pcretest will instead treat the initial "<" as a pattern delimiter.
3341              After reading in the pattern, pcretest goes on to read data lines as
3342              usual.
3343    
3344        (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3345              and 16-bit fields in a compiled pattern, to simulate a pattern that
3346              was compiled on a host of opposite endianness.
3347    
3348    21. The pcre-exec() function can now cope with patterns that were compiled on
3349        hosts of opposite endianness, with this restriction:
3350    
3351          As for any compiled expression that is saved and used later, the tables
3352          pointer field cannot be preserved; the extra_data field in the arguments
3353          to pcre_exec() should be used to pass in a tables address if a value
3354          other than the default internal tables were used at compile time.
3355    
3356    22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3357        now diagnosed as an error. Previously, most of the time, a negative number
3358        would have been treated as zero, but if in addition "ovector" was passed as
3359        NULL, a crash could occur.
3360    
3361    23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3362        new versions from the libtool 1.5 distribution (the last one is a copy of
3363        a file called libtool.m4). This seems to have fixed the need to patch
3364        "configure" to support Darwin 1.3 (which I used to do). However, I still
3365        had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3366        workstation).
3367    
3368    24. Changed the PCRE licence to be the more standard "BSD" licence.
3369    
3370    
3371    Version 4.5 01-Dec-03
3372    ---------------------
3373    
3374     1. There has been some re-arrangement of the code for the match() function so
3375        that it can be compiled in a version that does not call itself recursively.
3376        Instead, it keeps those local variables that need separate instances for
3377        each "recursion" in a frame on the heap, and gets/frees frames whenever it
3378        needs to "recurse". Keeping track of where control must go is done by means
3379        of setjmp/longjmp. The whole thing is implemented by a set of macros that
3380        hide most of the details from the main code, and operates only if
3381        NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3382        "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3383        operating.
3384    
3385        To make it easier for callers to provide specially tailored get/free
3386        functions for this usage, two new functions, pcre_stack_malloc, and
3387        pcre_stack_free, are used. They are always called in strict stacking order,
3388        and the size of block requested is always the same.
3389    
3390        The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3391        PCRE has been compiled to use the stack or the heap for recursion. The
3392        -C option of pcretest uses this to show which version is compiled.
3393    
3394        A new data escape \S, is added to pcretest; it causes the amounts of store
3395        obtained and freed by both kinds of malloc/free at match time to be added
3396        to the output.
3397    
3398     2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3399        what's available on my current Linux desktop machine.
3400    
3401     3. When matching a UTF-8 string, the test for a valid string at the start has
3402        been extended. If start_offset is not zero, PCRE now checks that it points
3403        to a byte that is the start of a UTF-8 character. If not, it returns
3404        PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3405        this is necessary because there may be backward assertions in the pattern.
3406        When matching the same subject several times, it may save resources to use
3407        PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3408    
3409     4. The code for checking the validity of UTF-8 strings has been tightened so
3410        that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3411        containing "overlong sequences".
3412    
3413     5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3414        I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3415        should have been "&", but it just so happened that all the cases this let
3416        through by mistake were picked up later in the function.
3417    
3418     6. I had used a variable called "isblank" - this is a C99 function, causing
3419        some compilers to warn. To avoid this, I renamed it (as "blankclass").
3420    
3421     7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3422        prompting; (b) run "./pcretest /dev/null" at the start of the test script
3423        so the version is shown; (c) stop "make test" echoing "./RunTest".
3424    
3425     8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3426    
3427     9. The prototype for memmove() for systems that don't have it was using
3428        size_t, but the inclusion of the header that defines size_t was later. I've
3429        moved the #includes for the C headers earlier to avoid this.
3430    
3431    10. Added some adjustments to the code to make it easier to compiler on certain
3432        special systems:
3433    
3434          (a) Some "const" qualifiers were missing.
3435          (b) Added the macro EXPORT before all exported functions; by default this
3436              is defined to be empty.
3437          (c) Changed the dftables auxiliary program (that builds chartables.c) so
3438              that it reads its output file name as an argument instead of writing
3439              to the standard output and assuming this can be redirected.
3440    
3441    11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3442        class containing characters with values greater than 255, PCRE compilation
3443        went into a loop.
3444    
3445    12. A recursive reference to a subpattern that was within another subpattern
3446        that had a minimum quantifier of zero caused PCRE to crash. For example,
3447        (x(y(?2))z)? provoked this bug with a subject that got as far as the
3448        recursion. If the recursively-called subpattern itself had a zero repeat,
3449        that was OK.
3450    
3451    13. In pcretest, the buffer for reading a data line was set at 30K, but the
3452        buffer into which it was copied (for escape processing) was still set at
3453        1024, so long lines caused crashes.
3454    
3455    14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3456        "internal error: code overflow...". This applied to any character class
3457        that was followed by a possessive quantifier.
3458    
3459    15. Modified the Makefile to add libpcre.la as a prerequisite for
3460        libpcreposix.la because I was told this is needed for a parallel build to
3461        work.
3462    
3463    16. If a pattern that contained .* following optional items at the start was
3464        studied, the wrong optimizing data was generated, leading to matching
3465        errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3466        matching string must start with a or b or c. The correct conclusion for
3467        this pattern is that a match can start with any character.
3468    
3469    
3470    Version 4.4 13-Aug-03
3471    ---------------------
3472    
3473     1. In UTF-8 mode, a character class containing characters with values between
3474        127 and 255 was not handled correctly if the compiled pattern was studied.
3475        In fixing this, I have also improved the studying algorithm for such
3476        classes (slightly).
3477    
3478     2. Three internal functions had redundant arguments passed to them. Removal
3479        might give a very teeny performance improvement.
3480    
3481     3. Documentation bug: the value of the capture_top field in a callout is *one
3482        more than* the number of the hightest numbered captured substring.
3483    
3484     4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3485        in incorrectly linking with a previously installed version. They now link
3486        explicitly with libpcre.la.
3487    
3488     5. configure.in no longer needs to recognize Cygwin specially.
3489    
3490     6. A problem in pcre.in for Windows platforms is fixed.
3491    
3492     7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3493        pcretest, it used to include the size of the study block as part of its
3494        output. Unfortunately, the structure contains a field that has a different
3495        size on different hardware architectures. This meant that the tests that
3496        showed this size failed. As the block is currently always of a fixed size,
3497        this information isn't actually particularly useful in pcretest output, so
3498        I have just removed it.
3499    
3500     8. Three pre-processor statements accidentally did not start in column 1.
3501        Sadly, there are *still* compilers around that complain, even though
3502        standard C has not required this for well over a decade. Sigh.
3503    
3504     9. In pcretest, the code for checking callouts passed small integers in the
3505        callout_data field, which is a void * field. However, some picky compilers
3506        complained about the casts involved for this on 64-bit systems. Now
3507        pcretest passes the address of the small integer instead, which should get
3508        rid of the warnings.
3509    
3510    10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3511        both compile and run time, and gives an error if an invalid UTF-8 sequence
3512        is found. There is a option for disabling this check in cases where the
3513        string is known to be correct and/or the maximum performance is wanted.
3514    
3515    11. In response to a bug report, I changed one line in Makefile.in from
3516    
3517            -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3518        to
3519            -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3520    
3521        to look similar to other lines, but I have no way of telling whether this
3522        is the right thing to do, as I do not use Windows. No doubt I'll get told
3523        if it's wrong...
3524    
3525    
3526    Version 4.3 21-May-03
3527    ---------------------
3528    
3529    1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3530       Makefile.
3531    
3532    2. Some refactoring to improve the quality of the code:
3533    
3534       (i)   The utf8_table... variables are now declared "const".
3535    
3536       (ii)  The code for \cx, which used the "case flipping" table to upper case
3537             lower case letters, now just substracts 32. This is ASCII-specific,
3538             but the whole concept of \cx is ASCII-specific, so it seems
3539             reasonable.
3540    
3541       (iii) PCRE was using its character types table to recognize decimal and
3542             hexadecimal digits in the pattern. This is silly, because it handles
3543             only 0-9, a-f, and A-F, but the character types table is locale-
3544             specific, which means strange things might happen. A private
3545             table is now used for this - though it costs 256 bytes, a table is
3546             much faster than multiple explicit tests. Of course, the standard
3547             character types table is still used for matching digits in subject
3548             strings against \d.
3549    
3550       (iv)  Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3551             ending in _t are). So I've renamed it as ESC_tee.
3552    
3553    3. The first argument for regexec() in the POSIX wrapper should have been
3554       defined as "const".
3555    
3556    4. Changed pcretest to use malloc() for its buffers so that they can be
3557       Electric Fenced for debugging.
3558    
3559    5. There were several places in the code where, in UTF-8 mode, PCRE would try
3560       to read one or more bytes before the start of the subject string. Often this
3561       had no effect on PCRE's behaviour, but in some circumstances it could
3562       provoke a segmentation fault.
3563    
3564    6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3565       to try to read one or more bytes before the start of the subject string.
3566    
3567    7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3568       UTF-8 support could misbehave in various ways if the subject string
3569       contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3570       area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3571       back over UTF-8 characters.)
3572    
3573    
3574    Version 4.2 14-Apr-03
3575    ---------------------
3576    
3577    1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3578    
3579    2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3580         [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3581         [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3582         [WIN_PREFIX]: new variable, "cyg" for Cygwin
3583         * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3584           and BUILD_EXEEXT
3585         Note: automatic setting of the BUILD variables is not yet working
3586         set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3587           compile-time but not at link-time
3588         [LINK]: use for linking executables only
3589         make different versions for Windows and non-Windows
3590         [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3591           libraries
3592         [LINK_FOR_BUILD]: new variable
3593         [OBJEXT]: use throughout
3594         [EXEEXT]: use throughout
3595         <winshared>: new target
3596         <wininstall>: new target
3597         <dftables.o>: use native compiler
3598         <dftables>: use native linker
3599         <install>: handle Windows platform correctly
3600         <clean>: ditto
3601         <check>: ditto
3602         copy DLL to top builddir before testing
3603    
3604       As part of these changes, -no-undefined was removed again. This was reported
3605       to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3606       in any case.
3607    
3608    3. Some tidies to get rid of compiler warnings:
3609    
3610       . In the match_data structure, match_limit was an unsigned long int, whereas
3611         match_call_count was an int. I've made them both unsigned long ints.
3612    
3613       . In pcretest the fact that a const uschar * doesn't automatically cast to
3614         a void * provoked a warning.
3615    
3616       . Turning on some more compiler warnings threw up some "shadow" variables
3617         and a few more missing casts.
3618    
3619    4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3620       option, a class that contained a single character with a value between 128
3621       and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3622    
3623    5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3624       option, a class that contained several characters, but with at least one
3625       whose value was between 128 and 255 caused PCRE to crash.
3626    
3627    
3628    Version 4.1 12-Mar-03
3629    ---------------------
3630    
3631    1. Compiling with gcc -pedantic found a couple of places where casts were
3632    needed, and a string in dftables.c that was longer than standard compilers are
3633    required to support.
3634    
3635    2. Compiling with Sun's compiler found a few more places where the code could
3636    be tidied up in order to avoid warnings.
3637    
3638    3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3639    first of these names is deprecated in the latest Autoconf in favour of the name
3640    CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3641    compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3642    analogy I have changed it to CFLAGS_FOR_BUILD.
3643    
3644    4. Added -no-undefined to the linking command in the Makefile, because this is
3645    apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3646    linking step for the pcreposix library.
3647    
3648    5. PCRE was failing to diagnose the case of two named groups with the same
3649    name.
3650    
3651    6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3652    literal character that is needed in the subject for a match, and scans along to
3653    ensure that it is present before embarking on the full matching process. This
3654    saves time in cases of nested unlimited repeats that are never going to match.
3655    Problem: the scan can take a lot of time if the subject is very long (e.g.
3656    megabytes), thus penalizing straightforward matches. It is now done only if the
3657    amount of subject to be scanned is less than 1000 bytes.
3658    
3659    7. A lesser problem with the same optimization is that it was recording the
3660    first character of an anchored pattern as "needed", thus provoking a search
3661    right along the subject, even when the first match of the pattern was going to
3662    fail. The "needed" character is now not set for anchored patterns, unless it
3663    follows something in the pattern that is of non-fixed length. Thus, it still
3664    fulfils its original purpose of finding quick non-matches in cases of nested
3665    unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3666    
3667    
3668    Version 4.0 17-Feb-03
3669    ---------------------
3670    
3671    1. If a comment in an extended regex that started immediately after a meta-item
3672    extended to the end of string, PCRE compiled incorrect data. This could lead to
3673    all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3674    
3675    2. Moved to autoconf 2.53 and libtool 1.4.2.
3676    
3677    3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3678    the special perltest8 script is no longer needed - all the tests can be run
3679    from a single perltest script.
3680    
3681    4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3682    by \s. It has now been removed in PCRE. This means it isn't recognized as
3683    whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3684    class [:space:] *does* include VT, thereby creating a mess.
3685    
3686    5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3687    space and tab.
3688    
3689    6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3690    its new features into the main test script, reducing the number of scripts.
3691    
3692    7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3693    were backward compatible, and made the (?i) apply to the whole pattern, as if
3694    /i were given. Now it behaves more logically, and applies the option setting
3695    only to what follows. PCRE has been changed to follow suit. However, if it
3696    finds options settings right at the start of the pattern, it extracts them into
3697    the global options, as before. Thus, they show up in the info data.
3698    
3699    8. Added support for the \Q...\E escape sequence. Characters in between are
3700    treated as literals. This is slightly different from Perl in that $ and @ are
3701    also handled as literals inside the quotes. In Perl, they will cause variable
3702    interpolation. Note the following examples:
3703    
3704        Pattern            PCRE matches      Perl matches
3705    
3706        \Qabc$xyz\E        abc$xyz           abc followed by the contents of $xyz
3707        \Qabc\$xyz\E       abc\$xyz          abc\$xyz
3708        \Qabc\E\$\Qxyz\E   abc$xyz           abc$xyz
3709    
3710    For compatibility with Perl, \Q...\E sequences are recognized inside character
3711    classes as well as outside them.
3712    
3713    9. Re-organized 3 code statements in pcretest to avoid "overflow in
3714    floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3715    (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3716    signed/unsigned warnings.
3717    
3718    10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3719    option for pcretest, so I've replaced it by a simple function that does just
3720    that job.
3721    
3722    11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3723    "pcregrep -".
3724    
3725    12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3726    Java package. This provides some syntactic sugar for simple cases of what my
3727    documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3728    as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3729    item, you can use this simplified notation. Note that only makes sense with
3730    greedy quantifiers. Consequently, the use of the possessive quantifier forces
3731    greediness, whatever the setting of the PCRE_UNGREEDY option.
3732    
3733    13. A change of greediness default within a pattern was not taking effect at
3734    the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3735    subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3736    was abstracted outside.
3737    
3738    14. PCRE now supports the \G assertion. It is true when the current matching
3739    position is at the start point of the match. This differs from \A when the
3740    starting offset is non-zero. Used with the /g option of pcretest (or similar
3741    code), it works in the same way as it does for Perl's /g option. If all
3742    alternatives of a regex begin with \G, the expression is anchored to the start
3743    match position, and the "anchored" flag is set in the compiled expression.
3744    
3745    15. Some bugs concerning the handling of certain option changes within patterns
3746    have been fixed. These applied to options other than (?ims). For example,
3747    "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3748    been the other way round. Some of this was related to change 7 above.
3749    
3750    16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3751    features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3752    and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3753    POSIX classes only within a class (e.g. /[[:alpha:]]/).
3754    
3755    17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3756    mode. Unlike ".", it always matches newline, whatever the setting of
3757    PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3758    assertions. Perl allows it, but it doesn't (in general) work because it can't
3759    calculate the length of the lookbehind. At least, that's the case for Perl
3760    5.8.0 - I've been told they are going to document that it doesn't work in
3761    future.
3762    
3763    18. Added an error diagnosis for escapes that PCRE does not support: these are
3764    \L, \l, \N, \P, \p, \U, \u, and \X.
3765    
3766    19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3767    reading past the end of the pattern in cases such as /[abcd/.
3768    
3769    20. PCRE was getting more memory than necessary for patterns with classes that
3770    contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3771    
3772    21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3773    compiling PCRE for use with Virtual Pascal.
3774    
3775    22. Small fix to the Makefile to make it work properly if the build is done
3776    outside the source tree.
3777    
3778    23. Added a new extension: a condition to go with recursion. If a conditional
3779    subpattern starts with (?(R) the "true" branch is used if recursion has
3780    happened, whereas the "false" branch is used only at the top level.
3781    
3782    24. When there was a very long string of literal characters (over 255 bytes
3783    without UTF support, over 250 bytes with UTF support), the computation of how
3784    much memory was required could be incorrect, leading to segfaults or other
3785    strange effects.
3786    
3787    25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3788    start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3789    there was a subsequent back reference to those brackets. This meant that, for
3790    example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3791    possible to check for precisely this case. All we can do is abandon the
3792    optimization if .* occurs inside capturing brackets when there are any back
3793    references whatsoever. (See below for a better fix that came later.)
3794    
3795    26. The handling of the optimization for finding the first character of a
3796    non-anchored pattern, and for finding a character that is required later in the
3797    match were failing in some cases. This didn't break the matching; it just
3798    failed to optimize when it could. The way this is done has been re-implemented.
3799    
3800    27. Fixed typo in error message for invalid (?R item (it said "(?p").
3801    
3802    28. Added a new feature that provides some of the functionality that Perl
3803    provides with (?{...}). The facility is termed a "callout". The way it is done
3804    in PCRE is for the caller to provide an optional function, by setting
3805    pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3806    global variable. By default it is unset, which disables all calling out. To get
3807    the function called, the regex must include (?C) at appropriate points. This
3808    is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3809    This provides a means of identifying different callout points. When PCRE
3810    reaches such a point in the regex, if pcre_callout has been set, the external
3811    function is called. It is provided with data in a structure called
3812    pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3813    matching continues; if it returns a non-zero value, the match at the current
3814    point fails. However, backtracking will occur if possible. [This was changed
3815    later and other features added - see item 49 below.]
3816    
3817    29. pcretest is upgraded to test the callout functionality. It provides a
3818    callout function that displays information. By default, it shows the start of
3819    the match and the current position in the text. There are some new data escapes
3820    to vary what happens:
3821    
3822        \C+         in addition, show current contents of captured substrings
3823        \C-         do not supply a callout function
3824        \C!n        return 1 when callout number n is reached
3825        \C!n!m      return 1 when callout number n is reached for the mth time
3826    
3827    30. If pcregrep was called with the -l option and just a single file name, it
3828    output "<stdin>" if a match was found, instead of the file name.
3829    
3830    31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3831    slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3832    pcre_exec(). This saves a malloc/free per call. The default value of
3833    POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3834    when configuring.
3835    
3836    32. The default maximum size of a compiled pattern is 64K. There have been a
3837    few cases of people hitting this limit. The code now uses macros to handle the
3838    storing of links as offsets within the compiled pattern. It defaults to 2-byte
3839    links, but this can be changed to 3 or 4 bytes by --with-link-size when
3840    configuring. Tests 2 and 5 work only with 2-byte links because they output
3841    debugging information about compiled patterns.
3842    
3843    33. Internal code re-arrangements:
3844    
3845    (a) Moved the debugging function for printing out a compiled regex into
3846        its own source file (printint.c) and used #include to pull it into
3847        pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3848        separate copies.
3849    
3850    (b) Defined the list of op-code names for debugging as a macro in
3851        internal.h so that it is next to the definition of the opcodes.
3852    
3853    (c) Defined a table of op-code lengths for simpler skipping along compiled
3854        code. This is again a macro in internal.h so that it is next to the
3855        definition of the opcodes.
3856    
3857    34. Added support for recursive calls to individual subpatterns, along the
3858    lines of Robin Houston's patch (but implemented somewhat differently).
3859    
3860    35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3861    allow it to read and process whole directories in Win32. This code was
3862    contributed by Lionel Fourquaux; it has not been tested by me.
3863    
3864    36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3865    used to name a group. Names consist of alphanumerics and underscores, and must
3866    be unique. Back references use the syntax (?P=name) and recursive calls use
3867    (?P>name) which is a PCRE extension to the Python extension. Groups still have
3868    numbers. The function pcre_fullinfo() can be used after compilation to extract
3869    a name/number map. There are three relevant calls:
3870    
3871      PCRE_INFO_NAMEENTRYSIZE        yields the size of each entry in the map
3872      PCRE_INFO_NAMECOUNT            yields the number of entries
3873      PCRE_INFO_NAMETABLE            yields a pointer to the map.
3874    
3875    The map is a vector of fixed-size entries. The size of each entry depends on
3876    the length of the longest name used. The first two bytes of each entry are the
3877    group number, most significant byte first. There follows the corresponding
3878    name, zero terminated. The names are in alphabetical order.
3879    
3880    37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3881    case instead of 255. Making it the same both with and without UTF-8 support
3882    means that the same test output works with both.
3883    
3884    38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3885    calling malloc() with a zero argument.
3886    
3887    39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3888    optimization. I've improved things by keeping a bitmap of backreferences with
3889    numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3890    fact referenced, the optimization can be applied. It is unlikely that a
3891    relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3892    the match to follow \n) will appear inside brackets with a number greater than
3893    31, but if it does, any back reference > 31 suppresses the optimization.
3894    
3895    40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3896    of disabling numbered capturing parentheses. Any opening parenthesis that is
3897    not followed by ? behaves as if it were followed by ?: but named parentheses
3898    can still be used for capturing (and they will acquire numbers in the usual
3899    way).
3900    
3901    41. Redesigned the return codes from the match() function into yes/no/error so
3902    that errors can be passed back from deep inside the nested calls. A malloc
3903    failure while inside a recursive subpattern call now causes the
3904    PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3905    
3906    42. It is now possible to set a limit on the number of times the match()
3907    function is called in a call to pcre_exec(). This facility makes it possible to
3908    limit the amount of recursion and backtracking, though not in a directly
3909    obvious way, because the match() function is used in a number of different
3910    circumstances. The count starts from zero for each position in the subject
3911    string (for non-anchored patterns). The default limit is, for compatibility, a
3912    large number, namely 10 000 000. You can change this in two ways:
3913    
3914    (a) When configuring PCRE before making, you can use --with-match-limit=n
3915        to set a default value for the compiled library.
3916    
3917    (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3918        a different value is set. See 45 below.
3919    
3920    If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3921    
3922    43. Added a new function pcre_config(int, void *) to enable run-time extraction
3923    of things that can be changed at compile time. The first argument specifies
3924    what is wanted and the second points to where the information is to be placed.
3925    The current list of available information is:
3926    
3927      PCRE_CONFIG_UTF8
3928    
3929    The output is an integer that is set to one if UTF-8 support is available;
3930    otherwise it is set to zero.
3931    
3932      PCRE_CONFIG_NEWLINE
3933    
3934    The output is an integer that it set to the value of the code that is used for